Nothing Special   »   [go: up one dir, main page]
Next Issue
Volume 7, October
Previous Issue
Volume 7, August
You seem to have javascript disabled. Please note that many of the page functionalities won't work as expected without javascript enabled.
 
 
Search for Articles:
Title / Keyword
Author / Affiliation / Email
Journal
Article Type
 
 
Advanced Search
 
Section
Special Issue
Volume
Issue
Number
Page
 
8.3
4.2
Journals
Remote Sensing
Volume 7
Issue 9
remotesensing-logo

Journal Menu

► Journal Menu

Journal Browser

► Journal Browser
share announcement

Remote Sens., Volume 7, Issue 9 (September 2015) – 73 articles , Pages 11016-12587

  • Issues are regarded as officially published after their release is announced to the table of contents alert mailing list.
  • You may sign up for e-mail alerts to receive table of contents of newly released issues.
  • PDF is the official format for papers published in both, html and pdf forms. To view the papers in pdf format, click on the "PDF Full-text" link, and use the free Adobe Reader to open them.
Order results
Result details
Section
Show export options expand_more Show export options expand_less
Select all
Export citation of selected articles as:
1193 KiB  
Article
Mapping Forest Canopy Height Across Large Areas by Upscaling ALS Estimates with Freely Available Satellite Data
by Phil Wilkes, Simon D. Jones, Lola Suarez, Andrew Mellor, William Woodgate, Mariela Soto-Berelov, Andrew Haywood and Andrew K. Skidmore
Remote Sens. 2015, 7(9), 12563-12587; https://doi.org/10.3390/rs70912563 - 23 Sep 2015
Cited by 46 | Viewed by 10589
Abstract
Operational assessment of forest structure is an on-going challenge for land managers, particularly over large, remote or inaccessible areas. Here, we present an easily adopted method for generating a continuous map of canopy height at a 30 m resolution, demonstrated over 2.9 million [...] Read more.
Operational assessment of forest structure is an on-going challenge for land managers, particularly over large, remote or inaccessible areas. Here, we present an easily adopted method for generating a continuous map of canopy height at a 30 m resolution, demonstrated over 2.9 million hectares of highly heterogeneous forest (canopy height 0–70 m) in Victoria, Australia. A two-stage approach was utilized where Airborne Laser Scanning (ALS) derived canopy height, captured over ~18% of the study area, was used to train a regression tree ensemble method; random forest. Predictor variables, which have a global coverage and are freely available, included Landsat Thematic Mapper (Tasselled Cap transformed), Moderate Resolution Imaging Spectroradiometer Normalized Difference Vegetation Index time series, Shuttle Radar Topography Mission elevation data and other ancillary datasets. Reflectance variables were further processed to extract additional spatial and temporal contextual and textural variables. Modeled canopy height was validated following two approaches; (i) random sample cross validation; and (ii) with 108 inventory plots from outside the ALS capture extent. Both the cross validation and comparison with inventory data indicate canopy height can be estimated with a Root Mean Square Error (RMSE) of ≤ 31% (~5.6 m) at the 95th percentile confidence interval. Subtraction of the systematic component of model error, estimated from training data error residuals, rescaled canopy height values to more accurately represent the response variable distribution tails e.g., tall and short forest. Two further experiments were carried out to test the applicability and scalability of the presented method. Results suggest that (a) no improvement in canopy height estimation is achieved when models were constructed and validated for smaller geographic areas, suggesting there is no upper limit to model scalability; and (b) training data can be captured over a small percentage of the study area (~6%) if response and predictor variable variance is captured within the training cohort, however RMSE is higher than when compared to a stratified random sample. Full article
►▼ Show Figures

Graphical abstract

Graphical abstract
Full article ">Figure 1
<p>Study area in east Victoria, Australia. A mosaic of 5 Landsat TM false color composite images covering the study area (outlined in white) and location of the study area within Australia (inset) (<b>A</b>). Canopy height derived from ALS capture where canopy height values are aggregated into 10 × 10 km cells (grey indicates no data) (<b>B</b>). The extent of the ALS capture (<b>C</b>). Forest extent [<a href="#B35-remotesensing-07-12563" class="html-bibr">35</a>] and location of Victorian Forest Monitoring Programme forest inventory plots (VFMP) (<b>D</b>). Map coordinate system is the projected Map Grid of Australia (MGA) zone 55.</p> Full article ">Figure 2
<p>Relative importance of the 18 variables selected for the final random forest model. Confidence intervals (95th percentile) for variable importance were calculated in a bootstrap (<span class="html-italic">N</span> = 50). Increase mean square error (MSE) is the mean of the squared prediction error when the variable is permuted for a random variable [<a href="#B84-remotesensing-07-12563" class="html-bibr">84</a>]. Numbers in brackets indicate the kernel size.</p> Full article ">Figure 3
<p>ALS derived canopy height (<span class="html-italic">H</span><sub>95</sub>) compared to model residual error for random forest models (<span class="html-italic">H</span><sub>RF</sub>); constructed using 5000 ALS plots of untransformed response data (<b>A</b>), 5000 ALS plots where the response variable was resampled to a uniform distribution (<b>B</b>), and subtraction of systematic error from modeled canopy height (<b>C</b>). The coefficient of determination values (<span class="html-italic">r</span><sup>2</sup>) were calculated from a linear regression of measured canopy height and model residuals.</p> Full article ">Figure 4
<p>A comparison of the response variable distribution (<b>A</b>), range of Tasselled Cap wetness values (3 × 3 pixels) (<b>B</b>) and the distributions of modeled canopy height (random forest (<b>C</b>) and random forest-systematic error (<b>D</b>)) for different height classes. The solid arrows indicate the direction in which the random forest model output was “squeezed” by inequity in response variable distribution; the dashed arrow indicates the direction canopy height values were rescaled after correcting for systematic error.</p> Full article ">Figure 5
<p>Canopy height at a 30 m resolution (clipped to forest extent) generated using random forest−systematic error (<b>A</b>). Model output when compared to ALS derived canopy height (10 km × 10 km resolution) where error is represented as height difference (<b>B</b>) and percentage of height (<b>C</b>). Coordinate system is the projected Map Grid of Australia (MGA) zone 55.</p> Full article ">Figure 6
<p>A map of canopy height (30 m resolution) for an area of 16.5 km × 16.5 km, highlighting the land use history of an area of mixed-use forest (<b>left</b>). Three logging coupes that were clear-felled between 2002 and 2009 (<b>A</b>–<b>C</b>) and an area that has never been logged (<b>D</b>) are singled out (coupe extents outlined in black) (center column). Transects of canopy height across each coupe indicating regrowth since logging, coupe boundaries are also identified (<b>right</b>). Coordinate system is the projected Map Grid of Australia (MGA) zone 55.</p> Full article ">Figure 7
<p>A comparison of inventory measured canopy height (<span class="html-italic">H</span><sub>inv</sub>) with random forest (<span class="html-italic">H</span><sub>RF</sub>) (<b>A</b>) and random forest corrected for systematic error (<span class="html-italic">H</span><sub>RF-SE</sub>) (<b>B</b>) estimated canopy height, at 108 forest inventory plots.</p> Full article ">Figure 8
<p>A comparison of inventory (<span class="html-italic">H</span><sub>inv</sub>) measured canopy height and ALS (<span class="html-italic">H</span><sub>ALS</sub>) and random forest-systematic error (<span class="html-italic">H</span><sub>RF-SE</sub>) estimated canopy height, for 22 plots from within the ALS capture area. Vertical dotted lines link the same plot estimated with ALS or random forest.</p> Full article ">Figure 9
<p>Error in canopy height estimates when constructing random forest models from ALS data selected to represent a combination of a number of disparate (non-random) acquisitions. Model output was validated with ALS plots from outside the training area. For comparison (see boxplot), the results from bootstrapping (<span class="html-italic">N</span> = 50) random forest trained with a random stratified (by IBRA bioregion) sample from across the whole study area (~18% of forested area) is included.</p> Full article ">
1663 KiB  
Article
Flood Mapping Based on Multiple Endmember Spectral Mixture Analysis and Random Forest Classifier—The Case of Yuyao, China
by Quanlong Feng, Jianhua Gong, Jiantao Liu and Yi Li
Remote Sens. 2015, 7(9), 12539-12562; https://doi.org/10.3390/rs70912539 - 23 Sep 2015
Cited by 47 | Viewed by 8543
Abstract
Remote sensing is recognized as a valuable tool for flood mapping due to its synoptic view and continuous coverage of the flooding event. This paper proposed a hybrid approach based on multiple endmember spectral analysis (MESMA) and Random Forest classifier to extract inundated [...] Read more.
Remote sensing is recognized as a valuable tool for flood mapping due to its synoptic view and continuous coverage of the flooding event. This paper proposed a hybrid approach based on multiple endmember spectral analysis (MESMA) and Random Forest classifier to extract inundated areas in Yuyao City in China using medium resolution optical imagery. MESMA was adopted to tackle the mixing pixel problem induced by medium resolution data. Specifically, 35 optimal endmembers were selected to construct a total of 3111 models in the MESMA procedure to derive accurate fraction information. A multi-dimensional feature space was constructed including the normalized difference water index (NDWI), topographical parameters of height, slope, and aspect together with the fraction maps. A Random Forest classifier consisting of 200 decision trees was adopted to classify the post-flood image based on the above multi-features. Experimental results indicated that the proposed method can extract the inundated areas precisely with a classification accuracy of 94% and a Kappa index of 0.88. The inclusion of fraction information can help improve the mapping accuracy with an increase of 2.5%. Moreover, the proposed method also outperformed the maximum likelihood classifier and the NDWI thresholding method. This research provided a useful reference for flood mapping using medium resolution optical remote sensing imagery. Full article
(This article belongs to the Special Issue Remote Sensing in Flood Monitoring and Management)
►▼ Show Figures

Graphical abstract

Graphical abstract
Full article ">Figure 1
<p>Study area. Red-Green-Blue composition: near-infrared, red and green bands of the multispectral charge coupled device camera (HJ-CCD) after the flood.</p> Full article ">Figure 2
<p>Workflow of this study.</p> Full article ">Figure 3
<p>Spectral reflectance of endmembers including (<b>a</b>) water; (<b>b</b>) vegetation; (<b>c</b>) impervious surface; (<b>d</b>) soil.</p> Full article ">Figure 4
<p>Spatial distribution of training and testing points for “water” class.</p> Full article ">Figure 5
<p>Fraction maps derived from multiple endmember spectral analysis (MESMA) including (<b>a</b>) water; (<b>b</b>) vegetation; (<b>c</b>) impervious surface; (<b>d</b>) soil.</p> Full article ">Figure 6
<p>Out-of-bag (OOB) error <span class="html-italic">vs.</span> <span class="html-italic">ntree</span>.</p> Full article ">Figure 7
<p>Flooded areas generated through RF classifier.</p> Full article ">Figure 8
<p>Land cover types before the flood event.</p> Full article ">Figure 9
<p>Importance of input variables.</p> Full article ">
2334 KiB  
Article
Efficient Wetland Surface Water Detection and Monitoring via Landsat: Comparison with in situ Data from the Everglades Depth Estimation Network
by John W. Jones
Remote Sens. 2015, 7(9), 12503-12538; https://doi.org/10.3390/rs70912503 - 23 Sep 2015
Cited by 103 | Viewed by 10997
Abstract
The U.S. Geological Survey is developing new Landsat science products. One, named Dynamic Surface Water Extent (DSWE), is focused on the representation of ground surface inundation as detected in cloud-/shadow-/snow-free pixels for scenes collected over the U.S. and its territories. Characterization of DSWE [...] Read more.
The U.S. Geological Survey is developing new Landsat science products. One, named Dynamic Surface Water Extent (DSWE), is focused on the representation of ground surface inundation as detected in cloud-/shadow-/snow-free pixels for scenes collected over the U.S. and its territories. Characterization of DSWE uncertainty to facilitate its appropriate use in science and resource management is a primary objective. A unique evaluation dataset developed from data made publicly available through the Everglades Depth Estimation Network (EDEN) was used to evaluate one candidate DSWE algorithm that is relatively simple, requires no scene-based calibration data, and is intended to detect inundation in the presence of marshland vegetation. A conceptual model of expected algorithm performance in vegetated wetland environments was postulated, tested and revised. Agreement scores were calculated at the level of scenes and vegetation communities, vegetation index classes, water depths, and individual EDEN gage sites for a variety of temporal aggregations. Landsat Archive cloud cover attribution errors were documented. Cloud cover had some effect on model performance. Error rates increased with vegetation cover. Relatively low error rates for locations of little/no vegetation were unexpectedly dominated by omission errors due to variable substrates and mixed pixel effects. Examined discrepancies between satellite and in situ modeled inundation demonstrated the utility of such comparisons for EDEN database improvement. Importantly, there seems no trend or bias in candidate algorithm performance as a function of time or general hydrologic conditions, an important finding for long-term monitoring. The developed database and knowledge gained from this analysis will be used for improved evaluation of candidate DSWE algorithms as well as other measurements made on Everglades surface inundation, surface water heights and vegetation using radar, lidar and hyperspectral instruments. Although no other sites have such an extensive in situ network or long-term records, the broader applicability of this and other candidate DSWE algorithms is being evaluated in other wetlands using this work as a guide. Continued interaction among DSWE producers and potential users will help determine whether the measured accuracies are adequate for practical utility in resource management. Full article
(This article belongs to the Special Issue Towards Remote Long-Term Monitoring of Wetland Landscapes)
►▼ Show Figures

Graphical abstract

Graphical abstract
Full article ">Figure 1
<p>The Greater Everglades region of South Florida, USA. The digital elevation model covering the EDEN model domain, shows elevation in meters above sea level [<a href="#B8-remotesensing-07-12503" class="html-bibr">8</a>] and illustrates the extremely flat nature of the Everglades. Subtle differences in topography create variations in hydroperiod that dictate vegetation composition. Reliable information on inundation dynamics would be very beneficial to restoration science and resource management. Non-tidally influenced areas of Water Conservation Areas (WCA) 1–3 are labelled simply as 1, 2A, 2B, 3AN, 3AS, and 3B.</p> Full article ">Figure 2
<p>Spectra for dominant Everglades marshland covers as well as common woody plants that border Everglades marshlands. The inundation status of each vegetation type as noted at the time of collection and number of samples are provided in parentheses.</p> Full article ">Figure 3
<p>The result of resampling the spectra shown in <a href="#remotesensing-07-12503-f002" class="html-fig">Figure 2</a> to TM5 band reflectances. The MNDWI values calculated from the resampled spectra and sample size for each are shown in parentheses. Inundated sawgrass and periphyton are distinguished from dry sawgrass and periphyton based on the −0.5 MNDWI threshold. However, additional thresholds on TM band 4 and TM band 7 are needed to prevent the incorrect labeling of other dry cover types such as the willow sample as inundated.</p> Full article ">Figure 4
<p>Open (<span class="html-italic">i.e.</span>, vegetation free) water is relatively rare in the Everglades. Therefore, Landsat-based classification techniques that identify pure water pixels suggest the vast Everglades “river of grass” is overly dry. In contrast, the application of Equations (2) and (3) yields maps of inundation for the driest (<b>left</b>) and wettest (<b>right</b>) image dates in the study image database. Blue pixels are those that met the trial model criteria and were not masked as cloud/cloud shadow (white pixels).</p> Full article ">Figure 5
<p>Procedure used to combine publicly available EDEN gage and ancillary data and associate them with values sampled from publicly available Landsat Land Surface Reflectance data.</p> Full article ">Figure 6
<p>Histogram of overall agreement at the scene level across all 50 study dates (<a href="#sec3dot1-remotesensing-07-12503" class="html-sec">Section 3.1</a>). The Landsat data themselves, <span class="html-italic">in situ</span> collected data on vegetation cover and high resolution imagery all provide insights regarding sources of error.</p> Full article ">Figure 7
<p>False color composite (Band 5: Red; Band 4: Green, and Band 3: Blue) for a Landsat scene with cloud contamination that is grossly underestimated in the Archive database but which was used in the analysis (<b>Left</b>), and a Landsat scene with nearly no cloud cover over the wetlands of interest (<b>Right</b>) that was excluded from analysis given its Archive attribute of 28% cloud cover.</p> Full article ">Figure 8
<p>Landsat cloud scores attributed to scenes in Earth Explorer (Y-axis) <span class="html-italic">versus</span> percent of the EDEN study domain covered by clouds on the corresponding image date as calculated for Landsat Surface Reflectance provisional data generation. This demonstrates both the inaccuracy of the information accessed when querying Earth Explorer and the broad range of actual cloudiness represented by the study input data.</p> Full article ">Figure 9
<p>Scene-level overall agreement <span class="html-italic">versus</span> the number of gages obscured by clouds as assessed by sampling the cloud mask associated with each date. Although the goal is to detect inundation for all cloud/cloud shadow free pixels, overall model performance is degraded by the presence of a cloudy atmosphere.</p> Full article ">Figure 10
<p>Agreement rate for each sample point as a function of its average NDVI value across all study image dates. While scene-level agreement decreases with increasing vegetation cover, NDVI discerned from the satellite data themselves fails to explain model performance at this scale.</p> Full article ">Figure 11
<p>The distribution of NDVI for all clear (non-cloudy, not shaded) observations on EDEN gages across all study image dates was used to partition the samples into low, medium and high vegetation cover classes. While negative NDVI values are sometimes used to detect water, few Everglades locations would ever seem to be inundated based on that criterion.</p> Full article ">Figure 12
<p>Mean agreement rates and agreement variation for gages grouped by field-determined community type. Values in parentheses represent the number of gages classified as that type. The number of observations for that type is larger given the 50-date study image database. Lowest agreements occurred in both upland (forest) and wetland (marsh) classes in which a minority of EDEN gages was located.</p> Full article ">Figure 13
<p>Mean omission rates and variation for gages grouped by field-determined community type. Values in parentheses represent the number of gages classified as that type. The number of observations for that type is larger given the 50-date study image database. Highest omission error rates occur in areas of dense vegetation cover with higher variability in flooding, that is, forests and ridges or emergent marshes. Note that a “ridge” in the Everglades is routinely less than a meter above surrounding slough environments.</p> Full article ">Figure 14
<p>Frequency of agreement rate across all individual observations in the study data base irrespective of date. The distribution favors high agreement rates on-the-whole. Poorly performing sites tend to be persistent.</p> Full article ">Figure 15
<p>A multivariate map showing the locations of individual EDEN gages, the number of observations at each gage displayed on top of agreement rates for each gage given the 50 scene database spanning 11 years and a false color composite Landsat image backdrop (Band 5: Red; Band 4: Green; and Band 3: Blue).</p> Full article ">Figure 16
<p>Frequency of mean agreement rate for only those gages that had more than 30 observations and for which the agreement rate was below 50%. Concentration on these provides explanation of model performance. Both the model and <span class="html-italic">in situ</span> data shortcomings are responsible.</p> Full article ">Figure 17
<p>Temporal traces of scaled Landsat band reflectances, water depths modeled using EDEN stage and ground elevation estimates, and agreement (1 = yes, 0 = no, co-located with the Depth axis) given the thresholds applied to MNDWI, TM Band 4 and Band 7 reflectance.</p> Full article ">Figure 18
<p>Overall agreement (<span class="html-italic">i.e.</span>, total agreement at the level of individual study dates) as a function of time for every scene in the 10-year study period. The lack of any significant trend in the relationship suggests the accuracy of the method is relatively stable through time, an important characteristic for the application of time series data on inundation.</p> Full article ">Figure 19
<p>Regression of agreement against mean depth for all 50 image study dates. Model performance is not related to overall inundation conditions—again suggesting the general approach is appropriate for long-term monitoring.</p> Full article ">Figure 20
<p>Overall agreement (<span class="html-italic">i.e.</span>, total agreement at the level of individual study dates) as a function of time for the top 25 dates ranked in terms of overall agreement given the 10-year study period. The residuals are dramatically reduced (<span class="html-italic">i.e.</span>, are on the order of ± 0.05). Unfortunately, cloud cover is not the sole determinant of model performance (<a href="#remotesensing-07-12503-t008" class="html-table">Table 8</a>).</p> Full article ">
2494 KiB  
Article
Seasonal Variation of Colored Dissolved Organic Matter in Barataria Bay, Louisiana, Using Combined Landsat and Field Data
by Ishan Joshi and Eurico J. D’Sa
Remote Sens. 2015, 7(9), 12478-12502; https://doi.org/10.3390/rs70912478 - 23 Sep 2015
Cited by 49 | Viewed by 9751
Abstract
Coastal bays, such as Barataria Bay, are important transition zones between the terrigenous and marine environments that are also optically complex due to elevated amounts of particulate and dissolved constituents. Monthly field data collected over a period of 15 months in 2010 and [...] Read more.
Coastal bays, such as Barataria Bay, are important transition zones between the terrigenous and marine environments that are also optically complex due to elevated amounts of particulate and dissolved constituents. Monthly field data collected over a period of 15 months in 2010 and 2011 in Barataria Bay were used to develop an empirical band ratio algorithm for the Landsat-5 TM that showed a good correlation with the Colored Dissolved Organic Matter (CDOM) absorption coefficient at 355 nm (ag355) (R2 = 0.74). Landsat-derived CDOM maps generally captured the major details of CDOM distribution and seasonal influences, suggesting the potential use of Landsat imagery to monitor biogeochemistry in coastal water environments. An investigation of the seasonal variation in ag355 conducted using Landsat-derived ag355 as well as field data suggested the strong influence of seasonality in the different regions of the bay with the marine end members (lower bay) experiencing generally low but highly variable ag355 and the freshwater end members (upper bay) experiencing high ag355 with low variability. Barataria Bay experienced a significant increase in ag355 during the freshwater release at the Davis Pond Freshwater Diversion (DPFD) following the Deep Water Horizon oil spill in 2010 and following the Mississippi River (MR) flood conditions in 2011, resulting in a weak linkage to salinity in comparison to the other seasons. Tree based statistical analysis showed the influence of high river flow conditions, high- and low-pressure systems that appeared to control ag355 by ~28%, 29% and 43% of the time duration over the study period at the marine end member just outside the bay. An analysis of CDOM variability in 2010 revealed the strong influence of the MR in controlling CDOM abundance in the lower bay during the high flow conditions, while strong winds associated with cold fronts significantly increase CDOM abundance in the upper bay, thus revealing the important role these events play in the CDOM dynamics of the bay. Full article
(This article belongs to the Special Issue Remote Sensing of Water Resources)
►▼ Show Figures

Graphical abstract

Graphical abstract
Full article ">Figure 1
<p>Barataria Bay, Louisiana, USA. Sampling stations are plotted along the transect from the marine end member (Station 1) to the freshwater end member (Station 15). The black squares represent the approximate locations of the Davis Pond Freshwater Diversion (DPFD), Baton Rouge, and Belle Chasse; Little Lake is represented by a triangle. MR is the Mississippi River and LP is Lake Pontchartrain.</p> Full article ">Figure 2
<p>Mean daily discharge (m<sup>3</sup>∙s<sup>−1</sup>) of the Mississippi River at Baton Rouge (red line) and Belle Chasse (blue line) (<b>1</b>), mean daily discharge (m<sup>3</sup>∙s<sup>−1</sup>) at Davis Pond Freshwater Diversion (DPFD) (<b>2</b>), mean daily wind speed (ms<sup>−1</sup>) and mean daily wind direction (in degrees from the North) (<b>3</b>). Black boxes represent the MR flood—The Mississippi River flood condition (May 2011) (1), DPFD opening—A large amount of freshwater release at Davis Pond Freshwater Diversion (April and July, 2010) (2), and the evidences of the high-pressure systems (e.g., cold fronts) during the field observations (February and November, 2010) (panel-3); In (3), the blue, red, green, and black colors represent the winter (November, December, January, and February), the spring (March, April, May, and June), the summer (July and August) and the fall (September and October), respectively.</p> Full article ">Figure 3
<p>Mean (<b>a</b>) salinity, (<b>b</b>) temperature (solid lines represent trends), (<b>c</b>) CDOM absorption coefficient (<span class="html-italic">a<sub>g</sub>355</span>), and (<b>d</b>) CDOM spectral slope (<span class="html-italic">S<sub>275–295</sub></span>) shown for Stations 1–15 (marine to freshwater end member) for summer (filled circles), winter (open circles), spring (inverted triangles) and fall (open triangle). Vertical dashed lines represent the different regions in Barataria Bay, namely: lower bay (Stations 1–5), central bay (Stations 6–10) and upper bay (Stations 11–15).</p> Full article ">Figure 4
<p>CDOM absorption coefficient (355nm) <span class="html-italic">vs.</span> salinity in (<b>a</b>) spring (the solid ellipse encloses measurements on 1 June 2010 associated with freshwater release at DPFD; the dashed ellipse encloses measurements at the marine stations on 6 April and 24 May 2011 associated with MR flood water), (<b>b</b>) summer, (<b>c</b>) fall, and (<b>d</b>) winter (solid ellipses correspond to measurements in Barataria Bay on 23 February 2010 associated with a high-pressure system (e.g., cold front)).</p> Full article ">Figure 5
<p>CDOM spectral slope <span class="html-italic">vs.</span> salinity relationship (<b>a</b>) spring, (<b>b</b>) summer, (<b>c</b>) fall, and (<b>d</b>) winter (the solid ellipse encloses similar measurements described in <a href="#remotesensing-07-12478-f004" class="html-fig">Figure 4</a>).</p> Full article ">Figure 6
<p><b>(a)</b> Decision tree for <span class="html-italic">a<sub>g</sub>355</span> at Station 1. The groups are represented by mean, number of samples (months) and percentage (%) of total measurements. <b>(b)</b> Part of the unpruned tree shows the influence of northerly winds at Node-5.</p> Full article ">Figure 7
<p>(<b>a</b>) Power law relationship between the CDOM absorption coefficient and the Landsat band ratio (Green/Red). The <span class="html-italic">a<sub>g</sub>355</span> band ratio algorithm: <span class="html-italic">a<sub>g</sub>355</span> = 6.68 × (B2/B3)<sup>−3.12</sup> (R<sup>2</sup> = 0.74, N = 50). (<b>b</b>) Validation of a band ratio algorithm (R<sup>2</sup> = 0.76, N = 28).</p> Full article ">Figure 8
<p>Time-series of salinity (Station 1: dots; Station 15: diamonds), the Mississippi river discharge at Baton Rouge (Q<sub>BR</sub>: diamonds) and Belle Chasse (Q<sub>BC</sub>: dots) (m<sup>3</sup>/s), discharge at Davis Pond Freshwater Diversion (Q<sub>DPFD</sub>) (m<sup>3</sup>∙s<sup>−1</sup>), wind speed (m/s) and wind direction (degrees from north) corresponding to monthly-observed <span class="html-italic">a<sub>g</sub>355</span> (m<sup>−1</sup>) in Barataria Bay in 2010 and 2011. The boxes (<b>a</b>–<b>d</b>) represent the meteorological and hydrological conditions corresponding to the CDOM maps illustrated in <a href="#remotesensing-07-12478-f009" class="html-fig">Figure 9</a>.</p> Full article ">Figure 9
<p>Landsat-5 TM-derived CDOM imagery of the Barataria Bay and Louisiana coastal waters obtained using Equation 4 during (<b>a</b>) a strong frontal system (25 February 2010), (<b>b</b>) a weak high-pressure system (8 November 2010), (<b>c</b>) elevated freshwater release at DPFD (12 February 2011) and (<b>d</b>) high MR flow condition (1 April 2011) (black arrows represent approximate daily mean wind direction, with longer arrows indicating stronger winds and <span class="html-italic">vice versa</span>).</p> Full article ">
7717 KiB  
Article
Mapping Impervious Surface Distribution with Integration of SNNP VIIRS-DNB and MODIS NDVI Data
by Wei Guo, Dengsheng Lu, Yanlan Wu and Jixian Zhang
Remote Sens. 2015, 7(9), 12459-12477; https://doi.org/10.3390/rs70912459 - 22 Sep 2015
Cited by 72 | Viewed by 8543
Abstract
Data from the U.S. Defense Meteorological Satellite Program’s Operational Line-scan System are often used to map impervious surface area (ISA) distribution at regional and global scales, but its coarse spatial resolution and data saturation produce high inaccuracy in ISA estimation. Suomi National Polar-orbiting [...] Read more.
Data from the U.S. Defense Meteorological Satellite Program’s Operational Line-scan System are often used to map impervious surface area (ISA) distribution at regional and global scales, but its coarse spatial resolution and data saturation produce high inaccuracy in ISA estimation. Suomi National Polar-orbiting Partnership (SNPP) Visible Infrared Imaging Radiometer Suite’s Day/Night Band (VIIRS-DNB) with its high spatial resolution and dynamic data range may provide new insights but has not been fully examined in mapping ISA distribution. In this paper, a new variable—Large-scale Impervious Surface Index (LISI)—is proposed to integrate VIIRS-DNB and Moderate Resolution Imaging Spectroradiometer (MODIS) normalized difference vegetation index (NDVI) data for mapping ISA distribution. A regression model was established, in which LISI was used as an independent variable and the reference ISA from Landsat images was a dependent variable. The results indicated a better estimation performance using LISI than using a single VIIRS-DNB or MODIS NDVI variable. The LISI-based approach provides accurate spatial patterns from high values in core urban areas to low values in rural areas, with an overall root mean squared error of 0.11. The LISI-based approach is recommended for fractional ISA estimation in a large area. Full article
►▼ Show Figures

Graphical abstract

Graphical abstract
Full article ">Figure 1
<p>The study area—all of China and six selected cities.</p> Full article ">Figure 2
<p>Framework of mapping ISA distribution using the Large-scale Impervious Surface Index (LISI).</p> Full article ">Figure 3
<p>Framework of extracting ISA data from Landsat 8 OLI imagery.</p> Full article ">Figure 4
<p>A comparison of four datasets in Beijing City; (<b>a</b>) VIIRS-DNB with 750 m spatial resolution; (<b>b</b>) 1-NDVI<sub>max</sub> with 250 m spatial resolution; (<b>c</b>) LISI with improved spatial resolution (250 m); and (<b>d</b>) Landsat 8 OLI color composite with 15 m spatial resolution.</p> Full article ">Figure 5
<p>The relationships between reference (<b>a</b>) ISA and DNB<sub>nor</sub>, (<b>b</b>) ISA and 1-NDVI<sub>max</sub>, and (<b>c</b>) ISA and LISI.</p> Full article ">Figure 6
<p>ISA distribution in China using the LISI-based model, highlighting the ISA spatial patterns from high ISA proportions in core urban areas to low proportions in rural regions; (<b>a</b>) Urumqi; (<b>b</b>) Beijing; and (<b>c</b>) Wuhan.</p> Full article ">Figure 7
<p>A comparison of the ISA distributions from three datasets for three cities, highlighting the better spatial patterns from the LISI-based model than the other two results; (<b>a1</b>–<b>a3</b>) represent ISA distribution using the DNB<sub>nor</sub> in Beijing, Wuhan, and Urumqi; <b>(b1</b>–<b>b3</b>) represent ISA distribution using 1-NDVI<sub>max</sub> in Beijing, Wuhan, and Urumqi; and (<b>c1</b>–<b>c3</b>) represent ISA distribution using LISI in Beijing, Wuhan, and Urumqi.</p> Full article ">Figure 8
<p>The relationships between ISA estimates and corresponding reference data among three datasets; (<b>a</b>) ISA estimates from DNB<sub>nor</sub>; (<b>b</b>) ISA estimates from 1-NDVI<sub>max</sub>; and (<b>c</b>) ISA estimates from LISI.</p> Full article ">Figure 9
<p>The relationship between RMSE and economic conditions, showing the effects of different economic conditions on ISA estimation performance through the comparison of different variables.</p> Full article ">
11418 KiB  
Article
Mapping Two-Dimensional Deformation Field Time-Series of Large Slope by Coupling DInSAR-SBAS with MAI-SBAS
by Liming He, Lixin Wu, Shanjun Liu, Zhi Wang, Chang Su and Sheng-Nan Liu
Remote Sens. 2015, 7(9), 12440-12458; https://doi.org/10.3390/rs70912440 - 22 Sep 2015
Cited by 45 | Viewed by 7624
Abstract
Mapping deformation field time-series, including vertical and horizontal motions, is vital for landslide monitoring and slope safety assessment. However, the conventional differential synthetic aperture radar interferometry (DInSAR) technique can only detect the displacement component in the satellite-to-ground direction, i.e., line-of-sight (LOS) direction [...] Read more.
Mapping deformation field time-series, including vertical and horizontal motions, is vital for landslide monitoring and slope safety assessment. However, the conventional differential synthetic aperture radar interferometry (DInSAR) technique can only detect the displacement component in the satellite-to-ground direction, i.e., line-of-sight (LOS) direction displacement. To overcome this constraint, a new method was developed to obtain the displacement field time series of a slope by coupling DInSAR based small baseline subset approach (DInSAR-SBAS) with multiple-aperture InSAR (MAI) based small baseline subset approach (MAI-SBAS). This novel method has been applied to a set of 11 observations from the phased array type L-band synthetic aperture radar (PALSAR) sensor onboard the advanced land observing satellite (ALOS), spanning from 2007 to 2011, of two large-scale north–south slopes of the largest Asian open-pit mine in the Northeast of China. The retrieved displacement time series showed that the proposed method can detect and measure the large displacements that occurred along the north–south direction, and the gradually changing two-dimensional displacement fields. Moreover, we verified this new method by comparing the displacement results to global positioning system (GPS) measurements. Full article
►▼ Show Figures

Graphical abstract

Graphical abstract
Full article ">Figure 1
<p>(<b>a</b>) Location of Fushun west open-pit mine (FWOPM). The blue box marked the coverage of the frame provided by the phased array type L-band synthetic aperture radar (PALSAR) sensor onboard the advanced land observing satellite (ALOS) in this study. The background is a shaded image of Shuttle Radar Topography Mission (SRTM) digital elevation model (DEM). The open-pit mine is specified by a red rectangle. The left-upper inset is a sketch map to mark the location of the study area in China. (<b>b</b>) The three-dimensional (3-D) topography presents the detailed morphology of the open-pit mine, which was obtained by 3-D laser scanning technology in 2013.</p> Full article ">Figure 2
<p>The social economic environment around FWOPM. The green lines indicate the boundaries of residential zones, industrial zone, and factories. The base image was captured by China’s Gaofen-1 satellite in 2014.</p> Full article ">Figure 3
<p>Flowchart of the differential synthetic aperture radar interferometry based small baseline subset (DInSAR-SBAS) algorithm.</p> Full article ">Figure 4
<p>Flowchart of the multiple-aperture synthetic aperture radar interferometry based small baseline subset (MAI-SBAS) algorithm.</p> Full article ">Figure 5
<p>Monitoring vectors for line-of-sight (LOS) differential synthetic aperture radar interferometry (DInSAR) and azimuthal multiple-aperture synthetic aperture radar interferometry (MAI) in the ascending mode.</p> Full article ">Figure 6
<p>The distribution of ALOS PALSAR image pairs in the spatial and temporal baseline plane of two small baseline subsets.</p> Full article ">Figure 7
<p>Vertical displacement maps in FWOPM obtained by DInSAR-SBAS approach. The color represents the vertical displacements for each time period. CP01 and CP02 represent two locations of global positioning system (GPS) measurements. The background is a shaded relief map of SRTM DEM.</p> Full article ">Figure 8
<p>North–south displacement maps in FWOPM obtained by MAI-SBAS approach. The red color represents the southward displacements for each time period, while the blue color represents the northward displacements for each time period. CP01 and CP02 represent two locations of GPS measurements. The background is a shaded relief map of SRTM DEM.</p> Full article ">Figure 9
<p>Integrated displacement velocity field of FWOPM landslide for the period of January 2007–January 2011. The background is a shaded relief map of SRTM DEM. The color represents the vertical displacement rate for the studied period. The red arrows represent the horizontal displacement velocity with different directions in the north and south parts of FWOPM.</p> Full article ">Figure 10
<p>(<b>a</b>) Displacement time series along vertical and north–south directions obtained by GPS and synthetic aperture radar interferometry (InSAR) of CP01. (<b>b</b>) Displacement time series along vertical and north–south directions obtained by GPS and InSAR of CP02.</p> Full article ">Figure 11
<p>The yellow dashed line marks the boundary of the south landslide in FWOPM [<a href="#B22-remotesensing-07-12440" class="html-bibr">22</a>]. (<b>a</b>) The color represents the integrated displacement rate for the studied period. The background is a shaded relief map of SRTM DEM; (<b>b</b>) The base image was captured by China’s Gaofen-1 satellite in 2014, several ground fissures can be seen clearly in the south part of FWOPM; (<b>c</b>) The base image was an aerial picture of the study area from [<a href="#B22-remotesensing-07-12440" class="html-bibr">22</a>].</p> Full article ">
2192 KiB  
Article
Mapping Urban Areas with Integration of DMSP/OLS Nighttime Light and MODIS Data Using Machine Learning Techniques
by Wenlong Jing, Yaping Yang, Xiafang Yue and Xiaodan Zhao
Remote Sens. 2015, 7(9), 12419-12439; https://doi.org/10.3390/rs70912419 - 22 Sep 2015
Cited by 58 | Viewed by 7384
Abstract
Mapping urban areas at global and regional scales is an urgent and crucial task for detecting urbanization and human activities throughout the world and is useful for discerning the influence of urban expansion upon the ecosystem and the surrounding environment. DMSP-OLS stable nighttime [...] Read more.
Mapping urban areas at global and regional scales is an urgent and crucial task for detecting urbanization and human activities throughout the world and is useful for discerning the influence of urban expansion upon the ecosystem and the surrounding environment. DMSP-OLS stable nighttime lights have provided an effective way to monitor human activities on a global scale. Threshold-based algorithms have been widely used for extracting urban areas and estimating urban expansion, but the accuracy can decrease because of the empirical and subjective selection of threshold values. This paper proposes an approach for extracting urban areas with the integration of DMSP-OLS stable nighttime lights and MODIS data utilizing training sample datasets selected from DMSP-OLS and MODIS NDVI based on several simple strategies. Four classification algorithms were implemented for comparison: the classification and regression tree (CART), k-nearest-neighbors (k-NN), support vector machine (SVM), and random forests (RF). A case study was carried out on the eastern part of China, covering 99 cities and 1,027,700 km2. The classification results were validated using an independent land cover dataset, and then compared with an existing contextual classification method. The results showed that the new method can achieve results with comparable accuracies, and is easier to implement and less sensitive to the initial thresholds than the contextual method. Among the four classifiers implemented, RF achieved the most stable results and the highest average Kappa. Meanwhile CART produced highly overestimated results compared to the other three classifiers. Although k-NN and SVM tended to produce similar accuracy, less-bright areas around the urban cores seemed to be ignored when using SVM, which led to the underestimation of urban areas. Furthermore, quantity assessment showed that the results produced by k-NN, SVM, and RFs exhibited better agreement in larger cities and low consistency in small cities. Full article
►▼ Show Figures

Graphical abstract

Graphical abstract
Full article ">Figure 1
<p>Case study area of eastern China includes seven provinces and three municipalities (Beijing, Tianjin, and Shanghai), which are of provincial-level.</p> Full article ">Figure 2
<p>Flowchart of urban area extraction method.</p> Full article ">Figure 3
<p>Urban area extraction results: (<b>a</b>) Beijing, (<b>b</b>) Tianjin, (<b>c</b>) Qingdao, (<b>d</b>) Shenyang, (<b>e</b>) Cangzhou, (<b>f</b>) Guangzhou, and (<b>g)</b> Shanghai.</p> Full article ">Figure 4
<p>Scatter plot of urban pixel count of each city between predicted results and FROM-GLC for four classifiers: (<b>a</b>) CART, (<b>b</b>) k-NN, (<b>c</b>) SVM, and (<b>d</b>) RF.</p> Full article ">Figure 5
<p>Spatial distribution of Kappa at the city level: (<b>a</b>) CART, (<b>b</b>) k-NN, (<b>c</b>) SVM, and (<b>d</b>) RF.</p> Full article ">Figure 6
<p>(<b>a</b>) DMSP-OLS NTL image of study area. (<b>b</b>) Mean NTL DN of 99 cities.</p> Full article ">Figure 7
<p>Relationship between the Kappa variance and mean NTL DN of pixels with values greater than 30, for 99 cities.</p> Full article ">Figure 8
<p>Kappa coefficient for the changed initial threshold of the NTL DN value for the urban extraction results.</p> Full article ">Figure 9
<p>Kappa coefficient for the changed initial threshold of the NDVI value for the urban extraction results.</p> Full article ">Figure 10
<p>Average kappa coefficient with different sample percentages.</p> Full article ">
820 KiB  
Article
Assimilation of Two Variables Derived from Hyperspectral Data into the DSSAT-CERES Model for Grain Yield and Quality Estimation
by Zhenhai Li, Jihua Wang, Xingang Xu, Chunjiang Zhao, Xiuliang Jin, Guijun Yang and Haikuan Feng
Remote Sens. 2015, 7(9), 12400-12418; https://doi.org/10.3390/rs70912400 - 22 Sep 2015
Cited by 61 | Viewed by 9380
Abstract
The combination of remote sensing and crop growth models has become an effective tool for yield estimation and a potential method for grain quality estimation. In this study, two assimilation variables (derived from a hyperspectral sensor), called leaf area index (LAI) and canopy [...] Read more.
The combination of remote sensing and crop growth models has become an effective tool for yield estimation and a potential method for grain quality estimation. In this study, two assimilation variables (derived from a hyperspectral sensor), called leaf area index (LAI) and canopy nitrogen accumulation (CNA), were jointly used to calibrate the sensitive parameters and initial states of the DSSAT-CERES crop model, to improve simulated output of the grain yield and protein content of winter wheat. The results show that the modified simple ratio (MSR) and normalized difference red edge (NDRE) better estimated LAI and CNA, respectively, compared with the other possible vegetation indices. The integration of both LAI and CNA resulted in a more robust DSSAT-CERES models with than each one alone. The R2 and RMSE values, respectively, of the regression between the simulated (using the two assimilation variables method) and measured LAI were 0.828 and 0.494, and for CNA were 0.808 and 20.26 kg N∙ha−1. These two assimilation variables resulted in grain yield and protein content estimates of winter wheat with a high precision and R2 and RMSE values of 0.698 and 0.726 ton∙ha−1, and 0.758% and 1.16%, respectively. This study provides a more robust method for estimating the grain yield and protein content of winter wheat based on the integration of the DSSAT-CERES crop model and remote sensing data. Full article
(This article belongs to the Special Issue Remote Sensing in Precision Agriculture)
►▼ Show Figures

Graphical abstract

Graphical abstract
Full article ">Figure 1
<p>Flowchart of the particle swarm optimization (PSO) assimilation scheme for integrating remote sensing data with the Decision Support System for Agrotechnology Transfer–Crop Environment Resource Synthesis (DSSAT-CERES) model.</p> Full article ">Figure 2
<p>Relationships between simulated and measured values of (<b>a</b>) leaf area index (LAI) with LAI with a state variable (SV<sub>LAI</sub>), (<b>b</b>) canopy N accumulation (CNA) with SV<sub>LAI</sub>, (<b>c</b>) LAI with CNA with a state variable (SV<sub>CNA</sub>), (<b>d</b>) CNA with SV<sub>CNA</sub>, (<b>e</b>) LAI with SV<sub>LAI +</sub> <sub>CNA</sub>, and (<b>f</b>) CNA with SV<sub>LAI + CNA</sub>.</p> Full article ">Figure 3
<p>Relationships between measured and simulated values of (<b>a</b>) yield with leaf area index as a state variable (SV<sub>LAI</sub>), (<b>b</b>) grain protein content (GPC) with SV<sub>LAI</sub>, (<b>c</b>) yield with canopy N accumulation as a state variable (SV<sub>CNA</sub>), (<b>d</b>) GPC with SV<sub>CNA</sub>, (<b>e</b>) yield with SV<sub>LAI + CNA</sub>, and (<b>f</b>) GPC with SV<sub>LAI + CNA.</sub></p> Full article ">
1727 KiB  
Article
Remotely Sensed Nightlights to Map Societal Exposure to Hydrometeorological Hazards
by Agnes Jane Soto Gómez, Giuliano Di Baldassarre, Allan Rodhe and Veijo A. Pohjola
Remote Sens. 2015, 7(9), 12380-12399; https://doi.org/10.3390/rs70912380 - 22 Sep 2015
Cited by 6 | Viewed by 5938
Abstract
This study used remotely sensed maps of nightlights to investigate the etiology of increasing disaster losses from hydrometeorological hazards in a data-scarce area. We explored trends in the probability of occurrence of hazardous events (extreme rainfall) and exposure of the local population as [...] Read more.
This study used remotely sensed maps of nightlights to investigate the etiology of increasing disaster losses from hydrometeorological hazards in a data-scarce area. We explored trends in the probability of occurrence of hazardous events (extreme rainfall) and exposure of the local population as components of risk. The temporal variation of the spatial distribution of exposure to hydrometeorological hazards was studied using nightlight satellite imagery as a proxy. Temporal (yearly) and spatial (1 km) resolution make them more useful than official census data. Additionally, satellite nightlights can track informal (unofficial) human settlements. The study focused on the Samala River catchment in Guatemala. The analyses of disasters, using DesInventar Disaster Information Management System data, showed that fatalities caused by hydrometeorological events have increased. Such an increase in disaster losses can be explained by trends in both: (i) catchment conditions that tend to lead to more frequent hydrometeorological extremes (more frequent occurrence of days with wet conditions); and (ii) increasing human exposure to hazardous events (as observed by amount and intensity of nightlights in areas close to rivers). Our study shows the value of remote sensing data and provides a framework to explore the dynamics of disaster risk when ground data are spatially and temporally limited. Full article
(This article belongs to the Special Issue Remote Sensing in Flood Monitoring and Management)
►▼ Show Figures

Graphical abstract

Graphical abstract
Full article ">Figure 1
<p>Samala River catchment location, topography and administrative divisions. The physical boundary of the catchment (black line) does not coincide with municipal boundaries (grey dotted lines).</p> Full article ">Figure 2
<p>Temporal evolution of nighttime lights in the Samala River catchment. Images show the intercalibrated values of nightlights in different years to expose how the brightness has increased over the years. Images for 1994, 1998, 2002, and 2004 show the averaged intercalibrated values of the two available satellites (the boundary of the catchment corresponds to <a href="#remotesensing-07-12380-f001" class="html-fig">Figure 1</a>). Original image and data processing by NOAA’s National Geophysical Data Center. DMSP data collected by US Air Force Weather Agency. Intercalibrated images: the authors.</p> Full article ">Figure 3
<p>Relative impact of disasters over time. The relative impact of disasters (DI<sub>Rel</sub>) is measured by the number of fatalities caused by all the disasters (classified by hydrometeorological and non hydrometeorological causes) during each period, divided by the number of years in the period and the yearly average population in that period (<b>a</b>) in the Samala River catchment, (<b>b</b>) in the highlands and (<b>c</b>) in the lowlands of the catchment. The fatalities resulting from non hydrometeorological causes for disasters correspond to earthquakes and have been included in the graph to show the importance of the hydrometeorological causes in the occurrences of disasters in the case study.</p> Full article ">Figure 3 Cont.
<p>Relative impact of disasters over time. The relative impact of disasters (DI<sub>Rel</sub>) is measured by the number of fatalities caused by all the disasters (classified by hydrometeorological and non hydrometeorological causes) during each period, divided by the number of years in the period and the yearly average population in that period (<b>a</b>) in the Samala River catchment, (<b>b</b>) in the highlands and (<b>c</b>) in the lowlands of the catchment. The fatalities resulting from non hydrometeorological causes for disasters correspond to earthquakes and have been included in the graph to show the importance of the hydrometeorological causes in the occurrences of disasters in the case study.</p> Full article ">Figure 4
<p>Precipitation trends 1978–2011, Retalhuleu station, Guatemala. Three approaches identifying trends in extreme precipitation: (<b>a</b>) yearly maxima of daily precipitation, (<b>b</b>) yearly maxima of 10-day period wetness index according to Soto <span class="html-italic">et al.</span> [<a href="#B32-remotesensing-07-12380" class="html-bibr">32</a>], (<b>c</b>) and yearly frequency of days when the 10-day period wetness index was equal to or higher than the aforementioned threshold. Data source: INSIVUMEH.</p> Full article ">Figure 4 Cont.
<p>Precipitation trends 1978–2011, Retalhuleu station, Guatemala. Three approaches identifying trends in extreme precipitation: (<b>a</b>) yearly maxima of daily precipitation, (<b>b</b>) yearly maxima of 10-day period wetness index according to Soto <span class="html-italic">et al.</span> [<a href="#B32-remotesensing-07-12380" class="html-bibr">32</a>], (<b>c</b>) and yearly frequency of days when the 10-day period wetness index was equal to or higher than the aforementioned threshold. Data source: INSIVUMEH.</p> Full article ">Figure 5
<p>Nightlights in the close proximity of the Samala River over time. The evolution in time of nightlight cells corresponding with the river path (close proximity) is shown in terms of (<b>a</b>) distribution of cells per brightness ranges, (<b>b</b>) ratio of lit cells (cells that have recorded some light) to total cells, and (<b>c</b>) characteristic average digital number (CB<sub>avg</sub>).</p> Full article ">Figure 6
<p>Nightlights in the highlands and lowlands of the Samala River catchment over time. The evolution in time of nightlight cells is shown separately for the highlands and the lowlands in terms of (<b>a1</b>,<b>b1</b>) distribution of cells per brightness ranges and (<b>a2</b>,<b>b2</b>) characteristic average digital number (CBavg).</p> Full article ">Figure 7
<p>Changes in the Samala River catchment regarding (<b>a</b>) relative disaster impact DI<sub>rel</sub> associated with hydrometeorological causes (a proxy for disaster risk), (<b>b</b>) average number of days in the year of I<sub>wet</sub> over the identified threshold ([<a href="#B32-remotesensing-07-12380" class="html-bibr">32</a>] (a proxy for natural hazards), and (<b>c</b>) average brightness (a proxy for vulnerability in terms of exposure). Average brightness (c) corresponds to the values of the final year of each period (1994, 2002, and 2011) and is shown separately for the close proximities of the river and the rest of the catchment (non river area).</p> Full article ">
1967 KiB  
Article
Assessment of an Operational System for Crop Type Map Production Using High Temporal and Spatial Resolution Satellite Optical Imagery
by Jordi Inglada, Marcela Arias, Benjamin Tardy, Olivier Hagolle, Silvia Valero, David Morin, Gérard Dedieu, Guadalupe Sepulcre, Sophie Bontemps, Pierre Defourny and Benjamin Koetz
Remote Sens. 2015, 7(9), 12356-12379; https://doi.org/10.3390/rs70912356 - 22 Sep 2015
Cited by 286 | Viewed by 15555
Abstract
Crop area extent estimates and crop type maps provide crucial information for agricultural monitoring and management. Remote sensing imagery in general and, more specifically, high temporal and high spatial resolution data as the ones which will be available with upcoming systems, such as [...] Read more.
Crop area extent estimates and crop type maps provide crucial information for agricultural monitoring and management. Remote sensing imagery in general and, more specifically, high temporal and high spatial resolution data as the ones which will be available with upcoming systems, such as Sentinel-2, constitute a major asset for this kind of application. The goal of this paper is to assess to what extent state-of-the-art supervised classification methods can be applied to high resolution multi-temporal optical imagery to produce accurate crop type maps at the global scale. Five concurrent strategies for automatic crop type map production have been selected and benchmarked using SPOT4 (Take5) and Landsat 8 data over 12 test sites spread all over the globe (four in Europe, four in Africa, two in America and two in Asia). This variety of tests sites allows one to draw conclusions applicable to a wide variety of landscapes and crop systems. The results show that a random forest classifier operating on linearly temporally gap-filled images can achieve overall accuracies above 80% for most sites. Only two sites showed low performances: Madagascar due to the presence of fields smaller than the pixel size and Burkina Faso due to a mix of trees and crops in the fields. The approach is based on supervised machine learning techniques, which need in situ data collection for the training step, but the map production is fully automatic. Full article
(This article belongs to the Special Issue Lessons Learned from the SPOT4 (Take5): Experiment in Preparation for Sentinel-2)
►▼ Show Figures

Graphical abstract

Graphical abstract
Full article ">Figure 1
<p>Choices of algorithms leading to strategy comparisons.</p> Full article ">Figure 2
<p>Example of Sentinel-2 tracks: in red, the acquisitions of Day D; in yellow, those of Day D + 7 (with one satellite) or Day D + 2 (with 2 satellites). Background image <math display="inline"> <msup> <mrow/> <mi>©</mi> </msup> </math>2015 Google Imagery.</p> Full article ">Figure 3
<p>Block diagram of the crop type map production.</p> Full article ">Figure 4
<p>F-score and OA results for RF and SVM for the France test site. (<b>a</b>) RF; (<b>b</b>) SVM.</p> Full article ">Figure 5
<p>Comparison of the crop map type obtained using the RF classifier and field surveys with the RPG database in a region 40 km north to the field data collection site. Numbers indicate the crop type in the reference data and colors correspond to the output of the classifier.</p> Full article ">Figure 6
<p>Comparison of the crop map type obtained using the RF classifier and field surveys with the Registre Parcellaire Graphique (RPG) database on a high altitude region far from the field data collection site. Numbers indicate the crop type in the reference data, and colors correspond to the output of the classifier.</p> Full article ">Figure 7
<p>F-score and OA results for RF and SVM for the Ukraine test site. (<b>a</b>) RF; (<b>b</b>) SVM.</p> Full article ">Figure 8
<p>F-score and OA results for RF and SVM for the Burkina Faso test site. (<b>a</b>) RF; (<b>b</b>) SVM.</p> Full article ">Figure 9
<p>Intra-plot variability in the Burkina Faso site. <span class="html-italic">In situ</span> data plots are overlayed in red. The fields are highly heterogeneous and difficult to classify as a single crop. <math display="inline"> <msup> <mrow/> <mi>©</mi> </msup> </math>2015 Google Imagery, <math display="inline"> <msup> <mrow/> <mi>©</mi> </msup> </math>2015 DigitalGlobe.</p> Full article ">
2058 KiB  
Article
High-Resolution Mapping of Urban Surface Water Using ZY-3 Multi-Spectral Imagery
by Fangfang Yao, Chao Wang, Di Dong, Jiancheng Luo, Zhanfeng Shen and Kehan Yang
Remote Sens. 2015, 7(9), 12336-12355; https://doi.org/10.3390/rs70912336 - 22 Sep 2015
Cited by 89 | Viewed by 8598
Abstract
Accurate information of urban surface water is important for assessing the role it plays in urban ecosystem services under the content of urbanization and climate change. However, high-resolution monitoring of urban water bodies using remote sensing remains a challenge because of the limitation [...] Read more.
Accurate information of urban surface water is important for assessing the role it plays in urban ecosystem services under the content of urbanization and climate change. However, high-resolution monitoring of urban water bodies using remote sensing remains a challenge because of the limitation of previous water indices and the dark building shadow effect. To address this problem, we proposed an automated urban water extraction method (UWEM) which combines a new water index, together with a building shadow detection method. Firstly, we trained the parameters of UWEM using ZY-3 imagery of Qingdao, China. Then we verified the algorithm using five other sub-scenes (Aksu, Fuzhou, Hanyang, Huangpo and Huainan) ZY-3 imagery. The performance was compared with that of the Normalized Difference Water Index (NDWI). Results indicated that UWEM performed significantly better at the sub-scenes with kappa coefficients improved by 7.87%, 32.35%, 12.64%, 29.72%, 14.29%, respectively, and total omission and commission error reduced by 61.53%, 65.74%, 83.51%, 82.44%, and 74.40%, respectively. Furthermore, UWEM has more stable performances than NDWI’s in a range of thresholds near zero. It reduces the over- and under-estimation issues which often accompany previous water indices when mapping urban surface water under complex environmental conditions. Full article
(This article belongs to the Special Issue Remote Sensing of Water Resources)
►▼ Show Figures

Graphical abstract

Graphical abstract
Full article ">Figure 1
<p>Reflectance and water index value distributions of major land cover types from pure pixels. Each box plot shows the location of the 10th, 25th, 50th, 75th, and 90th percentiles with horizontal lines (boxes and whiskers) and the circles are 5th and 95th percentiles (The dashed boxes show the spectral contrast between building and dark shadow).(<b>a–d</b>): the reflectance distributions of major land cover types in blue, green, red, NIR bands, respectively; <b>e</b>: the HRWI value distributions of major land cover types; <b>f</b>: the NDWI value distributions of major land cover types.</p> Full article ">Figure 2
<p>Sketch maps of the building and its building shadow (Edge stands for the edge between sun side and shade side, inside edge pixels stand for the shaded pixels which are adjacent to the Edge, outside edge pixels stand for the building pixels which are adjacent to the Edge).</p> Full article ">Figure 3
<p>Flowchart of automated dark building shadow detection method.</p> Full article ">Figure 4
<p>Flowchart of urban water extraction method.</p> Full article ">Figure 5
<p>Comparison of water extraction results using NDWI and UWEM at the five test sites (NVP stands for the number of verification pixels).</p> Full article ">Figure 6
<p>The accuracy of UWEM and NDWI at each optimal threshold.</p> Full article ">Figure 7
<p>The accuracy of the UWEM and NDWI at five test sites in a range of thresholds near zero.</p> Full article ">Figure 8
<p>The extracted shadow mask at the test site in Huainan using the method proposed in <a href="#sec3dot3dot4-remotesensing-07-12336" class="html-sec">Section 3.3.4</a>.<b>a</b>: the ZY-3 color composite (NIR, Red, and Green); <b>b1</b>: enlarged view of the highlighted region b in <b>a</b>; <b>b2</b>: the extracted shadow results of b1; <b>c1</b>: enlarged view of the highlighted region c in <b>a</b>; <b>c2</b>: the extracted shadow results of c1; <b>d1</b>: enlarged view of the highlighted region d in <b>a</b>; <b>d2</b>: the extracted shadow results of d1.</p> Full article ">
1905 KiB  
Article
A Phenology-Based Method for Monitoring Woody and Herbaceous Vegetation in Mediterranean Forests from NDVI Time Series
by David Helman, Itamar M. Lensky, Naama Tessler and Yagil Osem
Remote Sens. 2015, 7(9), 12314-12335; https://doi.org/10.3390/rs70912314 - 21 Sep 2015
Cited by 66 | Viewed by 11039
Abstract
We present an efficient method for monitoring woody (i.e., evergreen) and herbaceous (i.e., ephemeral) vegetation in Mediterranean forests at a sub pixel scale from Normalized Difference Vegetation Index (NDVI) time series derived from the Moderate Resolution Imaging Spectroradiometer (MODIS). The method is based [...] Read more.
We present an efficient method for monitoring woody (i.e., evergreen) and herbaceous (i.e., ephemeral) vegetation in Mediterranean forests at a sub pixel scale from Normalized Difference Vegetation Index (NDVI) time series derived from the Moderate Resolution Imaging Spectroradiometer (MODIS). The method is based on the distinct development periods of those vegetation components. In the dry season, herbaceous vegetation is absent or completely dry in Mediterranean forests. Thus the mean NDVI in the dry season was attributed to the woody vegetation (NDVIW). A constant NDVI value was assumed for soil background during this period. In the wet season, changes in NDVI were attributed to the development of ephemeral herbaceous vegetation in the forest floor and its maximum value to the peak green cover (NDVIH). NDVIW and NDVIH agreed well with field estimates of leaf area index and fraction of vegetation cover in two differently structured Mediterranean forests. To further assess the method’s assumptions, understory NDVI was retrieved form MODIS Bidirectional Reflectance Distribution Function (BRDF) data and compared with NDVIH. After calibration, leaf area index and woody and herbaceous vegetation covers were assessed for those forests. Applicability for pre- and post-fire monitoring is presented as a potential use of this method for forest management in Mediterranean-climate regions. Full article
►▼ Show Figures

Graphical abstract

Graphical abstract
Full article ">Figure 1
<p>(<b>a</b>) Location of the two study sites (Yatir forest and Mt. Carmel woodlands). (<b>b</b>) Aerial (Google Earth <sup>®</sup>) and (<b>c</b>) field views of the planted pine forest of Yatir in the semiarid region of Israel. Photo: Eugene Ivanov.</p> Full article ">Figure 2
<p>(<b>a</b>) Aerial (Google Earth <sup>®</sup>) and (<b>b</b>) field views of the mixed pine-oak evergreen woodlands at Mt. Carmel. The burnt area from the wildfire of 2010 (red line) and the location of the 22 survey plots (dots) are indicated in (<b>a</b>). Photo: Naama Tessler.</p> Full article ">Figure 3
<p>Schematic representation of the distinct growth and senescence periods of evergreen woody vegetation (dashed blue line) and ephemeral herbaceous plants (red line) in Mediterranean forests. Phenological stages are shown as the relative Normalized Difference Vegetation Index (<span class="html-italic">i.e.</span>, NDVI/NDVI<sub>max</sub>) of each of those two vegetation components. The wet and dry periods are also indicated.</p> Full article ">Figure 4
<p>Examples of the original and decomposed NDVI time series in one representative pixel (250 m) at the evergreen pine forest of Yatir (left column) and the pine-oak woodlands of Mt. Carmel (right column). Original and smoothed time series of NDVI<sub>Ecos</sub> are shown in (<b>a</b>,<b>b</b>), NDVI<sub>W</sub> in (<b>c</b>,<b>d</b>), NDVI<sub>Seas</sub> in (<b>e</b>,<b>f</b>) and NDVI<sub>H</sub> in (<b>g</b>,<b>h</b>) for Yatir and Mt. Carmel, respectively.</p> Full article ">Figure 5
<p>(<b>a</b>) NDVI<sub>Ecos</sub>, NDVIu (retrieved from Bidirectional Reflectance Distribution Function product), and monthly <span class="html-italic">in situ</span> overstory LAI (LAIo) in a 4 km<sup>2</sup> area at the Yatir pine forest. Scatterplots showing the relationships between (<b>b</b>) monthly LAIo <span class="html-italic">vs.</span> NDVI<sub>Ecos</sub> and (<b>c</b>) mean annual LAIo <span class="html-italic">vs.</span> NDVI<sub>W</sub> (2000–2006) for the same area.</p> Full article ">Figure 6
<p>Spatial distribution of (<b>a</b>) the mean annual and (<b>b</b>) trends of overstory LAI retrieved from NDVI<sub>W</sub> at Yatir for 2000–2014. Significant trends are indicated in (<b>a</b>) as + (positive) and • (negative), while in (<b>b</b>) all significant trends are indicated as +.</p> Full article ">Figure 7
<p>NDVI<sub>Ecos</sub> and NDVIu (from BRDF) in a 4-km<sup>2</sup> area of the pine-oak woodlands at Mt. Carmel.</p> Full article ">Figure 8
<p>Scatterplots of woody (<b>left</b>) and herbaceous (<b>right</b>) vegetation covers (%) assessed in field <span class="html-italic">vs.</span> NDVI<sub>W</sub> and NDVI<sub>H</sub> in 14-MODIS pixels. Each dot in the graph is the four-year averaged vegetation cover within one MODIS pixel (see <a href="#sec2dot2dot2-remotesensing-07-12314" class="html-sec">Section 2.2.2</a>). For the herbaceous vegetation cover, an exponential function was fitted following the assumption that NDVI<sub>H</sub> equals 0 in the dry season when herbaceous vegetation is absent (0% cover).</p> Full article ">Figure 9
<p>The 14-year mean woody (<b>a</b>–<b>c</b>) and herbaceous (<b>d</b>–<b>f</b>) vegetation cover (%) and NDVI trends in Mt. Carmel. Vegetation cover was estimated from NDVI<sub>W</sub> and NDVI<sub>H</sub> using (<b>a</b>,<b>d</b>) NDVI-field regression functions (see <a href="#remotesensing-07-12314-f008" class="html-fig">Figure 8</a>) and (<b>b</b>,<b>e</b>) the two-end members FVC equation (Equation (2)). The strong NDVI<sub>W</sub> decline in (<b>c</b>) and contrast NDVI<sub>H</sub> increase in (<b>f</b>) are the result of the 2010 wildfire (compare with wildfire area in <a href="#remotesensing-07-12314-f002" class="html-fig">Figure 2</a>a).</p> Full article ">Figure 10
<p>(<b>a</b>) A fuel-based fire risk map produced for the year 2009 (prior to the 2010 wildfire) from the woody vegetation cover (mean NDVI<sub>W</sub>) and its dryness status (NDVI<sub>W</sub> trends). Superimposed is the area of the wildfire; histograms of (<b>b</b>) the total number of pixels in Mt. Carmel with their respective risk levels and (<b>c</b>) the ratio between the number of pixels with a specific risk level in the burnt zone to that in the entire Mt. Carmel area (in %). The dashed line in (<b>c</b>) indicates the percent area of burnt zone from the entire Mt. Carmel area (<span class="html-italic">i.e.</span>, 8.6%).</p> Full article ">Figure 11
<p>Maps of (<b>a</b>) low, medium and high severity burnt areas classified in field and extended with high-resolution aerial photograph; and (<b>b</b>) the difference between post- and pre-fire NDVI<sub>W</sub> (ΔNDVI<sub>W</sub>). (<b>c</b>) Box plot of mean, first and third quartiles (with respective standard deviations) of ΔNDVI<sub>W</sub> in the low, medium and high severity areas mapped at field (shown in a). Different letters indicate statistically significant differences at <span class="html-italic">p</span> &lt; 0.001 using a two-tailed Student’s t-test, after a Bonferroni correction.</p> Full article ">Figure 12
<p>Changes in woody and herbaceous vegetation cover following the wildfire of 2010 as estimated from field (open bars) and from NDVI (solid bars) using a field-based calibration function (NDVI-field) and two-end members fraction of vegetation cover (FVC) equation. Asterisks indicate statistically significant differences between the NDVI’s components and the field estimates for each specific year at <span class="html-italic">p</span> &lt; 0.05. Different letters denote statistically significant differences in vegetation covers between years at <span class="html-italic">p</span> &lt; 0.05.</p> Full article ">
1476 KiB  
Article
World’s Largest Macroalgal Blooms Altered Phytoplankton Biomass in Summer in the Yellow Sea: Satellite Observations
by Qianguo Xing, Chuanmin Hu, Danling Tang, Liqiao Tian, Shilin Tang, Xiao Hua Wang, Mingjing Lou and Xuelu Gao
Remote Sens. 2015, 7(9), 12297-12313; https://doi.org/10.3390/rs70912297 - 21 Sep 2015
Cited by 63 | Viewed by 7639
Abstract
Since 2008, the world’s largest blooms of the green macroalgae, Ulva prolifera, have occurred every summer in the Yellow Sea, posing the question of whether these macroalgal blooms (MABs) have changed the phytoplankton biomass due to their perturbations of nutrient dynamics. We [...] Read more.
Since 2008, the world’s largest blooms of the green macroalgae, Ulva prolifera, have occurred every summer in the Yellow Sea, posing the question of whether these macroalgal blooms (MABs) have changed the phytoplankton biomass due to their perturbations of nutrient dynamics. We have attempted to address this question using long-term Moderate Resolution Imaging Spectroradiometer (MODIS) observations. A new MODIS monthly time-series of chlorophyll-a concentrations (Chl-a, an index of phytoplankton biomass) was generated after removing the macroalgae-contaminated pixels that were characterized by unexpectedly high values in the daily Chl-a products. Compared with Chl-a during July of 2002–2006 (the pre-MAB period), Chl-a during July of 2008–2012 (the MAB period) exhibited significant increases in the offshore Yellow Sea waters (rich in macroalgae), with mean Chl-a increased by 98% from 0.64 µg/L to 1.26 µg/L in the study region. In contrast, no significant Chl-a changes were observed during June between the two periods. After analyzing sea surface temperature, photosynthetically available radiation, and nutrient availability, we speculate that the observed Chl-a changes are due to nutrient competition between macroalgae and phytoplankton: during the MAB period, the fast-growing macroalgae would uptake the increased nutrients from the origin of Jiangsu Shoal; thus, the nutrients available to phytoplankton were reduced, leading to no apparent increases in biomass in the offshore Yellow Sea in June. Full article
(This article belongs to the Special Issue Remote Sensing of Biogeochemical Cycles)
►▼ Show Figures

Graphical abstract

Graphical abstract
Full article ">Figure 1
<p>Study region showing the area impacted by MABs (green slicks in and near the purple dashed circle) during summer 2008 in the Yellow Sea. The arrow shows the main drifting pathway of macroalgae from the origin of Jiangsu Shoal, the most nutrient-polluted region in the Yellow Sea. The green slicks of macroalgae were extracted from Moderate Resolution Imaging Spectroradiometer (MODIS) measurements on day 136, 151, 170, and 180 in 2008 using the data of MODIS Normalized Difference Vegetation Index (NDVI) [<a href="#B6-remotesensing-07-12297" class="html-bibr">6</a>]. The green dots show the macroalgae origin sites validated from a cruise survey in 2009. Box A is a pilot region (35.5°N –36°N, 121.25°E –121.75°E) selected for this study. The light yellow blocks show the locations with significant increases in the five-year average of Chl-a during July from the pre-MAB period (2002–2006) to the MAB period (2008–2012) (see the section of results for details). The background image is a MODIS quasi-true color Red-Green-Blue composite image on 11 April 2011 (R: band 1; G: band 4; B: band 3).</p> Full article ">Figure 2
<p>Data-flow chart showing the removal of macroalgae-contaminated pixels. D0, daily Level-2 standard MODIS aqua Chl-a; D1, the mapped D0 with an equidistant cylindrical projection; D2, the averaged D1 with a moving window of 9 × 9 pixels; D3, D1 deducted by D2; D4, macroalgae-contaminated pixels identified in D3 (D3 &gt; 0.5 µg/L); D5, D1 after the removal of D4; D6, the resampled D5 with a 9 km × 9 km bin average.</p> Full article ">Figure 3
<p>(<b>a</b>) MODIS quasi-true color Red-Green-Blue composite image on 15 July 2009 (R: band 1; G: band 4; B: band 3); (<b>b</b>) color composite image of MODIS Level-2 reflectance: R (Band 7, 2130 nm) G (Band 2, 859 nm) B (Band 1, 645 nm). Light-green slicks show the pixels containing floating macroalgae. (<b>c</b>) Standard Chl-a data product; (<b>d</b>) standard Chl-a data product after the removal of the macroalgae-contaminated pixels. White color indicates land, clouds, and invalid data. All the west-east profiles in pink are the same as the P-P’ line in <a href="#remotesensing-07-12297-f003" class="html-fig">Figure 3</a>a.</p> Full article ">Figure 4
<p>(<b>a</b>) NDVI image corresponding to <a href="#remotesensing-07-12297-f003" class="html-fig">Figure 3</a>; the white slicks pointed out by arrows are macroalgae; the west-east profile (P-P’) is plotted in pink; (<b>b</b>) NDVI and the Level-2 standard Chl-a along the profile (P-P’). Data gaps in Chl-a indicate the pixels with invalid data flagged as “CHLFAIL” or “CLDICE” in the standard product due to the presence of floating macroalgae.</p> Full article ">Figure 5
<p>Changes in the five-year average of the water-column Chl-a for June and July between the pre-MAB period (2002–2006) and the MAB period (2008–2012), generated from the standard level-2 daily Chl-a product after the removal of the macroalgae-contaminated pixels (<a href="#remotesensing-07-12297-f003" class="html-fig">Figure 3</a>). (<b>a</b>) Difference in Chl-a for June (MAB minus pre-MAB), and (<b>b</b>) significance of the difference. (<b>c</b>) Difference in Chl-a for July (MAB minus pre-MAB) and (<b>d</b>) significance of the difference. Green <math display="inline"> <semantics> <mrow> <mstyle mathcolor="lime"> <mo>■</mo> </mstyle> </mrow> </semantics> </math> and red <math display="inline"> <semantics> <mrow> <mstyle mathcolor="red"> <mo>■</mo> </mstyle> </mrow> </semantics> </math> for significant increases and decreases as indicated by <span class="html-italic">t</span>-test (<span class="html-italic">p</span> &lt; 0.05), respectively; blue <math display="inline"> <semantics> <mrow> <mstyle mathcolor="#00B0F0"> <mo>■</mo> </mstyle> </mrow> </semantics> </math> for non-significant changes; white pixels for no results.</p> Full article ">Figure 6
<p>(<b>a</b>) Difference in Chl-a (July minus June) during the pre-MAB period (2002–2006), and (<b>b</b>) significance of the difference. (<b>c</b>) Difference in Chl-a (July minus June) during the MAB period (2008–2012) and (<b>d</b>) significance of the difference. The white dotted line shows the latitude line of 34.5°N. Green <math display="inline"> <semantics> <mrow> <mstyle mathcolor="lime"> <mo>■</mo> </mstyle> </mrow> </semantics> </math> and red <math display="inline"> <semantics> <mrow> <mstyle mathcolor="red"> <mo>■</mo> </mstyle> </mrow> </semantics> </math> for significant increases and decreases as indicated by <span class="html-italic">t</span>-test, respectively; blue <math display="inline"> <semantics> <mrow> <mstyle mathcolor="#00B0F0"> <mo>■</mo> </mstyle> </mrow> </semantics> </math> for non-significant changes; white pixels for no results.</p> Full article ">Figure 7
<p>(<b>a</b>) Monthly Chl-a for June and July between 2002 and 2013. The two solid lines show the overall trends for June and July, respectively; (<b>b</b>) SeaWiFS standard monthly Chl-a of June and July before the super macroalgal blooms in the Yellow Sea (pilot region: 35.5°N–36°N, 121.25°E–121.75°E); (<b>c</b>,<b>d</b>) Monthly Chl-a changes for 2002–2006 and 2008–2012 with and without the removal of macroalgae pixels, respectively. Chl-a values for April, May, August, and September were obtained from Level-3 standard MODIS aqua Chl-a products. Vertical bars show the standard deviations (S.D.) of multi-year monthly Chl-a. 0: standard Chl-a with macroalgae included; 0.5: standard Chl-a with macroalgae excluded using a threshold of 0.5 µg/L (see the Methods section for details of data-processing).</p> Full article ">Figure 8
<p>(<b>a</b>) Monthly sea surface temperature (SST) and (<b>b</b>) photosynthetically active radiation (PAR) for June and July during the 2000–2013 period in the pilot region (box A in <a href="#remotesensing-07-12297-f001" class="html-fig">Figure 1</a>); there is no statistically significant difference between SeaWiFS and MODIS data for the overlapping period. (<b>c</b>) Unclean waters (mainly nutrient polluted waters with water quality levels II, III, IV, and V) in the Jiangsu Shoal (JSS) and the Yellow Sea (YS) during the 2003–2011 period [<a href="#B6-remotesensing-07-12297" class="html-bibr">6</a>]; “YS-JSS” represents the annual area of unclean waters of the YS excluding that of the JSS; “JSS:YS” represents the ratio of unclean waters of the JSS to that of the YS. (<b>d</b>) Level-3 standard MODIS aqua and terra Chl-a for June in the pilot region <span class="html-italic">versus</span> the annual average of nutrient concentrations in the JSS [<a href="#B6-remotesensing-07-12297" class="html-bibr">6</a>].</p> Full article ">Figure 9
<p>Schematic chart showing scenarios of how phytoplankton biomass could be modulated by nutrient supplies and MABs in the May–August period. Se0 (A→F→C→D): the real situation (2008–2012) with MABs and increased nutrient supply; Se1 (A→F→E→D): without increases in nutrient supply and occurrence of MABs; Se2 (A→B→G→D): without occurrence of MABs but with increases in nutrient supply; Se0’(A→F→G→D): the same as scenario Se0 but without nutrient release from macroalgae to the water column in July. Nutrients of phytoplankton biomass (BF) correspond to those consumed by macroalgae in June, and biomass (CG) corresponds to the potential nutrients released from macroalgae die-off for July.</p> Full article ">
1197 KiB  
Article
Mineral Classification of Makhtesh Ramon in Israel Using Hyperspectral Longwave Infrared (LWIR) Remote-Sensing Data
by Gila Notesco, Yaron Ogen and Eyal Ben-Dor
Remote Sens. 2015, 7(9), 12282-12296; https://doi.org/10.3390/rs70912282 - 21 Sep 2015
Cited by 29 | Viewed by 7946
Abstract
Hyperspectral remote-sensing techniques offer an efficient procedure for mineral mapping, with a unique hyperspectral remote-sensing fingerprint in the longwave infrared spectral region enabling identification of the most abundant minerals in the continental crust—quartz and feldspars. This ability was examined by acquiring airborne data [...] Read more.
Hyperspectral remote-sensing techniques offer an efficient procedure for mineral mapping, with a unique hyperspectral remote-sensing fingerprint in the longwave infrared spectral region enabling identification of the most abundant minerals in the continental crust—quartz and feldspars. This ability was examined by acquiring airborne data with the AisaOWL sensor over the Makhtesh Ramon area in Israel. The at-sensor radiance measured from each pixel in a longwave infrared image represents the emissivity, expressing chemical and physical properties such as surface mineralogy, and the atmospheric contribution which is expressed differently during the day and at night. Therefore, identifying similar features in day and night radiance enabled identifying the major minerals in the surface—quartz, silicates (feldspars and clay minerals), gypsum and carbonates—and mapping their spatial distribution. Mineral identification was improved by applying the radiance of an in situ surface that is featureless for minerals but distinctive for the atmospheric contribution as a gain spectrum to each pixel in the image, reducing the atmospheric contribution and emphasizing the mineralogical features. The results were in agreement with the mineralogy of selected rock samples collected from the study area as derived from laboratory X-ray diffraction analysis. The resulting mineral map of the major minerals in the surface was in agreement with the geological map of the area. Full article
(This article belongs to the Special Issue Remote Sensing in Geology)
►▼ Show Figures

Graphical abstract

Graphical abstract
Full article ">Figure 1
<p>(<b>a</b>) The study area (image source: Esri, DigitalGlobe, GeoEye, Earthstar Geographics, CNES/Airbus DS, USDA, USGS, AEX, Getmapping, Aerogrid, IGN, IGP, swisstopo, and the GIS User Community); arrow indicates the acquired LWIR image (band 10.60 μm, day image); (<b>b</b>) Geological map of the area (source: Geological Survey of Israel). The photographs show the surface of about 4 square meters of three sites, out of seventeen sites, which were sampled for the ground truth dataset, as described in <a href="#sec2dot3-remotesensing-07-12282" class="html-sec">Section 2.3</a>.</p> Full article ">Figure 2
<p>At-sensor radiance (Ls; solid curves) and fitted tangent blackbody radiation (Lb; dashed curves) of three ROIs (representing the three sites from <a href="#remotesensing-07-12282-f001" class="html-fig">Figure 1</a>b) in (<b>a</b>) day and (<b>b</b>) night images. Gray shading marks the absorption ranges of water vapor (H<sub>2</sub>O) and ozone (O<sub>3</sub>).</p> Full article ">Figure 3
<p>(<b>a</b>) Day and (<b>b</b>) night surface temperature images as derived from the LWIR images. The hot surface in the daytime and the cold surface at night, compared to air temperatures of 32–22 °C, respectively, indicate a surface with low thermal inertia, suitable for applying the suggested procedure.</p> Full article ">Figure 4
<p>Day and night Ls/Lb spectra of (<b>a</b>) ROI 1 and (<b>b</b>) ROI 2. Different atmospheric contributions are noticeable: ozone’s daytime absorption feature <span class="html-italic">vs.</span> nighttime emission feature ((<b>a</b>); dotted lines) and water vapor’s daytime absorption feature <span class="html-italic">vs.</span> nighttime emission feature ((<b>b</b>); dotted lines). Similar daytime and nighttime absorption features are noticeable ((<b>a</b>) and (<b>b</b>); thick lines).</p> Full article ">Figure 5
<p>Emissivity spectra of minerals based on [<a href="#B14-remotesensing-07-12282" class="html-bibr">14</a>] and resampled to the AisaOWL spectral configuration. The minerals in the figure are mentioned in the text. Thick lines represent the main mineralogical features.</p> Full article ">Figure 6
<p>Day (primary y-axis) and night (secondary y-axis) Ls/Lb spectra of ROI 3. The radiation emitted from the warm surface during the day (60 °C, field measurement) was absorbed by the atmosphere, and the atmospheric emission adds to the radiation emitted from the cold surface at night (22 °C). Dotted lines emphasize daytime absorption features <span class="html-italic">vs.</span> nighttime emission features. There were no similar daytime and nighttime absorption features at any point along the spectrum.</p> Full article ">Figure 7
<p>Day and night approximate (App.) emissivity spectra of the two ROIs, 1 and 2. Thick lines represent the mineralogical features.</p> Full article ">Figure 8
<p>Mineral map of the bottom surface of Makhtesh Ramon. Sampled ROIs are shown and numbered (including ROIs 1, 2 and 3 from <a href="#sec2dot4-remotesensing-07-12282" class="html-sec">Section 2.4</a>).</p> Full article ">Figure 9
<p>Mineral map of the bottom surface of Makhtesh Ramon. The “mafic” minerals refer to mineralogy similar to ROI 3 which was mapped by applying the Spectral Angle Mapper algorithm [<a href="#B15-remotesensing-07-12282" class="html-bibr">15</a>] with the ROI 3 Ls/Lb spectrum as the endmember spectrum.</p> Full article ">Figure 10
<p>(<b>a</b>) Ls/Lb spectra and (<b>b</b>) approximate emissivity spectrum of ROI 4. The relevant features are emphasized with thick lines for the identified minerals.</p> Full article ">Figure 11
<p>Approximate emissivity spectrum of two ROIs representing two types of silicates-rich rocks. The relevant features are emphasized with thick lines for the identified minerals.</p> Full article ">
2397 KiB  
Article
Characteristics of Surface Deformation Detected by X-band SAR Interferometry over Sichuan-Tibet Grid Connection Project Area, China
by Yunshan Meng, Hengxing Lan, Langping Li, Yuming Wu and Quanwen Li
Remote Sens. 2015, 7(9), 12265-12281; https://doi.org/10.3390/rs70912265 - 21 Sep 2015
Cited by 19 | Viewed by 5573
Abstract
The Sichuan-Tibet grid connection project is a national key project implemented in accordance with the developmental needs of Tibet and the living requirements of 700 thousand local residents. It is the first grid project with special high voltage that passes through eastern margin [...] Read more.
The Sichuan-Tibet grid connection project is a national key project implemented in accordance with the developmental needs of Tibet and the living requirements of 700 thousand local residents. It is the first grid project with special high voltage that passes through eastern margin of the Tibetan Plateau. The ground deformation due to widely distributed landslides and debris flow in this area is the major concern to the safety of the project. The multi-temporal interferometry technique is applied to retrieve the surface deformation information using high resolution X-band SAR imagery. The time series of surface deformation is obtained through the sequential high spatial and temporal resolution TerraSAR images (20 scenes of X-band TerraSAR SLC images acquired from 5 January 2014 to 12 December 2014). The results have been correlated with the permafrost activities and intensive precipitation. They show that the study area is prone to slow to moderate ground motion with the range of −30 to +30 mm/year. Seasonal movement is observed due to the freeze-thaw cycle effect and intensive precipitation weather condition. Typical region analysis suggests that the deformation rate tends to increase dramatically during the late spring and late autumn while slightly during the winter time. The correlations of surface deformations with these two main trigger factors were further discussed. The deformation curves of persistent scatterers in the study area showing the distinct seasonal characteristics coincide well with the effect of freeze-thaw cycle and intensive precipitation. The movement occurring at late spring is dominated by the freeze-thaw cycle which is a common phenomenon in such a high-elevated area as the Tibetan Plateau. Intensive precipitation plays more important role in triggering landsides in the summer season. The combining effect of both factors results in fast slope movement in May. The results also suggest that the movement often occur at the middle to toe part of the slope where the combining effect of freeze-thaw cycle and precipitation plays an important role. Therefore the majority of transmission towers are not threatened significantly by geological hazards since they are located on the higher elevation which is beyond the boundary of slope movement. The comparison between field observations and the persistent scatterers interferometry (PSI) results reveals good agreement in obvious deformation accumulations. High uncertainty still exists due to issue of SAR imagery quality and the persistent scatterers interferometry technique. Nevertheless, this study provides an insight into understanding the characteristics of ground movement trend in the complicated eastern Tibet area. Full article
►▼ Show Figures

Graphical abstract

Graphical abstract
Full article ">Figure 1
<p>Overview of the grid project area with elevation displayed as background. Grid lines with different colors represent different voltage levels. Major tectonic faults adjacent to the grid lines are also shown. The red frame indicates the SAR study area.</p> Full article ">Figure 2
<p>The locations of transmission towers, geological hazards, residential sites, and major rivers in the study area are shown on the SAR intensity map.</p> Full article ">Figure 3
<p>Meteorological data of the study area in 2014.</p> Full article ">Figure 4
<p>The PSI monitoring results of the whole study area. (<b>a</b>) The reference point is located on the top of a temple (shown as a black start in the upper figure). (<b>b</b>) The rectangular frames show the locations of the Xiangdui township and (<b>c</b>) a small scale landslide.</p> Full article ">Figure 5
<p>TerraSAR-X PSI monitoring results for landslides with different terrains. The deformation histories of the persistent scatterers sited on landslide and the average seven-days min temperature as well as intensive precipitation time periods are also shown. The locations of landslides are shown in the <a href="#remotesensing-07-12265-f004" class="html-fig">Figure 4</a>.</p> Full article ">Figure 6
<p>TerraSAR-X PSI monitoring results for an unstable surface on the banks of Changqu River. (<b>a</b>) The location of this unstable surface is shown in the insert figure. (<b>b</b>) The time series deformation histories of three points with different elevations are also shown.</p> Full article ">Figure 7
<p>The PSI results around transmission tower 74A are shown using Inverse Distance Weighting interpolation. (<b>a</b>) The tower shown with a black point lies on one side of a ridge. (<b>b</b>) The photo shows the deposits located near one leg of the tower. The location of this tower is shown in the insert figure.</p> Full article ">Figure 8
<p>The surface deformation field along the grid line is shown. Photos (<b>a</b>), (<b>b</b>) and (<b>c</b>) show the landscapes of different areas.</p> Full article ">
6228 KiB  
Article
SPOT-4 (Take 5): Simulation of Sentinel-2 Time Series on 45 Large Sites
by Olivier Hagolle, Sylvia Sylvander, Mireille Huc, Martin Claverie, Dominique Clesse, Cécile Dechoz, Vincent Lonjou and Vincent Poulain
Remote Sens. 2015, 7(9), 12242-12264; https://doi.org/10.3390/rs70912242 - 21 Sep 2015
Cited by 73 | Viewed by 9074
Abstract
This paper presents the SPOT-4 (Take 5) experiment, aimed at providing time series of optical images simulating the repetitivity, the resolution and the large swath of Sentinel-2 images. The aim was to help users set up and test their applications and methods, before [...] Read more.
This paper presents the SPOT-4 (Take 5) experiment, aimed at providing time series of optical images simulating the repetitivity, the resolution and the large swath of Sentinel-2 images. The aim was to help users set up and test their applications and methods, before Sentinel-2 mission data become available. In 2016, when both Sentinel-2 satellites are operational, and for at least fifteen years, users will have access to high resolution time series of images systematically acquired every five days, over the whole Earth land surfaces. Thanks to Sentinel-2’s high revisit frequency, a given surface should be observed without clouds at least once a month, except in the most cloudy periods and regions. In 2013, the Centre National d’Etudes Spatiales (CNES) lowered the orbit altitude of SPOT-4, to place it on a five-day repeat cycle orbit for a duration of five months. This experiment started on 31 January 2013 and lasted until 19 June 2013. SPOT-4 images were acquired every fifth day, over 45 sites scattered in nearly all continents and covering very diverse biomes for various applications. Two ortho-rectified products were delivered for each acquired image that was not fully cloudy, expressed either as top of atmosphere reflectance (Level 1C) or as surface reflectance (Level 2A). An extensive validation campaign was held to check the performances of these products with regard to the multi-temporal registration, the quality of cloud masks, the accuracy of aerosol optical thickness estimates and the quality of surface reflectances. Despite high a priori geo-location errors, it was possible to register the images with an accuracy better than 0.5 pixels in the large majority of cases. Despite the lack of a blue band on the SPOT-4 satellite, the cloud and shadow detection yielded good results, while the aerosol optical thickness was measured with a root mean square error better than 0.06. The surface reflectances after atmospheric correction were compared with in situ data and other satellite data showing little bias and the standard deviation of surface reflectance errors in the range (0.01–0.02). The Take 5 experiment is being repeated in 2015 with the SPOT-5 satellite with an enhanced resolution. Full article
(This article belongs to the Special Issue Lessons Learned from the SPOT4 (Take5): Experiment in Preparation for Sentinel-2)
►▼ Show Figures

Graphical abstract

Graphical abstract
Full article ">Figure 1
<p>The green polygons show the situation of the sites in Europe and North Africa.</p> Full article ">Figure 2
<p>Example of a Level 1C product (<b>left</b>) and a Level 2A (<b>right</b>) product on the Morocco site on 7 March 2013. A (red, green, blue) composite is made with (NIR, red, green) SPOT-4 reflectances, with the same color scale factor in both products. On the Level 2A product, the clouds and shadows, water and snow are respectively circled in green and black, blue and pink.</p> Full article ">Figure 3
<p>Geo-location errors before using ground control points, measured for all of the SPOT-4 (Take 5) scenes with enough cloud-free surface to perform a significant measurement. Although most of the measurements are below 500 m, several scenes with geo-location errors above 1000 m were observed.</p> Full article ">Figure 4
<p>Multi-temporal registration performances obtained for the South Africa, Morocco, Versailles and Sumatra sites, using as a reference the image of the date for which performances are shown equal to zero. The maximum registration error is provided for the best 50%, 70%, 80% and 95% of measurements. In some cases and for the high percentages, the maximum error might exceed one pixel and is not visible in the plot. This happens only for images with a very large cloud cover, but it is always the case for Sumatra.</p> Full article ">Figure 5
<p>On this SPOT-4 (Take 5) image acquired in Provence, the cloud mask is outlined in green, the cloud shadow mask in black, the water mask in blue and the snow mask in pink (in the northeast corner). One can note that faint clouds and cloud shadows are well detected. Image ©CNES (2013), all rights reserved.</p> Full article ">Figure 6
<p>Example of a SPOT-4 (Take 5) image acquired in Argentina on 29 April with small undetected clouds. The detected clouds are circled in green and their shadows in black.</p> Full article ">Figure 7
<p>The left plot shows a map of the number of cloud-free observations (between four and nine), according to the Multi-sensor Atmospheric Correction and Cloud Screening (MACCS) cloud mask, for the site in Belgium, from February–June 2013. The plot on the right shows in blue the percentage of cloud-free pixels for each date. The missing dates are dates with too few cloud-free pixels to allow the ortho-rectification. It should be noted that on that site, the first image of this site with a sufficient number of cloud-free pixels to allow the L1C processing was obtained on 10 March, six weeks after the start of the experiment.</p> Full article ">Figure 8
<p>Comparison of MACCS aerosol optical thickness (AOT) estimates at 550 nm (lines) with the AOT measured <span class="html-italic">in situ</span> on the AeroNet site in Ouarzazate (Morocco) (dots), as a function of time.</p> Full article ">Figure 9
<p>Validation of SPOT-4 (Take 5) AOT estimates with regard to AeroNet <span class="html-italic">in situ</span> measurements over 13 sites in four continents. The blue dots correspond to stable cases, while the red triangles correspond to unstable cases.</p> Full article ">Figure 10
<p>Comparison of SPOT-4 (Take 5) surface reflectances with Robotic Station for Characterizing Atmosphere and Surface (ROSAS) reflectances for the four bands of SPOT-4 (Take 5). The blue circles correspond to ROSAS <span class="html-italic">in situ</span> measurements, while the red squares correspond to SPOT-4 (Take 5). Three of them, observed after rain events, are marked by red diamonds.</p> Full article ">Figure 11
<p>Comparison of SPOT-4 (Take 5) surface reflectances with MODIS reflectances averaged at 5-km resolution, for all SPOT-4 (Take 5) acquisition dates; top, before directional correction; bottom, after directional correction. From left to right, the columns correspond to the green, red, NIR and SWIR bands.</p> Full article ">
6940 KiB  
Article
Quality Assessment of S-NPP VIIRS Land Surface Temperature Product
by Yuling Liu, Yunyue Yu, Peng Yu, Frank M. Göttsche and Isabel F. Trigo
Remote Sens. 2015, 7(9), 12215-12241; https://doi.org/10.3390/rs70912215 - 21 Sep 2015
Cited by 61 | Viewed by 8506
Abstract
The VIIRS Land Surface Temperature (LST) Environmental Data Record (EDR) has reached validated (V1 stage) maturity in December 2014. This study compares VIIRS v1 LST with the ground in situ observations and with heritage LST product from MODIS Aqua and AATSR. Comparisons against [...] Read more.
The VIIRS Land Surface Temperature (LST) Environmental Data Record (EDR) has reached validated (V1 stage) maturity in December 2014. This study compares VIIRS v1 LST with the ground in situ observations and with heritage LST product from MODIS Aqua and AATSR. Comparisons against U.S. SURFRAD ground observations indicate a similar accuracy among VIIRS, MODIS and AATSR LST, in which VIIRS LST presents an overall accuracy of −0.41 K and precision of 2.35 K. The result over arid regions in Africa suggests that VIIRS and MODIS underestimate the LST about 1.57 K and 2.97 K, respectively. The cross comparison indicates an overall close LST estimation between VIIRS and MODIS. In addition, a statistical method is used to quantify the VIIRS LST retrieval uncertainty taking into account the uncertainty from the surface type input. Some issues have been found as follows: (1) Cloud contamination, particularly the cloud detection error over a snow/ice surface, shows significant impacts on LST validation; (2) Performance of the VIIRS LST algorithm is strongly dependent on a correct classification of the surface type; (3) The VIIRS LST quality can be degraded when significant brightness temperature difference between the two split window channels is observed; (4) Surface type dependent algorithm exhibits deficiency in correcting the large emissivity variations within a surface type. Full article
(This article belongs to the Collection Visible Infrared Imaging Radiometers and Applications)
►▼ Show Figures

Graphical abstract

Graphical abstract
Full article ">Figure 1
<p>(<b>a</b>) is Geographic landscape in Gobabeb station in Namibia and (<b>b</b>) is the instrumentation for LST measurement: two radiometers measure the surface-leaving radiance (9.6–11.5 μm) from the gravel plain, which is highly homogenous over at least 2500 km<sup>2</sup>. A third radiometer measures sky radiance.</p> Full article ">Figure 2
<p>Scatter plots of the VIIRS LSTs (<b>a</b>) and MODIS LSTs (<b>b</b>) against the SURFRAD LSTs compared in the period from February 2012 to April 2015. Overall accuracy and precision of the satellite LSTs referring the SURFRAD LSTs are noted, as well as the daytime and nighttime cases. Some VIIRS LST plots are circled as suspicious cloud contaminated plots (red).</p> Full article ">Figure 3
<p>Scatter plots of the VIIRS LSTs (blue) and the AATSR LSTs (red) against the SURFRAD LSTs compared in the period from 1 February 2012 to 8 April 2012. Overall accuracy and precision of the satellite LSTs referring the SURFRAD LSTs are noted. Some VIIRS LST plots are circled as suspicious cloud contaminated plots.</p> Full article ">Figure 4
<p>Validation result against the data in Gobabeb, Namibia in 2012: VIIRS LST (<b>a</b>) and MODIS LST V5 (<b>b</b>).</p> Full article ">Figure 5
<p>Cross-comparison results between VIIRS and AQUA for the whole period and area under analysis. (<b>a</b>) all comparison results under cloud clear condition ; (<b>b</b>) based on a, spatial variation tests are added ; (<b>c</b>) based on b, angle difference is added ; (<b>d</b>) based on c, VIIRS LST is calculated using MODIS data as input and then compare to MODIS LST.</p> Full article ">Figure 6
<p>Cross-comparison results between VIIRS and AQUA of the case study on 28 December 2013. (<b>a</b>) Overall comparison results under cloud clear condition; (<b>b</b>) Brightness temperature comparison of VIIRS band 15 and MODIS Aqua band; (<b>c</b>) the BT difference comparison between VIIRS (BT15-BT16) and MODIS (BT31-BT32); (<b>d</b>) 31 based on a, VIIRS LST is calculated using MODIS data as input and then compare to MODIS LST</p> Full article ">Figure 7
<p>Global BT difference distribution map for 19 December 2014 at daytime (<b>a</b>) and nighttime (<b>b</b>); 4 July 2014 at daytime (<b>c</b>) and nighttime (<b>d</b>).</p> Full article ">Figure 8
<p>LST uncertainty associated with the uncertainty in surface type classification. These values are estimated using the simulation dataset for all surface types and day/night conditions.</p> Full article ">Figure 9
<p>Impact of surface type accuracy (blue line, ranging from 0 to 1) on LST uncertainty (red line, in K) for daytime (<b>a</b>) and nighttime (<b>b</b>).</p> Full article ">
2960 KiB  
Article
Retrieval of Mangrove Aboveground Biomass at the Individual Species Level with WorldView-2 Images
by Yuanhui Zhu, Kai Liu, Lin Liu, Shugong Wang and Hongxing Liu
Remote Sens. 2015, 7(9), 12192-12214; https://doi.org/10.3390/rs70912192 - 21 Sep 2015
Cited by 109 | Viewed by 9063
Abstract
Previous research studies have demonstrated that the relationship between remote sensing-derived parameters and aboveground biomass (AGB) could vary across different species types. However, there are few studies that calibrate reliable statistical models for mangrove AGB. This study quantifies the differences of accuracy in [...] Read more.
Previous research studies have demonstrated that the relationship between remote sensing-derived parameters and aboveground biomass (AGB) could vary across different species types. However, there are few studies that calibrate reliable statistical models for mangrove AGB. This study quantifies the differences of accuracy in AGB estimation between the results obtained with and without the consideration of species types using Worldview-2 images and field surveys. A Back Propagation Artificial Neural Network (BP ANN) based model is developed for the accurate estimation of uneven-aged and dense mangrove forest biomass. The contributions of the input variables are further quantified using a “Weights” method based on BP ANN model. Two types of mangrove species, Sonneratia apetala (S. apetala) and Kandelia candel (K. candel), are examined in this study. Results show that the species type information is the most important variable for AGB estimation, and the red edge band and the associated vegetation indices from WorldView-2 images are more sensitive to mangrove AGB than other bands and vegetation indices. The RMSE of biomass estimation at the incorporation of species as a dummy variable is 19.17% lower than that of the mixed species level. The results demonstrate that species type information obtained from the WorldView-2 images can significantly improve of the accuracy of the biomass estimation. Full article
(This article belongs to the Special Issue Remote Sensing of Mangroves: Observation and Monitoring)
►▼ Show Figures

Graphical abstract

Graphical abstract
Full article ">Figure 1
<p>Location, planting sequence, and field sampling sites in the Qi’ao Island mangroves overlaid on a Worldview-2 images (band 7, 5, 3 false color combination) dated 11 November 2010. (Coordinate system: Universal Transverse Mercator Zone 49N, WGS84)</p> Full article ">Figure 2
<p>Field investigation of mangrove areas: (<b>a</b>) quadrate investigation for mangrove species and (<b>b</b>) DBH measurement for <span class="html-italic">S. apetala</span>.</p> Full article ">Figure 3
<p>Flow chart of mangrove species classification based on WorldView-2 images.</p> Full article ">Figure 4
<p>Subarea division and classified map of the mangrove vegetation on Qi’ao Island.</p> Full article ">Figure 5
<p>Ranked variable importance based on “Weights” method based on the built Dummy Species Model.</p> Full article ">Figure 6
<p>Spatial distribution of mangrove vegetation biomass at the mixed species level based on the artificial neural network (ANN) model in Expt.1.</p> Full article ">Figure 7
<p>Spatial distribution of mangrove vegetation biomass at the incorporation of species as a dummy variable based on the ANN model in Expt.2.</p> Full article ">Figure 8
<p>Spatial distribution of mangrove vegetation biomass at the individual species level based on the ANN model in Expt.3.</p> Full article ">
7171 KiB  
Article
Comparative Analysis of MODIS Time-Series Classification Using Support Vector Machines and Methods Based upon Distance and Similarity Measures in the Brazilian Cerrado-Caatinga Boundary
by Natanael Antunes Abade, Osmar Abílio de Carvalho Júnior, Renato Fontes Guimarães and Sandro Nunes De Oliveira
Remote Sens. 2015, 7(9), 12160-12191; https://doi.org/10.3390/rs70912160 - 18 Sep 2015
Cited by 40 | Viewed by 9385
Abstract
We have mapped the primary native and exotic vegetation that occurs in the Cerrado-Caatinga transition zone in Central Brazil using MODIS-NDVI time series (product MOD09Q1) data over a two-year period (2011–2013). Our methodology consists of the following steps: (a) the development of a [...] Read more.
We have mapped the primary native and exotic vegetation that occurs in the Cerrado-Caatinga transition zone in Central Brazil using MODIS-NDVI time series (product MOD09Q1) data over a two-year period (2011–2013). Our methodology consists of the following steps: (a) the development of a three-dimensional cube composed of the NDVI-MODIS time series; (b) the removal of noise; (c) the selection of reference temporal curves and classification using similarity and distance measures; and (d) classification using support vector machines (SVMs). We evaluated different temporal classifications using similarity and distance measures of land use and land cover considering several combinations of attributes. Among the classification using distance and similarity measures, the best result employed the Euclidean distance with the NDVI-MODIS data by considering more than one reference temporal curve per class and adopting six mapping classes. In the majority of tests, the SVM classifications yielded better results than other methods. The best result among all the tested methods was obtained using the SVM classifier with a fourth-degree polynomial kernel; an overall accuracy of 80.75% and a Kappa coefficient of 0.76 were obtained. Our results demonstrate the potential of vegetation studies in semiarid ecosystems using time-series data. Full article
►▼ Show Figures

Graphical abstract

Graphical abstract
Full article ">Figure 1
<p>Location map of the study area, with the spatial distribution of Caatinga and Cerrado domains indicated.</p> Full article ">Figure 2
<p>Methodological flowchart of the digital image processing.</p> Full article ">Figure 3
<p>Procedure for the temporal-signature selection of land use and vegetal cover considering both the MODIS-NDVI data dimension and the actual dimensions of images from signal components of the MNF transform.</p> Full article ">Figure 4
<p>(<b>a</b>) Landsat-5 TM image from 12 September 2011, RGB color composite of TM bands 4 (red), 5 (green), and 3 (blue) and (<b>b</b>) high-resolution images from Google Earth from 2013.</p> Full article ">Figure 5
<p>MODIS-NDVI time series, (<b>A</b>) original data with noise, (<b>B</b>) time profile softened by the median filter, and (<b>C</b>) refined by the S-G.</p> Full article ">Figure 6
<p>Temporal signatures from the MODIS-NDVI time series and the MNF signal components: (<b>a</b>) deciduous seasonal forest, (<b>b</b>) semi-deciduous seasonal forest, (<b>c</b>) savanna woodland (Cerrado <span class="html-italic">stricto sensu</span>), (<b>d</b>) grassland formations, (<b>e</b>) pasture, (<b>f</b>) annual crops, and (<b>g</b>) perennial crops.</p> Full article ">Figure 7
<p>Classified maps built from (<b>a</b>) one or (<b>b</b>) three temporal curves and holding the remaining factors constant: ED, six classes (land use and land cover), and NDVI-MODIS input data. (<b>c</b>) RGB color composite of NDVI-MODIS images (12/27/2011–04/30/2012–09/05/2012). (<b>d</b>) MNF (components 1-2-4 as RGB), and (<b>e</b>) Landsat-TM (bands 4-5-3 as RGB).</p> Full article ">Figure 8
<p>Classified images considering two types of input data (NDVI-MODIS time series and MNF signal components), three types of similarity and distance measures (SAM, SCM, and ED), and the more specific set of use classes (eight classes).</p> Full article ">Figure 9
<p>MNF signal-component curve for the water (green line) and pasture (red line) classes, which are characterized by different values and similar shapes.</p> Full article ">Figure 10
<p>SVM classification maps considering two types of input data (NDVI-MODIS time series and MNF signal components) and three types of polynomial kernel functions (fourth, fifth, and sixth degrees).</p> Full article ">Figure 11
<p>Comparison between the classifications considering six classes: (<b>a</b>) SVM (fourth-degree polynomial kernel) and (<b>b</b>) ED.</p> Full article ">
5680 KiB  
Article
Rooftop Surface Temperature Analysis in an Urban Residential Environment
by Qunshan Zhao, Soe W. Myint, Elizabeth A. Wentz and Chao Fan
Remote Sens. 2015, 7(9), 12135-12159; https://doi.org/10.3390/rs70912135 - 18 Sep 2015
Cited by 58 | Viewed by 13409
Abstract
The urban heat island (UHI) phenomenon is a significant worldwide problem caused by rapid population growth and associated urbanization. The UHI effect exacerbates heat waves during the summer, increases energy and water consumption, and causes the high risk of heat-related morbidity and mortality. [...] Read more.
The urban heat island (UHI) phenomenon is a significant worldwide problem caused by rapid population growth and associated urbanization. The UHI effect exacerbates heat waves during the summer, increases energy and water consumption, and causes the high risk of heat-related morbidity and mortality. UHI mitigation efforts have increasingly relied on wisely designing the urban residential environment such as using high albedo rooftops, green rooftops, and planting trees and shrubs to provide canopy coverage and shading. Thus, strategically designed residential rooftops and their surrounding landscaping have the potential to translate into significant energy, long-term cost savings, and health benefits. Rooftop albedo, material, color, area, slope, height, aspect and nearby landscaping are factors that potentially contribute. To extract, derive, and analyze these rooftop parameters and outdoor landscaping information, high resolution optical satellite imagery, LIDAR (light detection and ranging) point clouds and thermal imagery are necessary. Using data from the City of Tempe AZ (a 2010 population of 160,000 people), we extracted residential rooftop footprints and rooftop configuration parameters from airborne LIDAR point clouds and QuickBird satellite imagery (2.4 m spatial resolution imagery). Those parameters were analyzed against surface temperature data from the MODIS/ASTER airborne simulator (MASTER). MASTER images provided fine resolution (7 m) surface temperature data for residential areas during daytime and night time. Utilizing these data, ordinary least squares (OLS) regression was used to evaluate the relationships between residential building rooftops and their surface temperature in urban environment. The results showed that daytime rooftop temperature was closely related to rooftop spectral attributes, aspect, slope, and surrounding trees. Night time temperature was only influenced by rooftop spectral attributes and slope. Full article
►▼ Show Figures

Graphical abstract

Graphical abstract
Full article ">Figure 1
<p>Study area.</p> Full article ">Figure 2
<p>Airborne LIDAR point clouds.</p> Full article ">Figure 3
<p>Raw MASTER daytime imagery.</p> Full article ">Figure 4
<p>Raw MASTER night time imagery.</p> Full article ">Figure 5
<p>NAIP imagery.</p> Full article ">Figure 6
<p>Methodology framework (QuickBird image classification, LIDAR point clouds analysis and atmospheric and geometric correction of MASTER imagery).</p> Full article ">Figure 7
<p>Normalized height surface derived by airborne LIDAR point clouds.</p> Full article ">Figure 8
<p>Decision rules for object-oriented classification.</p> Full article ">Figure 9
<p>Land use/land cover map from QuickBird imagery.</p> Full article ">Figure 10
<p>Exposed rooftops.</p> Full article ">Figure 11
<p>Tree shade.</p> Full article ">Figure 12
<p>Rooftop height.</p> Full article ">Figure 13
<p>Rooftop slope.</p> Full article ">Figure 14
<p>Rooftop aspect.</p> Full article ">Figure 15
<p>Albedo estimated from QuickBird imagery.</p> Full article ">Figure 16
<p>Daytime temperature derived from MASTER imagery.</p> Full article ">Figure 17
<p>Night time temperature derived from MASTER imagery.</p> Full article ">
2555 KiB  
Article
To Fill or Not to Fill: Sensitivity Analysis of the Influence of Resolution and Hole Filling on Point Cloud Surface Modeling and Individual Rockfall Event Detection
by Michael J. Olsen, Joseph Wartman, Martha McAlister, Hamid Mahmoudabadi, Matt S. O’Banion, Lisa Dunham and Keith Cunningham
Remote Sens. 2015, 7(9), 12103-12134; https://doi.org/10.3390/rs70912103 - 18 Sep 2015
Cited by 32 | Viewed by 7617
Abstract
Monitoring unstable slopes with terrestrial laser scanning (TLS) has been proven effective. However, end users still struggle immensely with the efficient processing, analysis, and interpretation of the massive and complex TLS datasets. Two recent advances described in this paper now improve the ability [...] Read more.
Monitoring unstable slopes with terrestrial laser scanning (TLS) has been proven effective. However, end users still struggle immensely with the efficient processing, analysis, and interpretation of the massive and complex TLS datasets. Two recent advances described in this paper now improve the ability to work with TLS data acquired on steep slopes. The first is the improved processing of TLS data to model complex topography and fill holes. This processing step results in a continuous topographic surface model that seamlessly characterizes the rock and soil surface. The second is an advance in the automated interpretation of the surface model in such a way that a magnitude and frequency relationship of rockfall events can be quantified, which can be used to assess maintenance strategies and forecast costs. The approach is applied to unstable highway slopes in the state of Alaska, U.S.A. to evaluate its effectiveness. Further, the influence of the selected model resolution and degree of hole filling on the derived slope metrics were analyzed. In general, model resolution plays a pivotal role in the ability to detect smaller rockfall events when developing magnitude-frequency relationships. The total volume estimates are also influenced by model resolution, but were comparatively less sensitive. In contrast, hole filling had a noticeable effect on magnitude-frequency relationships but to a lesser extent than modeling resolution. However, hole filling yielded a modest increase in overall volumetric quantity estimates. Optimal analysis results occur when appropriately balancing high modeling resolution with an appropriate level of hole filling. Full article
(This article belongs to the Special Issue Use of LiDAR and 3D point clouds in Geohazards)
►▼ Show Figures

Graphical abstract

Graphical abstract
Full article ">Figure 1
<p>Flowchart of key steps of the methodology (BFP = Best Fit Plane). (<b>A</b>) Model creation; (<b>B</b>) Model analysis.</p> Full article ">Figure 2
<p>Topographic relief map showing the location of Study Sites <b>A</b>, <b>B</b> and <b>C</b> in Alaska, U.S.A. (Shaded relief map courtesy of the Earth Systems Research Institute, ESRI).</p> Full article ">Figure 3
<p>Perspective view of a sample point cloud for a 1 km section of the Parks Highway obtained with combined TLS and MLS data<span class="html-italic">.</span> The dashed line highlights Study Site C.</p> Full article ">Figure 4
<p>Triangulation scheme adopted from Olsen <span class="html-italic">et al.</span> [<a href="#B40-remotesensing-07-12103" class="html-bibr">40</a>].</p> Full article ">Figure 5
<p>Schematic illustrating the hole filling process using thin plate spline (TPS) interpolation.</p> Full article ">Figure 6
<p>Example of hole filling for the 2014 surface model for Site <b>A</b> using a grid cell size of 0.05 m and window size of 20 (<span class="html-italic">i.e.</span>, TPS search window of 41 × 41 cells (2 × 2 m) but hole fill size of 21 × 21 cells (1 × 1 m))<span class="html-italic">.</span> Inserts <b>B</b> and <b>C</b> show more detail of the TPS effectively filling both large and small holes<span class="html-italic">.</span> The filled holes are shown in magenta while scan data are shown with the natural RGB color.</p> Full article ">Figure 7
<p>Photograph of rock slope study Site A showing three well-defined discontinuity sets.</p> Full article ">Figure 8
<p>Results from an approximately 35 m section of exposed rock at Site A (0.05 m cell size surface model from 2013).</p> Full article ">Figure 9
<p>Rockfall clustering results between 2013 and 2014 with varying model resolution including (<b>A</b>) 0.01 m; (<b>B</b>), 0.05 m; (<b>C</b>) 0.10 m; (<b>D</b>) 0.20 m; and (<b>E</b>) 0.50 m for Site A (90 m). The grey indicates portions of the cliff that did not experience significant erosional change. (Note that these sections may have experienced accretion). Larger white sections indicate holes or spacing between centroid points. Other colors represent a rock cluster where material has dislodged from the cliff<span class="html-italic">.</span> Each cluster is given an ID and colored based on that ID with a repeating pattern so that they can be distinguished from one another<span class="html-italic">.</span> For simplicity of rendering, the point cloud was used rather than the surface model.</p> Full article ">Figure 10
<p>Comparison of magnitude–frequency curves (plotted based on exceedance) for various resolutions and with hole filling (<b>a</b>, HF) and without hole filling (<b>b</b>, NHF) for Site A between 2013 and 2014.</p> Full article ">Figure 11
<p>(<b>a</b>) Comparison of calculated erosion volumes for various resolutions and with (HF) and without (NHF) hole filling for Site A between 2013 and 2014; (<b>b</b>) Comparison of rockfall volumes for Site A with varying hole filling search window size. Computations are based on a cell size of 0.05 m. The plateau indicates that the majority of the holes for Site A were small and filled with a window size of 15 cells (0.775 m).</p> Full article ">Figure 12
<p>Photograph of overhanging rock at Site B. The crest-to-toe vertical height of the slope is approximately 11 m.</p> Full article ">Figure 13
<p>Sample results of surface modeling (0.05 m resolution) from the 2014 survey of the exposed rock at Site B with holes filled for the (<b>a</b>) west and (<b>b</b>) east portions of the site; (<b>c</b>) Close-up wireframe model of a section for detail.</p> Full article ">Figure 13 Cont.
<p>Sample results of surface modeling (0.05 m resolution) from the 2014 survey of the exposed rock at Site B with holes filled for the (<b>a</b>) west and (<b>b</b>) east portions of the site; (<b>c</b>) Close-up wireframe model of a section for detail.</p> Full article ">Figure 14
<p>Evaluation of the influence of varying the hole filling window size (ws) on the resulting rockfall clusters (2013–2014)<span class="html-italic">.</span> Note that the figure uses the same coloring scheme as described in <a href="#remotesensing-07-12103-f009" class="html-fig">Figure 9</a> with grey representing areas of no significant erosion and white indicating holes in the dataset<span class="html-italic">.</span> For reference, this section is 110 m long. (<b>A</b>) no holes filled; (<b>B</b>) holes filled (ws = 5); (<b>C</b>) holes filled (ws = 10); (<b>D</b>) holes filled (ws = 20); (<b>E</b>) holes filled (ws = 50);</p> Full article ">Figure 15
<p>Comparison of magnitude–frequency curves (plotted based on exceedance) for various resolutions and with (<b>a</b>, HF) and without (<b>b</b>, NHF) hole filling for Site B between 2013 and 2014.</p> Full article ">Figure 16
<p>(<b>a</b>) Comparison of calculated erosion volumes for various resolutions and with (HF) and without (NHF) hole filling for Site B between 2013 and 2014<span class="html-italic">.</span> Note that the hole filling window size was adjusted for each cell size analyzed such that it filled smaller holes, but did not fill excessively large holes; (<b>b</b>) Comparison of rockfall volumes for Site A with varying hole filling search window size<span class="html-italic">.</span> For this analysis, a constant cell size of 0.05 m was used<span class="html-italic">.</span> Note that at a window size of 50 (2.5 m), all holes are filled so the volume does not vary significantly below that<span class="html-italic">.</span></p> Full article ">Figure 17
<p>Photograph of Site C<span class="html-italic">.</span> This photograph captures approximately 0.5 km of the slope along the highway<span class="html-italic">.</span> The highway and river are obscured by the lower trees.</p> Full article ">Figure 18
<p>Perspective view of the distribution of comparison points obtained with a total station to the 2012 surface model (25 cm resolution).</p> Full article ">Figure 19
<p>Map of individual failure clusters between the 2013 and 2014 surveys at Site C (0.10 m resolution). These clusters are colored using the same scheme described in <a href="#remotesensing-07-12103-f009" class="html-fig">Figure 9</a> where grey indicates no significant erosion and white represents holes.</p> Full article ">Figure 20
<p>Comparison of magnitude–frequency curves for various resolutions and with (<b>a</b>, HF) and without (<b>b</b>, NHF) hole filling between 2013 and 2014.</p> Full article ">Figure 21
<p>Comparison of calculated erosion volumes for various resolutions and with (HF) and without (NHF) hole filling at site GG239 between 2013 and 2014.</p> Full article ">Figure 22
<p>Comparison of magnitude-frequency relationships (2013 to 2014, 0.05 cm holes filled models) and power law regressions for each site, normalized to an equivalent 1000 m<sup>2</sup> area based on events &gt;0.001 m<sup>3</sup> (<b>a</b>) with and (<b>b</b>) without hole filling and events &gt;0.01 m<sup>3</sup> (<b>c</b>) with and (<b>d</b>) without hole filling.</p> Full article ">Figure 22 Cont.
<p>Comparison of magnitude-frequency relationships (2013 to 2014, 0.05 cm holes filled models) and power law regressions for each site, normalized to an equivalent 1000 m<sup>2</sup> area based on events &gt;0.001 m<sup>3</sup> (<b>a</b>) with and (<b>b</b>) without hole filling and events &gt;0.01 m<sup>3</sup> (<b>c</b>) with and (<b>d</b>) without hole filling.</p> Full article ">
2578 KiB  
Article
Multi-Temporal Landsat Images and Ancillary Data for Land Use/Cover Change (LULCC) Detection in the Southwest of Burkina Faso, West Africa
by Benewinde J-B. Zoungrana, Christopher Conrad, Leonard K. Amekudzi, Michael Thiel, Evariste Dapola Da, Gerald Forkuor and Fabian Löw
Remote Sens. 2015, 7(9), 12076-12102; https://doi.org/10.3390/rs70912076 - 18 Sep 2015
Cited by 55 | Viewed by 9315
Abstract
Accurate quantification of land use/cover change (LULCC) is important for efficient environmental management, especially in regions that are extremely affected by climate variability and continuous population growth such as West Africa. In this context, accurate LULC classification and statistically sound change area estimates [...] Read more.
Accurate quantification of land use/cover change (LULCC) is important for efficient environmental management, especially in regions that are extremely affected by climate variability and continuous population growth such as West Africa. In this context, accurate LULC classification and statistically sound change area estimates are essential for a better understanding of LULCC processes. This study aimed at comparing mono-temporal and multi-temporal LULC classifications as well as their combination with ancillary data and to determine LULCC across the heterogeneous landscape of southwest Burkina Faso using accurate classification results. Landsat data (1999, 2006 and 2011) and ancillary data served as input features for the random forest classifier algorithm. Five LULC classes were identified: woodland, mixed vegetation, bare surface, water and agricultural area. A reference database was established using different sources including high-resolution images, aerial photo and field data. LULCC and LULC classification accuracies, area and area uncertainty were computed based on the method of adjusted error matrices. The results revealed that multi-temporal classification significantly outperformed those solely based on mono-temporal data in the study area. However, combining mono-temporal imagery and ancillary data for LULC classification had the same accuracy level as multi-temporal classification which is an indication that this combination is an efficient alternative to multi-temporal classification in the study region, where cloud free images are rare. The LULCC map obtained had an overall accuracy of 92%. Natural vegetation loss was estimated to be 17.9% ± 2.5% between 1999 and 2011. The study area experienced an increase in agricultural area and bare surface at the expense of woodland and mixed vegetation, which attests to the ongoing deforestation. These results can serve as means of regional and global land cover products validation, as they provide a new validated data set with uncertainty estimates in heterogeneous ecosystems prone to classification errors. Full article
(This article belongs to the Special Issue Carbon Cycle, Global Change, and Multi-Sensor Remote Sensing)
►▼ Show Figures

Graphical abstract

Graphical abstract
Full article ">Figure 1
<p>Situation of the study area.</p> Full article ">Figure 2
<p>Overall, user’s and producer’s accuracies of the mono-temporal classifications of 2011 derived from adjusted error matrix. User’s and producer’s accuracies are averages of the class-wise assessments.</p> Full article ">Figure 3
<p>Remotely sensed bands and ancillary data contributions to LULC classification based on mean decrease accuracy (MDA) score of RF mono-temporal image plus ancillary data classification.</p> Full article ">Figure 4
<p>LULC spatial distribution in the study area in 1999, 2006 and 2011.</p> Full article ">Figure 5
<p>Distribution of LULCC in the study area between 1999 and 2011.</p> Full article ">Figure 6
<p>Harvested crop intermixed with trees in the study area.</p> Full article ">Figure 7
<p>(<b>a</b>) Trees cut for new cropland; and (<b>b</b>) fuel wood collected in an area devastated by bushfire.</p> Full article ">Figure 8
<p>Rainfall variability in the study area (1981–2012) expressed as standardized precipitation index (SPI) distribution (Rainfall data collected from the national direction of meteorology of Burkina Faso).</p> Full article ">
3070 KiB  
Article
Evaluation of Medium Spatial Resolution BRDF-Adjustment Techniques Using Multi-Angular SPOT4 (Take5) Acquisitions
by Martin Claverie, Eric Vermote, Belen Franch, Tao He, Olivier Hagolle, Mohamed Kadiri and Jeff Masek
Remote Sens. 2015, 7(9), 12057-12075; https://doi.org/10.3390/rs70912057 - 18 Sep 2015
Cited by 24 | Viewed by 6133
Abstract
High-resolution sensor Surface Reflectance (SR) data are affected by surface anisotropy but are difficult to adjust because of the low temporal frequency of the acquisitions and the low angular sampling. This paper evaluates five high spatial resolution Bidirectional Reflectance Distribution Function (BRDF) adjustment [...] Read more.
High-resolution sensor Surface Reflectance (SR) data are affected by surface anisotropy but are difficult to adjust because of the low temporal frequency of the acquisitions and the low angular sampling. This paper evaluates five high spatial resolution Bidirectional Reflectance Distribution Function (BRDF) adjustment techniques. The evaluation is based on the noise level of the SR Time Series (TS) corrected to a normalized geometry (nadir view, 45° sun zenith angle) extracted from the multi-angular acquisitions of SPOT4 over three study areas (one in Arizona, two in France) during the five-month SPOT4 (Take5) experiment. Two uniform techniques (Cst, for Constant, and Av, for Average), relying on the Vermote–Justice–Bréon (VJB) BRDF method, assume no variation in space of the BRDF shape. Two methods (VI-dis, for NDVI-based disaggregation and LC-dis, for Land-Cover based disaggregation) are based on disaggregation of the MODIS-derived BRDF VJB parameters using vegetation index and land cover, respectively. The last technique (LUM, for Look-Up Map) relies on the MCD43 MODIS BRDF products and a crop type data layer. The VI-dis technique produced the lowest level of noise corresponding to the most effective adjustment: reduction from directional to normalized SR TS noises by 40% and 50% on average, for red and near-infrared bands, respectively. The uniform techniques displayed very good results, suggesting that a simple and uniform BRDF-shape assumption is good enough to adjust the BRDF in such geometric configuration (the view zenith angle varies from nadir to 25°). The most complex techniques relying on land cover (LC-dis and LUM) displayed contrasting results depending on the land cover. Full article
(This article belongs to the Special Issue Lessons Learned from the SPOT4 (Take5): Experiment in Preparation for Sentinel-2)
►▼ Show Figures

Graphical abstract

Graphical abstract
Full article ">Figure 1
<p>Maps of the study areas, including the SPOT4 satellite tracks (cyan lines, daytime only) during SPOT4 (Take5) experiment. Overlap areas for ProvLanguedoc and Sudmipy sites are shown with yellow polygons; the overlap area for Maricopa site corresponds to the entire footprint. Images are displayed using a false color composite (red = NIR, green = Red, blue = Green). Black boxes in A, B and C correspond to the spatial extend of the referred data frames.</p> Full article ">Figure 2
<p>Sun and view angles of the SPOT4 (Take5) data.</p> Full article ">Figure 3
<p>V and R VJB (Vermote-Justice-Bréon) coefficients retrieval for the constant technique. Each line refers to the relationship as defined in Equation (7) for one specific day<span class="html-italic"><sub>i</sub></span> (color bar in the right). (<b>a</b>) red band; (<b>b</b>) NIR band.</p> Full article ">Figure 4
<p>Red and NIR per-class Maricopa time series (TS) of the Surface Reflectance (SR) without BRDF adjustment (No-Adj) and time series of the normalized Surface Reflectance (SR, nadir view and 45° sun zenith angle) retrieved from one of the five BRDF-adjustment techniques: Constant model (Cst), average model (Av), NDVI-based disaggregation model (VI-dis), Land-cover-based disaggregation model (LC-dis), LUM technique (LUM). Two consecutive cloud-free measurements with one-day lag (<span class="html-italic">i.e.</span>, acquired from two satellite tracks) are connected with a line to highlight the deviation related to the anisotropy effect. Interpreted land-cover class and computed Noises of the displayed TS are reported in the right panels.</p> Full article ">Figure 5
<p>Scatterplots of Maricopa surface reflectance (<span class="html-italic">ρ</span>) between consecutive days: Day<span class="html-italic"><sub>i</sub></span> and day<span class="html-italic"><sub>i</sub></span> + 1 corresponding to the western and eastern acquisitions, respectively. The two subplot rows refer to the two spectral bands, the six subplot columns refer to the five BRDF-adjusted SR (see <a href="#remotesensing-07-12057-f004" class="html-fig">Figure 4</a> for naming) and the non-BRDF-adjusted SR. (<b>a</b>): No-Adj-Red; (<b>b</b>): Cst-Red; (<b>c</b>): AV-Red; (<b>d</b>): VI-dis-Red; (<b>e</b>): LC-dis-Red; (<b>f</b>): LUM-Red; (<b>g</b>) No-Adj-NIR; (<b>h</b>): Cst-NIR; (<b>i</b>): Av-NIR; (<b>j</b>): VI-dis-NIR; (<b>k</b>): LC-dis-NIR; (<b>l</b>): LUM-NIR.</p> Full article ">Figure 6
<p>Subsets of Maricopa surface reflectance (SR) time series noise. The two subplot rows refer to the two spectral bands, the six subplot columns refer to the non-BRDF-adjusted SR and the five BRDF-adjusted SR (see caption of <a href="#remotesensing-07-12057-f004" class="html-fig">Figure 4</a> for naming). Circle in subset-l identifies barley fields with issue in the BRDF-adjustments. (<b>a</b>): No-Adj-Red; (<b>b</b>): Cst-Red; (<b>c</b>): AV-Red; (<b>d</b>): VI-dis-Red; (<b>e</b>): LC-dis-Red; (<b>f</b>): LUM-Red; (<b>g</b>) No-Adj-NIR; (<b>h</b>): Cst-NIR; (<b>i</b>): Av-NIR; (<b>j</b>): VI-dis-NIR; (<b>k</b>): LC-dis-NIR; (<b>l</b>): LUM-NIR.</p> Full article ">Figure 7
<p>Noise (first column) and NoiseRatio (second column) distributions retrieved from Maricopa normalized SR TS (using five BRDF-adjustment techniques in colored lines, see caption of <a href="#remotesensing-07-12057-f004" class="html-fig">Figure 4</a> for naming) and from non-BRDF-adjusted SR TS (No-adj, black line).</p> Full article ">Figure 8
<p>Noise and NoiseRatio distributions retrieved from ProvLanguedoc site. See caption of <a href="#remotesensing-07-12057-f007" class="html-fig">Figure 7</a> for details.</p> Full article ">Figure 9
<p>Noise and NoiseRatio distributions retrieved from SudMipy site. See caption of <a href="#remotesensing-07-12057-f007" class="html-fig">Figure 7</a> for details.</p> Full article ">Figure 10
<p>Median NoiseRatio values for the five BRDF-adjustment techniques, the two spectral bands, the three sites (Mar = Maricopa, SMP = Sudmipy, PL = ProvLanguedoc), and five majors Land-Cover (LC, Agriculture, Grassland, Shrubland, Forest, Bare soil). Some LC-site combinations were empty and removed. LUM, which was applied only on Maricopa site, was removed from “All sites”. LC were deduced from the LC classification (<a href="#sec4dot2-remotesensing-07-12057" class="html-sec">Section 4.2</a>) based on visual interpretation. Cells are colored using the color bar located in the upper left and Median Noise values of each band-site-LC configuration, corresponding to half-a-row. Lower value of each configuration is written in bold with a star. Last column corresponds to the number of SPOT4 pixels before BRDF-adjustment (in millions).</p> Full article ">
4076 KiB  
Article
Temporal Monitoring of the Soil Freeze-Thaw Cycles over a Snow-Covered Surface by Using Air-Launched Ground-Penetrating Radar
by Khan Zaib Jadoon, Lutz Weihermüller, Matthew F. McCabe, Davood Moghadas, Harry Vereecken and Sebastíen Lambot
Remote Sens. 2015, 7(9), 12041-12056; https://doi.org/10.3390/rs70912041 - 18 Sep 2015
Cited by 16 | Viewed by 7370
Abstract
We tested an off-ground ground-penetrating radar (GPR) system at a fixed location over a bare agricultural field to monitor the soil freeze-thaw cycles over a snow-covered surface. The GPR system consisted of a monostatic horn antenna combined with a vector network analyzer, providing [...] Read more.
We tested an off-ground ground-penetrating radar (GPR) system at a fixed location over a bare agricultural field to monitor the soil freeze-thaw cycles over a snow-covered surface. The GPR system consisted of a monostatic horn antenna combined with a vector network analyzer, providing an ultra-wideband stepped-frequency continuous-wave radar. An antenna calibration experiment was performed to filter antenna and back scattered effects from the raw GPR data. Near the GPR setup, sensors were installed in the soil to monitor the dynamics of soil temperature and dielectric permittivity at different depths. The soil permittivity was retrieved via inversion of time domain GPR data focused on the surface reflection. Significant effects of soil dynamics were observed in the time-lapse GPR, temperature and dielectric permittivity measurements. In particular, five freeze and thaw events were clearly detectable, indicating that the GPR signals respond to the contrast between the dielectric permittivity of frozen and thawed soil. The GPR-derived permittivity was in good agreement with sensor observations. Overall, the off-ground nature of the GPR system permits non-invasive time-lapse observation of the soil freeze-thaw dynamics without disturbing the structure of the snow cover. The proposed method shows promise for the real-time mapping and monitoring of the shallow frozen layer at the field scale. Full article
►▼ Show Figures

Graphical abstract

Graphical abstract
Full article ">Figure 1
<p>Measured Green’s function in the time domain, where <math display="inline"> <msub> <mi>t</mi> <mrow> <mi>m</mi> <mi>i</mi> <mi>n</mi> </mrow> </msub> </math> and <math display="inline"> <msub> <mi>t</mi> <mrow> <mi>m</mi> <mi>a</mi> <mi>x</mi> </mrow> </msub> </math> define the time window focused on the surface reflection in which inversion is performed. Time <math display="inline"> <msub> <mi>t</mi> <mi>i</mi> </msub> </math> corresponds approximately to the surface interface in the space domain, and the maximum amplitude of Green’s function is defined by the peak-to-peak <math display="inline"> <mrow> <mi>P</mi> <mi>t</mi> <mi>P</mi> </mrow> </math>.</p> Full article ">Figure 2
<p>Decagon EC-TM capacitance probes were installed near the footprint of the ground-penetrating radar (GPR) antenna in a trench to monitor the vertical and lateral dynamics of soil temperature and dielectric permittivity (<b>a</b>), and the fully-automated off-ground GPR setup with an antenna fixed at 110 cm above the ground (<b>b</b>) (after [<a href="#B42-remotesensing-07-12041" class="html-bibr">42</a>]).</p> Full article ">Figure 3
<p>(<b>a</b>) Time series of air temperature and (<b>b</b>) soil temperature recorded at shallow depths (2, 3, 4, 5, 7, 8 cm) over a nine-day period. Arrows indicate the freeze-thaw cycles as the soil temperature dynamics is below 0 <math display="inline"> <msup> <mrow/> <mo>∘</mo> </msup> </math>C.</p> Full article ">Figure 4
<p>Time domain representation (b-scan) of radar measurement performed with the horn antenna at different heights above a perfect electrical conductor (copper sheet). (<b>a</b>) Measured raw GPR signal <math display="inline"> <mrow> <msub> <mi>S</mi> <mn>11</mn> </msub> <mrow> <mo>(</mo> <mi>t</mi> <mo>)</mo> </mrow> </mrow> </math> and (<b>b</b>) the filtered signal <math display="inline"> <mrow> <msubsup> <mi>g</mi> <mrow> <mi>x</mi> <mi>x</mi> </mrow> <mo>↑</mo> </msubsup> <mrow> <mo>(</mo> <mi>t</mi> <mo>)</mo> </mrow> </mrow> </math> using the antenna model.</p> Full article ">Figure 5
<p>(<b>a</b>) Radar signal in the time domain for a single measurement and (<b>b</b>) peak-to-peak (<math display="inline"> <mrow> <mi>P</mi> <mi>t</mi> <mi>P</mi> </mrow> </math>) maximum amplitude of the time-lapse GPR data, with the effect of the five freezing and thawing events clearly observed as marked by arrows.</p> Full article ">Figure 6
<p>Time series of the relative permittivity (<math display="inline"> <msub> <mtext>ε</mtext> <mi>r</mi> </msub> </math>) of the soil recorded at shallow depths (2, 3, 4, 5, 7, 8 cm) and estimated from time-lapse GPR data. Five freezing and thawing cycles can be clearly observed over a period of nine days, as indicated by arrows.</p> Full article ">
2390 KiB  
Article
Assessing the Suitability of Future Multi- and Hyperspectral Satellite Systems for Mapping the Spatial Distribution of Norway Spruce Timber Volume
by Sascha Nink, Joachim Hill, Henning Buddenbaum, Johannes Stoffels, Thomas Sachtleber and Joachim Langshausen
Remote Sens. 2015, 7(9), 12009-12040; https://doi.org/10.3390/rs70912009 - 18 Sep 2015
Cited by 20 | Viewed by 6160
Abstract
The availability of accurate and timely information on timber volume is important for supporting operational forest management. One option is to combine statistical concepts (e.g., small area estimates) with specifically designed terrestrial sampling strategies to provide estimations also on the level of administrative [...] Read more.
The availability of accurate and timely information on timber volume is important for supporting operational forest management. One option is to combine statistical concepts (e.g., small area estimates) with specifically designed terrestrial sampling strategies to provide estimations also on the level of administrative units such as forest districts. This may suffice for economic assessments, but still fails to provide spatially explicit information on the distribution of timber volume within these management units. This type of information, however, is needed for decision-makers to design and implement appropriate management operations. The German federal state of Rhineland-Palatinate is currently implementing an object-oriented database that will also allow the direct integration of Earth observation data products. This work analyzes the suitability of forthcoming multi- and hyperspectral satellite imaging systems for producing local distribution maps for timber volume of Norway spruce, one of the most economically important tree species. In combination with site-specific inventory data, fully processed hyperspectral data sets (HyMap) were used to simulate datasets of the forthcoming EnMAP and Sentinel-2 systems to establish adequate models for estimating timber volume maps. The analysis included PLS regression and the k-NN method. Root Mean Square Errors between 21.6% and 26.5% were obtained, where k-NN performed slightly better than PLSR. It was concluded that the datasets of both simulated sensor systems fulfill accuracy requirements to support local forest management operations and could be used in synergy. Sentinel-2 can provide meaningful volume distribution maps in higher geometric resolution, while EnMAP, due to its hyperspectral coverage, can contribute complementary information, e.g., on biophysical conditions. Full article
►▼ Show Figures

Graphical abstract

Graphical abstract
Full article ">Figure 1
<p>Study areas in the German federal state of Rhineland-Palatinate (black rectangles).</p> Full article ">Figure 2
<p>Simulated EnMAP (lines) and Sentinel-2 spectra (squares) of various tree species and stand structures in comparison to the spectral response functions used for generating the datasets based on interpolated HyMap reflectance spectra.</p> Full article ">Figure 3
<p>Comparison of timber volume from the plot-based Federal State Forest Inventory (FSFI 2003) and the corresponding expert assessments on stand level (FID, 1998–2009).</p> Full article ">Figure 4
<p>(<b>a</b>) Single band spectral response <span class="html-italic">vs.</span> timber volume, with identified outliers (red circles). (<b>b</b>) R² over the spectral range between 450 and 2500 nm for simulated EnMAP and Sentinel-2 data before (gray) and after (black) elimination of the outliers.</p> Full article ">Figure 5
<p>Images of the first three principal components of EnMAP data.</p> Full article ">Figure 6
<p>Loo-cv values of RMSE, AIC, and SBC for the PLSR with simulated EnMAP data (<b>a</b>) and Sentinel-2 data (<b>b</b>).</p> Full article ">Figure 7
<p>PLS regression coefficients for all EnMAP and Sentinel-2 spectral bands.</p> Full article ">Figure 8
<p>Estimated timber volume for EnMAP (<b>a</b>) and Sentinel-2 (<b>b</b>) in comparison to observed values on stand level, based on PLS-regression with all spectral bands.</p> Full article ">Figure 9
<p>Timber volume distribution map based on FID data.</p> Full article ">Figure 10
<p>Timber volume prediction map based on PLSR with simulated Sentinel-2 data.</p> Full article ">Figure 11
<p>RMSEcv as a function of the number of neighbors for EnMAP and Sentinel-2 based optimized <span class="html-italic">k</span>-NN estimation, using PC1 and PC3.</p> Full article ">Figure 12
<p><span class="html-italic">k</span>-NN-based estimations in comparison to observed values for EnMAP (<b>a</b>) and Sentinel-2 (<b>b</b>) on the level of administrative forest units.</p> Full article ">Figure 13
<p>Structural differences and their appearance in aerial photos during the development stages of Norway spruce forest stands.</p> Full article ">Figure 14
<p>Timber volume prediction map based on PLSR with simulated EnMAP data.</p> Full article ">Figure 15
<p>Timber volume prediction map based on <span class="html-italic">k</span>-NN with simulated EnMAP data.</p> Full article ">Figure 16
<p>Timber volume prediction map based on <span class="html-italic">k</span>-NN with simulated Sentinel-2 data.</p> Full article ">
1381 KiB  
Article
Remote Sensing Based Two-Stage Sampling for Accuracy Assessment and Area Estimation of Land Cover Changes
by Heinz Gallaun, Martin Steinegger, Roland Wack, Mathias Schardt, Birgit Kornberger and Ursula Schmitt
Remote Sens. 2015, 7(9), 11992-12008; https://doi.org/10.3390/rs70911992 - 18 Sep 2015
Cited by 15 | Viewed by 6168
Abstract
Land cover change processes are accelerating at the regional to global level. The remote sensing community has developed reliable and robust methods for wall-to-wall mapping of land cover changes; however, land cover changes often occur at rates below the mapping errors. In the [...] Read more.
Land cover change processes are accelerating at the regional to global level. The remote sensing community has developed reliable and robust methods for wall-to-wall mapping of land cover changes; however, land cover changes often occur at rates below the mapping errors. In the current publication, we propose a cost-effective approach to complement wall-to-wall land cover change maps with a sampling approach, which is used for accuracy assessment and accurate estimation of areas undergoing land cover changes, including provision of confidence intervals. We propose a two-stage sampling approach in order to keep accuracy, efficiency, and effort of the estimations in balance. Stratification is applied in both stages in order to gain control over the sample size allocated to rare land cover change classes on the one hand and the cost constraints for very high resolution reference imagery on the other. Bootstrapping is used to complement the accuracy measures and the area estimates with confidence intervals. The area estimates and verification estimations rely on a high quality visual interpretation of the sampling units based on time series of satellite imagery. To demonstrate the cost-effective operational applicability of the approach we applied it for assessment of deforestation in an area characterized by frequent cloud cover and very low change rate in the Republic of Congo, which makes accurate deforestation monitoring particularly challenging. Full article
►▼ Show Figures

Graphical abstract

Graphical abstract
Full article ">Figure 1
<p>General workflow.</p> Full article ">Figure 2
<p>Study site located in the northern part of the Republic of Congo with deforested areas highlighted in red.</p> Full article ">Figure 3
<p>Example of a six-month time lag between a Landsat ETM scene (left image) and a RapidEye scene (center image), and example of a geometric shift between a reference scene and the provided forest map (highlighted in yellow), seen in the right image.</p> Full article ">Figure 4
<p>Bootstrapping for estimating area of change from forest land to settlement, with <span class="html-italic">n</span> = 200 runs for first-phase sampling and <span class="html-italic">n</span> = 200 runs for second-phase sampling.</p> Full article ">
2125 KiB  
Article
The Use of Multi-Temporal Landsat Imageries in Detecting Seasonal Crop Abandonment
by Noryusdiana Mohamad Yusoff and Farrah Melissa Muharam
Remote Sens. 2015, 7(9), 11974-11991; https://doi.org/10.3390/rs70911974 - 18 Sep 2015
Cited by 27 | Viewed by 7524
Abstract
Abandonment of agricultural land is a global issue and a waste of resources and brings a negative impact on the local economy. It is also one of the key contributing factors in certain environmental problems, such as soil erosion and carbon sequestration. In [...] Read more.
Abandonment of agricultural land is a global issue and a waste of resources and brings a negative impact on the local economy. It is also one of the key contributing factors in certain environmental problems, such as soil erosion and carbon sequestration. In order to address such problems related to land abandonment, their spatial distribution must first be precisely identified. Hence, this study proposes the use of multi-temporal Landsat imageries, together with crop phenology information and an object-oriented classification technique, to identify abandoned paddy and rubber areas. Results indicate that Landsat time-series images were highly beneficial and, in fact, essential in identifying abandoned paddy and rubber areas, particularly due to the unique phenology of these seasonal crops. To differentiate between abandoned and non-abandoned paddy areas, a minimum of three time-series images, mainly acquired during the planting seasons is required. For rubber, multi-temporal images should be examined in order to confirm the wintering season. The study demonstrates the advantages of using multi-temporal Landsat imageries in identifying abandoned paddy and rubber areas wherein an accuracy of 93.33% ± 14% and 83.33% ± 1%, respectively, were achieved. Full article
(This article belongs to the Special Issue Mapping the Dynamics of Forest Plantations in Tropical and Subtropical Regions from Multi-Source Remote Sensing)
►▼ Show Figures

Graphical abstract

Graphical abstract
Full article ">Figure 1
<p>Study area in Mukim of Sungai Siput and Kuala Kangsar, Perak, Malaysia.</p> Full article ">Figure 2
<p>Workflow for abandoned and non-abandoned paddy and rubber feature extraction.</p> Full article ">Figure 3
<p>Phenology for paddy; (<b>i</b>) non-abandoned paddy; (<b>ii</b>) abandoned paddy.</p> Full article ">Figure 4
<p>Landsat time-series imagery for three selected paddy areas (<b>a</b>–<b>i</b>); (<b>4.1</b>) spectral reflectance for area I; (<b>4.2</b>) spectral reflectance for area II; and (<b>4.3</b>) spectral reflectance for area III.</p> Full article ">Figure 5
<p>Phenology for rubber; (<b>i</b>) non-abandoned rubber; (<b>ii</b>) abandoned rubber.</p> Full article ">Figure 6
<p>Multi-temporal Landsat OLI imagery used to choose the right satellite image in discriminate between non-abandoned rubber and oil palm; (<b>a</b>) images dated as 4 February 2014; (<b>b</b>) images dated as 28 June 2014; (<b>c</b>) images dated as 16 September 2014; and abandoned rubber appear permanently green using time series image (<b>d</b>) images dated as 4 February 2014; (<b>e</b>) images dated as 28 June 2014; (<b>f</b>) images dated as 16 September 2014.</p> Full article ">Figure 7
<p>Paddy feature extraction using time-series images and historical image.</p> Full article ">Figure 8
<p>Classification Map.</p> Full article ">
4132 KiB  
Article
Rapid Response to a Typhoon-Induced Flood with an SAR-Derived Map of Inundated Areas: Case Study and Validation
by Hsiao-Wei Chung, Cheng-Chien Liu, I-Fan Cheng, Yun-Ruei Lee and Ming-Chang Shieh
Remote Sens. 2015, 7(9), 11954-11973; https://doi.org/10.3390/rs70911954 - 18 Sep 2015
Cited by 35 | Viewed by 10161
Abstract
We report the successful case of a rapid response to a flash flood in I-Lan County of Taiwan with a map of inundated areas derived from COSMO-SkyMed 1 radar satellite imagery within 24 hours. The flood was caused by the intensive precipitation brought [...] Read more.
We report the successful case of a rapid response to a flash flood in I-Lan County of Taiwan with a map of inundated areas derived from COSMO-SkyMed 1 radar satellite imagery within 24 hours. The flood was caused by the intensive precipitation brought by Typhoon Soulik in July 2013. Based on the ensemble forecasts of trajectory, an urgent request of spaceborne SAR imagery was made 24 hours before Typhoon Soulik made landfall. Two COSMO-SkyMed images were successfully acquired when the center of Typhoon Soulik had just crossed the northern part of Taiwan. The standard level-1b product (radiometric-corrected, geometric-calibrated and orthorectified image) was generated by using the off-the-shelf SARscape software. Following the same approach used with the Expert Landslide and Shadow Area Delineating System, the regional threshold of each tile image was determined to delineate still water surface and quasi-inundated areas in a fully-automatic manner. The results were overlaid on a digital elevation model, and the same tile was visually compared to an optical image taken by Formosat-2 before this event. With this ancillary information, the inundated areas were accurately and quickly identified. The SAR-derived map of inundated areas was published on a web-based platform powered by Google Earth within 24 hours, with the aim of supporting the decision-making process of disaster prevention and mitigation. A detailed validation was made afterwards by comparing the map with in situ data of the water levels at 17 stations. The results demonstrate the feasibility of rapidly responding to a typhoon-induced flood with a spaceborne SAR-derived map of inundated areas. A standard operating procedure was derived from this work and followed by the Water Hazard Mitigation Center of the Water Resources Agency, Taiwan, in subsequent typhoon seasons, such as Typhoon Trami (August, 2013) and Typhoon Soudelor (August, 2015). Full article
(This article belongs to the Special Issue Remote Sensing in Flood Monitoring and Management)
►▼ Show Figures

Graphical abstract

Graphical abstract
Full article ">Figure 1
<p>Information about Typhoon Soulik and its moving tracks forecasted at 20:00, 11 July. Two COSMO-SkyMed 1 radar satellite images with a spatial resolution of 30 m and a 100 × 100 km<sup>2</sup> coverage area were scheduled at 5:46:59, 13 July [<a href="#B26-remotesensing-07-11954" class="html-bibr">26</a>]. The region of the image taken is shown by the green box. The different colors of the path lines show the forecasted typhoon moving tracks provided by different forecasting models, such as the Weather Research and Forecasting Model (WRF) and the Fifth-Generation Penn State/ National Center for Atmospheric Research (NCAR) Mesoscale Model (MM5).</p> Full article ">Figure 2
<p>(<b>a</b>) Formosat-2 true color satellite image of I-Lan County, which is located in the northwestern part of Taiwan. The region in the green triangle is Lanyang Plain. (<b>b</b>) The region denoted as the purple square is covered by the SAR image. The region denoted as the yellow square is the same geographic position as the areas shown in <a href="#remotesensing-07-11954-f003" class="html-fig">Figure 3</a>.</p> Full article ">Figure 3
<p>The positions of water level stations in the study area, with the geographic position being the same as that indicated by the yellow square in <a href="#remotesensing-07-11954-f002" class="html-fig">Figure 2</a>.</p> Full article ">Figure 4
<p>The flowchart of flood extent detection using the Expert Synthetic Aperture Radar Imagery Waterbody Delineation System (ESARIWDS).</p> Full article ">Figure 5
<p>An example of the low ratio of the dark region in one scene. (<b>a</b>) Dark regions determined by ESARIWDS (green polygons). (<b>b</b>) Histogram of the current scene, which exhibits a pattern of a single peak.</p> Full article ">Figure 6
<p>An example of the high ratio of the dark region in one scene. (<b>a</b>) Dark regions determined by ESARIWDS (green polygons). (<b>b</b>) Histogram of the current scene, which exhibits a bimodal pattern.</p> Full article ">Figure 7
<p>An example of flood extent detection using ESARIWDS. (<b>a</b>) The most recent optical image of the same area taken by Formosat-2; (<b>b</b>) like-polarization histogram; (<b>c</b>) cross-polarization histogram; (<b>d</b>) dark regions in the SAR image; (<b>e</b>) overlaying the SAR image and the boundary of dark regions onto the corresponding DEM to visually examine the topographical relationship of each dark region.</p> Full article ">Figure 8
<p>The flowchart of inferring the flood depths by combining flooding patterns with the inundation potential maps produced by WRA from 2007 to 2010.</p> Full article ">Figure 9
<p>The close-up photo of the water level gauge and the inundated areas (red shaded polygons) interpreted from the SAR imagery of twelve stations: (<b>a</b>) GJL12, (<b>b</b>) LML2, (<b>c</b>) MFL1, (<b>d</b>)MFL2, (<b>e</b>) ISR1, (<b>f</b>) ISR3, (<b>g</b>) ISR4, (<b>h</b>) ISR6, (<b>i</b>) ISR7, (<b>j</b>) ISR8, (<b>k</b>) ISR9, (<b>l</b>) ISR10, the name and location of each station is defined and illustrated in <a href="#remotesensing-07-11954-f003" class="html-fig">Figure 3</a> that were fully in accordance with the interpretation results.</p> Full article ">Figure 10
<p>The close-up photo of water level gauge and the inundated areas (red shaded polygons) interpreted from the SAR imagery of five stations: (<b>a</b>) GJL1, (<b>b</b>) GJL3, (<b>c</b>) KXL1, (<b>d</b>)ISR2, (<b>e</b>) ISR5, the name and location of each station is defined and illustrated in <a href="#remotesensing-07-11954-f003" class="html-fig">Figure 3</a> that were not consist with the interpretation results. These five stations are all located in a mixed pixel near the boundary of the inundated area.</p> Full article ">Figure 11
<p>The flooded regions in I-Lan County, with a focus on farmland areas.</p> Full article ">Figure 12
<p>The flood extent map compared to different inundation potential maps. Yellow regions show the common areas between the flood extent map and the inundation potential map; green regions show the areas overestimated by the inundation potential map; and red regions show the areas underestimated by the inundation potential map. (<b>a</b>) Comparison of the flood extent map to 200 mm rainfall map in one day. (<b>b</b>) Comparison of the flood extent map to the one-year recurrence period inundation potential map. (<b>c</b>) Comparison of the flood extent map to the two-year recurrence period inundation potential map.</p> Full article ">Figure 13
<p>The flood depth map derived in this research.</p> Full article ">
2011 KiB  
Article
The Impact of the Calibration Method on the Accuracy of Point Clouds Derived Using Unmanned Aerial Vehicle Multi-View Stereopsis
by Steve Harwin, Arko Lucieer and Jon Osborn
Remote Sens. 2015, 7(9), 11933-11953; https://doi.org/10.3390/rs70911933 - 17 Sep 2015
Cited by 155 | Viewed by 13493
Abstract
In unmanned aerial vehicle (UAV) photogrammetric surveys, the cameracan be pre-calibrated or can be calibrated "on-the-job" using structure-from-motion anda self-calibrating bundle adjustment. This study investigates the impact on mapping accuracyof UAV photogrammetric survey blocks, the bundle adjustment and the 3D reconstructionprocess under a [...] Read more.
In unmanned aerial vehicle (UAV) photogrammetric surveys, the cameracan be pre-calibrated or can be calibrated "on-the-job" using structure-from-motion anda self-calibrating bundle adjustment. This study investigates the impact on mapping accuracyof UAV photogrammetric survey blocks, the bundle adjustment and the 3D reconstructionprocess under a range of typical operating scenarios for centimetre-scale natural landformmapping (in this case, a coastal cliff). We demonstrate the sensitivity of the process tocalibration procedures and the need for careful accuracy assessment. For this investigation, vertical (nadir or near-nadir) and oblique photography were collected with 80%–90%overlap and with accurately-surveyed (σ ≤ 2 mm) and densely-distributed ground control.This allowed various scenarios to be tested and the impact on mapping accuracy to beassessed. This paper presents the results of that investigation and provides guidelines thatwill assist with operational decisions regarding camera calibration and ground control forUAV photogrammetry. The results indicate that the use of either a robust pre-calibration ora robust self-calibration results in accurate model creation from vertical-only photography,and additional oblique photography may improve the results. The results indicate thatif a dense array of high accuracy ground control points are deployed and the UAVphotography includes both vertical and oblique images, then either a pre-calibration or anon-the-job self-calibration will yield reliable models (pre-calibration RMSEXY = 7.1 mmand on-the-job self-calibration RMSEXY = 3.2 mm). When oblique photography was Remote Sens. 2015, 7 11934 excluded from the on-the-job self-calibration solution, the accuracy of the model deteriorated(by 3.3 mm horizontally and 4.7 mm vertically). When the accuracy of the ground controlwas then degraded to replicate typical operational practice (σ = 22 mm), the accuracyof the model further deteriorated (e.g., on-the-job self-calibration RMSEXY went from3.2–7.0 mm). Additionally, when the density of the ground control was reduced, the modelaccuracy also further deteriorated (e.g., on-the-job self-calibration RMSEXY went from7.0–7.3 mm). However, our results do indicate that loss of accuracy due to sparse groundcontrol can be mitigated by including oblique imagery. Full article
►▼ Show Figures

Graphical abstract

Graphical abstract
Full article ">Figure 1
<p>The study site is an eroding coastal scarp in a sheltered estuary in southeastern Tasmania, Australia. The map portrays the distribution of ground control and validation points.</p> Full article ">Figure 2
<p>A printed PhotoScan coded target as imaged in one of the UAV photographs from the nadir image set.</p> Full article ">Figure 3
<p>Calibration flight point cloud and camera network showing the 50 convergent camera stations and the 3D target array with some targets set up on tripods.</p> Full article ">Figure 4
<p>Box plots of the four calibration options for σ = 2 mm and with and without oblique imagery.</p> Full article ">Figure 5
<p>Pre-calibration <span class="html-italic">versus</span> on-the-job self-calibration scenario comparison using a strong control network (13 GCPs), with and without oblique imagery.</p> Full article ">Figure 6
<p>Pre-calibration <span class="html-italic">versus</span> on-the-job self-calibration scenario comparison using a sparse control network (five GCPs), with and without oblique imagery.</p> Full article ">
1228 KiB  
Article
Variability and climate change trend in vegetation phenology of recent decades in the Greater Khingan Mountain area, Northeastern China
by Huan Tang, Zhenwang Li, Zhiliang Zhu, Baorui Chen, Baohui Zhang and Xiaoping Xin
Remote Sens. 2015, 7(9), 11914-11932; https://doi.org/10.3390/rs70911914 - 16 Sep 2015
Cited by 64 | Viewed by 7286
Abstract
Vegetation phenology has been used in studies as an indicator of an ecosystem’s responses to climate change. Satellite remote sensing techniques can capture changes in vegetation greenness, which can be used to estimate vegetation phenology. In this study, a long-term vegetation phenology study [...] Read more.
Vegetation phenology has been used in studies as an indicator of an ecosystem’s responses to climate change. Satellite remote sensing techniques can capture changes in vegetation greenness, which can be used to estimate vegetation phenology. In this study, a long-term vegetation phenology study of the Greater Khingan Mountain area in Northeastern China was performed by using the Global Inventory Modeling and Mapping Studies (GIMMS) normalized difference vegetation index version 3 (NDVI3g) dataset from the years 1982–2012. After reconstructing the NDVI time series, the start date of the growing season (SOS), the end date of the growing season (EOS) and the length of the growing season (LOS) were extracted using a dynamic threshold method. The response of the variation in phenology with climatic factors was also analyzed. The results showed that the phenology in the study area changed significantly in the three decades between 1982 and 2012, including a 12.1-day increase in the entire region’s average LOS, a 3.3-day advance in the SOS and an 8.8-day delay in the EOS. However, differences existed between the steppe, forest and agricultural regions, with the LOSs of the steppe region, forest region and agricultural region increasing by 4.40 days, 10.42 days and 1.71 days, respectively, and a later EOS seemed to more strongly affect the extension of the growing season. Additionally, temperature and precipitation were closely correlated with the phenology variations. This study provides a useful understanding of the recent change in phenology and its variability in this high-latitude study area, and this study also details the responses of several ecosystems to climate change. Full article
►▼ Show Figures

Graphical abstract

Graphical abstract
Full article ">Figure 1
<p>Spatial patterns in the digital elevation model (DEM) of the Hulunber region.</p> Full article ">Figure 2
<p>Illustration of NDVI (normalized difference vegetation index) time series reconstruction using the HANTS algorithm.</p> Full article ">Figure 3
<p>Spatial distribution of phenological parameters in Hulunber during the period 1982–2012. (<b>a</b>) SOS; (<b>b</b>) EOS; (<b>c</b>) LOS.</p> Full article ">Figure 4
<p>Interannual phenology variability in the growing season (April–October) for the three main regions and the entire Hulunber area from 1982 to 2012. (<b>a</b>) SOS; (<b>b</b>) EOS; (<b>c</b>) LOS.</p> Full article ">Figure 5
<p>Spatial distributions of the variation ratio in Hulunber during 1982–2012. (<b>a</b>) SOS; (<b>b</b>) EOS; (<b>c</b>) LOS.</p> Full article ">Figure 6
<p>Relationships between phenological parameters and climatic variables for two regions of the study area. (<b>a</b>) SOS of the steppe region; (<b>b</b>) EOS of the steppe region; (<b>c</b>) SOS of the forest region; (<b>d</b>) EOS of the forest region.</p> Full article ">
Show export options expand_more Show export options expand_less
Select all
Export citation of selected articles as:
 
Displaying articles 1-73
Previous Issue
Next Issue
Back to TopTop