Hel Legers 2009
Hel Legers 2009
Hel Legers 2009
DOI 10.1007/s11269-009-9559-2
Abstract This paper demonstrates that combining spatial land surface data with
socio-economic analysis provides a number of indicators to strengthen decision mak-
ing in integrated water and environmental management. It provides a basis to: track
current water consumption in the Inkomati Basin in South-Africa; adjust irrigation
water management; select crop types; facilitate planning; estimate crop yields before
harvesting, and consequently to forecast market price development. Remote sensing
data and economic analysis can also be used to study the spatial distribution of
water consumption as an indicator of equity in access to water resources. It even
enables identification of farms that consume more irrigation water than formally
allocated. Finally, it provides a basis to assess the cost-effectiveness of various ways
to reduce agricultural water consumption. So, this approach is potentially useful for
determining water consumption, refining water allocation policies, and determining
the potential for water transfers through mechanisms such as water trading.
This paper is written in the framework of ‘A demonstration project in the Inkomati Basin’
(Soppe et al. 2006) funded by the ‘Partners for Water II’ program of the Dutch government.
P. J. G. J. Hellegers (B)
LEI-Wageningen UR, P.O. Box 29703, 2502 LS The Hague, The Netherlands
e-mail: petra.hellegers@wur.nl
R. Soppe · W. G. M. Bastiaanssen
WaterWatch, Generaal Foulkesweg 28, 6703 BS Wageningen, The Netherlands
C. J. Perry
Independent Consultant, 17 Storey Court, London NW8 8QX, UK
W. G. M. Bastiaanssen
Delft University of Technology, Stevinweg 1, 2600 GA Delft, The Netherlands
2420 P.J.G.J. Hellegers et al.
1 Introduction
weeds (Perry 2007). Remote sensing estimates the total consumed fraction ETact
and recent progress has been made to separate ET into evaporation E and tran-
spiration T, using the biomass production as a guideline (unpublished model from
WaterWatch). Of the non-consumed fraction, some will return to drains and rivers
as runoff or to an aquifer as recharge that can be exploited (the recoverable fraction).
The rest goes to saline sinks, contaminated groundwater or deep aquifers not capable
of economic exploitation (the non-recoverable fraction).
Agricultural production per unit of water is gradually becoming an accepted
framework for management of scarce river basin water resources (e.g. Kijne et al.
2003; Molden et al. 2007). The bio-physical Crop Water Productivity CWP (kg m−3 )
is calculated by dividing beneficial biomass by water consumed (Eq. 1). Here, the
Economic Water Productivity EWP ($ m−3 ) is calculated by multiplying beneficial
biomass by the market price, and deducting financial production costs of all inputs
(except water), all divided by water consumed (Eq. 2). In case of a negative EWP,
the financial costs of production exceed the gross production value (benefits).
EWPi = Pi∗ Yi − Bi∗ Yi − Ci / ETacti (2)
With
Yi Beneficial biomass or yield of crop i (kg ha−1 );
Pi Market price received for crop i ($ kg−1 );
Bi, Variable financial production cost of crop i ($ kg−1 );
Ci, Fixed financial production cost of crop i ($ ha−1 ).
This approach, which is known as the residual method (Young 2005), relies on
the fact that the value to a producer from producing a good is exhausted by the
summation of the values of the inputs required to produce it. If the value of one input
is unknown, it can be estimated by rearranging terms so that the unknown value is
a function of the price by quantity of the output, less the prices by quantities of all
known inputs, all divided by the quantity of the unknown input. This is the value of
water or the net return to water. Some earlier examples of gross return to water using
crop yield, market prices and total crop water consumption are given by Droogers
et al. (2003) and Soppe et al. (2006).
One basis for arguing for transfers of water among categories of uses—sectors,
crops, areas, and over time—to meet socio-political objectives will be considerably
strengthened if the CWP and EWP in existing and alternative uses is known. This
helps decision-makers to evaluate trade-offs between alternative water policies and
courses of social actions that change water use and the multiple services it provides.
The aim of this paper is to show that this method provides indicators to assist
decision making in integrated water management, by applying the method to the
Inkomati Basin in the eastern part of South Africa. South Africa’s National Water
Act (NWA, Act 36 of 1998) explicitly adopts Equity, Sustainability and Efficiency as
the principles for water management.
According to Dye and Versfeld (2007) the most fundamental shift in the way water
in the country is viewed has been to democratise the ownership of water, now seen as
a public good which can no longer be owned. All water users are required to register
2422 P.J.G.J. Hellegers et al.
and licence their use of water, to pay for that water, including river or groundwater
abstracted at their own expense and, in the case of plantation forestry, to pay for
the reduction in streamflow which is considered to result from the forestry activity.
Whilst ownership rights have been revoked, lawful users in 1998 have been allowed
to register that use and these volumes have been allocated to these users. Once
these volumes claimed have been verified as genuine use, such users may receive a
licence. Any new user of water since 1998 is obliged to apply for a water use licence,
which will only be granted if the resource is considered adequate to meet current and
projected demands. The only rights to water which now pertain are the right to Basic
Human Needs, and the rights of the Ecological Reserve. The Ecological Reserve
is the amount of water required for safeguarding and sustaining healthy stream and
river ecosystems. The national Department of Water Affairs (DWA) neatly captured
the spirit of the NWA in the slogan “Some, for all, forever”.
In the National Water Act there are currently 11 defined water uses. Forestry is
declared as a stream flow reducing activity, while rainfed sugarcane is for instance
not defined as a water use although it is known to use water.
In Section 2 spatial rainfall surplus patterns (rainfall minus ETact ) are analyzed—
which provides data essential for design and monitoring of advanced flexible water
allocation policies—as well as temporal rainfall surplus patterns—which can be used
to relate land use classes to streamflow patterns. Analyses of water stress and biomass
production can be used to estimate crop yield reduction before harvesting and the
implications for market prices. In Section 3 the spatial patterns of formally allocated
water rights and actual water consumption are compared, indicating the extent to
which legal entitlements are met and enforced. It will also be shown that the method
can be used as a compliance tool, to trace farms that consume more irrigation water
than formally allocated. In Section 4 procedures for assessing the cost-effectiveness
of various ways to reduce agricultural water consumption are presented.
2 Spatial and Temporal Variability in the Rainfall Surplus and Water Stress
The period analyzed is the year: July 2004–June 2005. The net rainfall during that
period—which is based on TRMM satellite data—was 563 mm, which is approxi-
mately 60 mm less than the 20-year average figure of 622 mm based on the climate
grids produced by the Climate and Research Unit of the University of East-Anglia.
Estimated ETact for the year1 was 694 mm, indicating that overall the basin con-
sumed more water than was replenished by rainfall (presumably from groundwater
withdrawals or reductions in soil moisture. This does not reveal anything about the
spatial or temporal variability of the consumption/supply balance.
On a spatial scale, rainfall surplus (rainfall minus ETact ) per 250 m pixel shows that
in the Inkomati catchment (Fig. 1), the western upper parts are net water producing
areas (shown in blue), while the eastern lower plains are net water consuming areas
(shown in red). These types of spatial rainfall surplus estimates were published
earlier for Sri Lanka by Bastiaanssen and Chandrapala (2003) and for Nepal by
Shilpaker et al. (2009). This provides data essential for proper water monitoring
and for design and monitoring of advanced flexible water allocation policies, such
as tradable water rights. The potential for transfers of water can be identified from
Fig. 1, but first we have to get the allocation system functioning well (i.e. basics right).
The method also allows estimation of ETact by land use class. Table 1 shows
water consumption by land use class in the Inkomti catchment, sorted by ETact in
millimeter. This is useful for estimating the likely impact of changes in land use. From
the table, it can be seen that although orchards, sugar cane and cash crops have the
highest ETact , they only consume 6% of the total volume of water consumed since
the area they cover within the Inkomati catchment is relatively small (4%). Pine and
eucalyptus plantations have both a lower ETact , but use in total about 12% of the
total volume of water consumed as they cover 11.5% of the area. It is also interesting
to note that Kruger Park is responsible for more than 25% of the water consumed in
the Inkomati catchment while accounting for less than 22% of the area—the park is
a slightly above average consumer of water.
Note that the rainfall and ETact data presented reflect annual average values:
seasonal effects are excluded. The total annual ETact of 694 mm can be broken down
into monthly values (see Fig. 2). When combined with monthly TRMM-based rainfall
2424 P.J.G.J. Hellegers et al.
120
70
20
Jul-04 Aug-04 Sep-04 Oct-04 Nov-04 Dec-04 Jan-05 Feb-05 Mar-05 Apr-05 May-05 Jun-05
-30
-80 Rainfall
ETact average
Rainfall surplus
-130 Accumulated rainfall surplus
Fig. 2 Monthly rainfall, average ETact and rainfall surplus (mm) in the Inkomati catchment
Indicators for Supporting Water Resources Management Decisions 2425
150
100
50
0
-50 Jul-04 Aug-04 Sep-04 Oct-04 Nov-04 Dec-04 Jan-05 Feb-05 Mar-05 Apr-05 May-05 Jun-05
-100
-150
-200
-250
-300 Rainfall
-350 ETact sugarcane
-400
Rainfall surplus
-450
Accumulated rainfall surplus
-500
Fig. 3 Monthly rainfall, ETact and rainfall surplus (mm) for sugar cane in the Inkomati catchment
measurements, monthly rainfall surplus can be calculated. The sign and magnitude
of the rainfall surplus indicates the impact of land use types either as net producers
of water, when local consumption is less than local rainfall, or net consumers of
water.
Overlaying rainfall minus ETact with detailed land use information, individual
crops can be extracted and analyzed. Figure 3, based on a total of 328,850 pixels (of
30 × 30 m) shows that sugarcane consumes more water than is supplied by rainfall
in almost every month (with a total annual ETact of 1,067 mm) and consequently
requires irrigation throughout the year. The contribution to ETact from i rainfall can
be estimated from the ET of a neighbouring rainfed pixel. The ET from irrigation
must originate from groundwater in the absence of surface water and is a good
indicator of net groundwater consumption in case no records are available (Ahmad
et al. 2005; Bastiaanssen and Hellegers 2007).
The distribution ETact each month can also be analyzed as a histogram (see Fig. 4).
March 2005 shows interesting results. This is a period that cane is harvested and
burned, and the histogram shows a much larger spread of values than in other
months, because of different harvest and burn times for different fields. Some fields
still have high values of ETact , since they are still growing, while other fields have low
ETact as a result of harvest or burning of the cane.
2.3 Spatial and Temporal Water Stress Based on ETpot and ETact per Pixel
Water stress occurs when actual evapotranspiration ETact 2 is lower than potential
evapotranspiration ETpot ,3 because the available water is limited. Spatial water stress
2 ET
act is affected by soil moisture conditions, vegetation development and meteorological
conditions.
3 ET is the evapotranspiration from a land surface whenthe crop is fully watered.
pot
2426 P.J.G.J. Hellegers et al.
0. 02 0. 02 0.02
normalized number of pixels (-)
0. 01 0. 01 0.01
0 0 0
0 50 100 150 200 250 0 50 100 150 200 250 0 50 100 150 200 250
Monthly ETa (mm) Monthly ETa (mm) Monthly ETa (mm)
0. 025 0. 025 0.025
0. 02 0. 02 0.02
normalized number of pixels (-)
0. 01 0. 01 0.01
0 0 0
0 50 100 150 200 250 0 50 100 150 200 250 0 50 100 150 200 250
Monthly ETa (mm) Monthly ETa (mm) Monthly ETa (mm)
0. 025 0. 025 0.025
0. 02 0. 02 0.02
normalized number of pixels (-)
0. 01 0. 01 0.01
0 0 0
0 50 100 150 200 250 0 50 100 150 200 250 0 50 100 150 200 250
Monthly ETa (mm) Monthly ETa (mm) Monthly ETa (mm)
0. 01 0. 01 0.01
0 0 0
0 50 100 15 0 200 250 0 50 100 150 200 250 0 50 100 150 200 250
Monthly ETa (mm) Monthly ETa (mm) Monthly ETa (mm)
Fig. 4 Variability in monthly ETact for sugar cane in the Inkomati catchment
maps such areas. Figure 5 shows that the forested area in the center have the highest
water stress. The map shows that water stress is least severe in the irrigated areas in
the lower part of the catchment, moderately stresses in the upper, rainfed catchment,
and most stressed in the central catchment.
The relationship between water stress, yield, irrigation, and water productivity is
complex. The basic relationship between biomass and transpiration is linear for a
crop that is not otherwise constrained (by nutrient deficiency, pest attack or other
constraints) (Steduto et al. 2007). However, several factors complicate this linear
Indicators for Supporting Water Resources Management Decisions 2427
simplicity: first, deliberate water stressing during the vegetative stage for some field
crops has only a small effect on grain yield so that the productivity of water can be en-
hanced.4 Further, stressing at certain times in the season can induce a better rooting
system so that more effective use of residual soil moisture is induced—so that the pro-
ductivity of irrigation water increases because total production reflects consumption
of both irrigation and residual moisture. These scenarios, under non-limiting con-
ditions in respect of other inputs, do not deviate substantially from the fundamental
linear relationship between crop water consumption and yield (Perry et al. 2009).
More significant variations in CWP result from variations in crop husbandry:
seed selection, planting density, land preparation, fertilization, and pesticide use
all contribute to determining the potential yield a crop can achieve if water is
adequate. The linear relationship alluded to above is the production frontier—and
many producers, for good reasons, or just by chance—will fall short of that frontier.
Figure 6, which is based on estimates of ETact and yield for over 300,000 sugar cane
pixels (236,595 pixels for commercial and 92,255 pixels for emerging farming), shows
that water maximum yield is achieved at about 1,190 mm. It also shows that many
farmers are (far) below the production frontier. Mapping of ETact against yield allows
identification of low and high productivity areas for further investigation.
Variations in water productivity result from a number of factors: management,
random events that cannot be controlled, and the natural productivity of the
farm resources. Poor water management (excessive deliveries cause an increase in
non-productive evaporation from wet soil, or inhibit growth due to waterlogging;
4 Some specialist crops—such as grapes—benefit positively from stress in terms of sugar content and
fruit quality, but are not of importance in this analysis.
2428 P.J.G.J. Hellegers et al.
Fig. 6 Sugar production per pixel as a function of actual water consumption for commercial (blue
cloud) and emerging farming (red cloud). Water consumption below 500 mm and sugar production
below 3,500 kg/ha are not taken into consideration. Source: Hellegers et al. (2009)
prolonged deficits, especially at critical growth stages) limit potential yield. Similarly,
seed selection and seeding rates, provision of nutrients and pest protection all affect
potential yield and water productivity. Severe droughts, storms, or pest attacks will
randomly affect productivity, while poor soils, salinity, or high water tables, will
continuously limit potential. Consideration of these various factors in relation to the
estimates of water productivity indicate that additional analysis is needed to establish
reasons for good or bad productivity levels: time series should show up random im-
pacts, while ground investigation, or relating productivity to mapped characteristics
of soils, water tables, etc may further explain continuous poor performance.
In this section the fraction of the total farm size to which water rights have been
allocated is used as an indicator of the pattern of authorized access to water. Further,
the volume of formally allocated water will be compared to actual consumed water
on commercial farms in two regions (Lomati and Komati) to trace trade of water
rights or detect unauthorized irrigation water use.
Data on farm and plot boundaries and formally allocated water rights per farm
were provided as GIS layers by the Department of Water Affairs and Forestry
(DWAF). Water rights (a) are based on a fraction of the farm, (b) are a specific
volume per hectare for that fraction, but (c) may be used on a greater or lesser area
than is authorized providing the abstraction volume implied by a and b is adhered to.
Figure 7 shows that the fraction of the total farm size to which water rights
have been allocated is larger closer to the river; probably due to easier access and
consequently higher historical usage. The allocated fraction varies considerably.
The figure also shows that farms in Lomati have a larger allocated fraction than
farms in Komati. The formally allocated volume per hectare is lower in Lomati
(8,500 m3 ha−1 year−1 ) than in Komati (9,950 m3 ha−1 year−1 ).
Indicators for Supporting Water Resources Management Decisions 2429
Fig. 7 Fraction of total farm size to which water rights have been allocated for commercial farming
in Lomati (on the left) and Komati (on the right)
Figure 8 shows that average water consumption per plot is generally higher
in Komati than in Lomati, which can be related to the intensity of sugarcane.
Comparison with Fig. 7 shows that those farms with the highest allocated fraction
are not necessarily the largest actual water consumers per plot. This may partly be
due to the difference in the allocated volume per hectare in Lomati and Komati—the
volume of formally allocated water is compared to the actual consumed water.
The volume of formally allocated water is compared to ETact in the left part of
Fig. 9 and to ETact minus rainfall (which is uniformly assumed to be 550 mm) in
the right part. Points on the 45 degree line correspond to the (unlikely) situation
where consumed (irrigation) water is equal to formally allocated (abstracted) water.
Points below the line can be explained by two factors: either the farmer did not
apply his full entitlement of water, or a significant proportion of the water applied
was not converted into ETact but instead went to drainage or deep percolation.
Trading of water rights would also be an explanation. Points above the 45 degree
line indicate farms where consumed irrigation water exceeds the formally allocated
volume. Excessive use suggests, if no water trading is taking place, that unauthorized
water consumption could be the case. In Komati there is more excessive use than
in Lomati. About 85% of the farms (n = 149) in Komati consumes more irrigation
water than they are entitled to. About half of the farms consume more than twice
the allocated volume and one-third uses even more than three times the allocated
volume. In Lomati about 35% of the farms (n = 88) consume more than they
are entitled to. A quarter is consuming more than twice the allocated volume and
2430 P.J.G.J. Hellegers et al.
Fig. 8 ETact per plot for commercial farming in Lomati (left part) and Komati (right part)
15% consumes more than three times the allocated volume. Obviously only a small
part of such excessive consumption can be explained by trading, because total use
substantially exceeds total authorised use.
12 4
Estimated consumed water from irrigation (Mm )
3
3
Actual consumed water (Mm3)
Komati Komati
Lomati Lomati
1:1 1:1
0 0
0 1 2 3 4 0 1 2 3 4
Entitled water (Mm3) Entitled water (Mm3)
Fig. 9 Comparison of volume entitlement and actual consumed water (left part) and comparison
of volume entitlement and consumed irrigation water assuming 550 mm of rainfall (right part) at
commercial farms in Komati and Lomati
Indicators for Supporting Water Resources Management Decisions 2431
According to the South-African Water Act, water has to be reserved for Basic
Human Needs and for the Ecological Reserve. It is, however, hard to quantify socio-
economic gains from achieving such socio-political objectives (see Hellegers 2006
and Hellegers and Perry 2006). It is therefore especially important to inform policy
makers of the costs of achieving such objectives. Cost-effectiveness curves—based on
the opportunity costs (foregone benefits to agriculture) of diverting water away from
agriculture—provide an indicator to show policy makers the cost of water diversions.
For the purposes of the analysis presented below, the caveats set out in Section 2
regarding the various causes for apparent poor production, and the need to collect
productivity information over time, are ignored. Correction and adjustment of
the data over time will lead to a revised distribution of performance, to which
the approaches defined below to assessing alternative means of reducing water
consumption can be applied.
Agricultural water consumption can be reduced in various ways. The four meth-
ods, which do not allow farmers to trade water, are described below and visually
illustrated in Appendix. Each row in Appendix represents one of the four methods
of reducing agricultural water consumption by 15%, 30% and 45%.
Method 1: Agricultural areas with lowest CWP are taken out of production first.
This cuts back the low-productive, left tail of the CWP histograms in
Appendix.
Method 2: Agricultural areas with lowest EWP are taken out of production first. It
is a reduction in economically low-performing areas—the left tail of the
EWP histograms in Appendix.
Method 3: Water use is rationed to a maximum volume per hectare—a quota. This
affects the largest ETact per hectare users most. Water use that exceeds
the quota can be either beneficial or non-beneficial consumption. It is
assumed that first the non-beneficial consumption is reduced by means
of improved management practices without biomass losses.
Method 4: Water is reduced proportionally (i.e. with an equal share of water
consumed). Again it is assumed that first non-beneficial consumption
is reduced, in this case by all water users. So if all farmers below the pro-
duction frontier shown in Fig. 6 can reduce water use without biomass
losses, agricultural water consumption can be reduced substantially with
zero costs. Whether this is a realistic assumption in reality depends on
the factors that determine the differences in productivity.
Figure 10 shows that method 1 and 2 have similar effects. Reducing 50% of water
consumption reduces about 45% of the biomass production and 40% of the net
production value of each crop (except for bananas). The results of method 1 and
2 would be different if applied to all crops (instead of to specific crops), as under
method 1 emerging sugar cane would be reduced first, while under method 2 bananas
would be reduced first.
Methods 1 and 2 have different effects for bananas. This is due to the fact that for
low CWP values of bananas there is no strong relationship (correlation) between the
CWP and EWP. This can be explained by the high fixed production costs of bananas,
which is a relatively heavy burden for low-productive users. There is a stronger
0. 5 technical efficient (method 1)
0.5 technical efficient (method 1) Sugar commercial 0. 5 technical efficient (method 1)
Banana Sugar emerging
economic efficient (method 2) economic efficient (method 2) economic efficient (method 2)
2432
largest users first (method 3) largest users first (method 3) largest users first (method 3)
socially equitable (method 4) socially equitable (method 4) socially equitable (method 4)
0. 4 0.4 0. 4
0. 3 0.3 0. 3
0. 2 0.2 0. 2
0 0 0
0 0. 1 0. 2 0.3 0. 4 0.5 0 0.1 0.2 0.3 0.4 0.5 0 0.1 0. 2 0. 3 0. 4 0.5
Fraction decrease in ETa (-) Fraction decrease in ETa (-) Fraction decrease in ETa (-)
0.4 0. 4 0. 4
0.3 0. 3 0. 3
0.2 0. 2 0. 2
0.1 0. 1 0. 1
0 0 0
0 0.1 0.2 0.3 0.4 0.5 0 0.1 0. 2 0.3 0. 4 0.5 0 0.1 0.2 0.3 0.4 0.5
Fraction decrease in ETa (-) Fraction decrease in ETa (-) Fraction decrease in ETa (-)
P.J.G.J. Hellegers et al.
Indicators for Supporting Water Resources Management Decisions 2433
Fig. 10 Fraction decrease in total biomass (top row) and in total net production value (bottom row)
as a function of the fraction decrease in actual water consumption for the four methods of reducing
water consumption for bananas, commercial and emerging sugar (left to right)
relationship between the CWP and EWP for sugar cane (see Fig. 3 of Hellegers et al.
2009).
Method 4 is more cost-effective than method 3. This might be counterintuitive,
but in our analysis method 4 provides all (not just the largest) water users incentives
to improve water management practices. This is due to the assumption that non-
beneficial consumption can be reduced by all water users without any costs in the
case of method 4, while in the case of method 3 it is assumed that non-beneficial
consumption is only reduced by those that exceed the quota. The two graphs
eventually converge, since under method 3 an increasing number of water users also
receive this incentive. It is a convex relationship, because initially most users can
meet the quota by reducing the non-beneficial consumed fraction, but eventually an
increasing number of users have to reduce the beneficial consumed fraction and costs
of biomass losses will increase at an increasing rate.
From this analysis it becomes clear that reducing all water use proportionally is the
most cost effective way of releasing small volumes of water, as it provides incentives
to all users to reduce their non-beneficial consumed fraction while maintaining their
biomass production. Taking low-performing agricultural areas (with a low CWP)
out of production is, however, a cheaper way of releasing larger volumes of water.
For bananas the most cost-effective way of reducing water consumption is to take
agricultural areas with negative EWP out of production.
5 Conclusions
In this paper the usefulness of remote sensing estimates of ETact , ETpot , biomass pro-
duction and rainfall in combination with economic analysis has been demonstrated
in the Inkomati Basin in the eastern part of South Africa. This method is widely
applicable to water resource policy-making and monitoring. It shows where water is
available and where it is consumed by what crops. It shows where there are surpluses,
and how these relate to land use. It also provides a basis to track current consumption
and potentially extend the definitions of water user to different categories of users
in South Africa. While a strong Water Act is in place, implementation of the water
policy envisaged remains a challenge—part of which relates to the availability of data.
Remote sensing is a useful source of important basic data—where water is being
consumed, how this relates to assigned water rights, how land use affects water
availability, and where water is being most and least productively used.
Finally, it has been demonstrated that the combination of remote sensing and
economic analysis allows comparison of the opportunity costs of allocating water
in a socially desirable way. This will strengthen the basis for arguments to transfer
water between users. This approach also provides insights into the cost-effectiveness
of various ways to divert water away from agriculture.
Open Access This article is distributed under the terms of the Creative Commons Attribution
Noncommercial License which permits any noncommercial use, distribution, and reproduction in
any medium, provided the original author(s) and source are credited.
2434 P.J.G.J. Hellegers et al.
Each row represents one of the methods of ET reduction as used in the study.
0 0 0
0 1 2 3 4 5 0 1 2 3 4 5 0 1 2 3 4 5
CWP Fresh Banana (kg/m3) CWP Fresh Banana (kg/m3) CWP Fresh Banana (kg/m3)
0 0 0
-3 -2 -1 0 1 2 3 -3 -2 -1 0 1 2 3 -3 -2 -1 0 1 2 3
ewp (R/m3) ewp (R/m3) ewp (R/m3)
0 0 0
400 800 1200 1600 400 800 1200 1600 400 800 1200 1600
ETa (mm) ETa (mm) ETa (mm)
0 0 0
340 740 1140 280 480 680 880 1080 220 420 620 820
ETa (mm) ETa (mm) ETa (mm)
References
Ahmad MD, Bastiaanssen WGM, Feddes RA (2005) A new technique to estimate net groundwater
use across large irrigated areas by combining remote sensing and water balance approaches,
Rechna Doab, Pakistan. J Hydrogeol J 13:653–664
Allen RG, Tasumi M, Morse A, Trezza R (2005) A Landsat-based energy balance and evapotran-
spiration model in Western US water rights regulation and planning. Irrig Drain Syst 19:251–268
Allen RG, Tasumi M, Morse A, Trezza R, Wright JL, Bastiaanssen W, Kramber W, Lorite I, Robison
CW (2007) Satellite-based energy balance for mapping evapotranspiration with internalized
calibration (METRIC)—applications. ASCE J Irrig Drain Eng 133(4):395–406
Barret EC (1988) Precipitation monitoring by satellites. Remote sensing for studies of global en-
vironmental changes. ISPRA Courses, RS/88/10, Commission of European communities, Joint
Research Center, Ispra Establishment, Italy
Barrett EC, Beaumont MJ (1994) Satellite rainfall monitoring: an overview. Remote Sens Rev 11:
23–48
Bastiaanssen WGM, Ali S (2003) A new crop yield forecasting model based on satellite measure-
ments applied across the Indus Basin, Pakistan. Agric Ecosyst Environ 94(3):321–340
Bastiaanssen WGM, Chandrapala L (2003) Water balance variability across Sri Lanka for assessing
agricultural and environmental water use. Agric Water Manag 58(2):171–192
Bastiaanssen WGM, Hellegers PJGJ (2007) Satellite measurements to assess and charge for ground-
water abstraction. In: Dinar A, Abdel Dayem S, Agwe J (eds) The role of technology and
institutions in the cost recovery of irrigation and drainage projects. Agriculture and Rural
Development discussion paper 33. The World Bank, Washington
Bastiaanssen WGM, Ahmed MD, Chemin Y (2002) Satellite surveillance of evaporative depletion
across the Indus Basin. Water Resour Res 38:1273–1282
Bastiaanssen WGM, Noordman EJM, Pelgrum H, Davids G, Allen RG (2005) SEBAL for spatially
distributed ET under actual management and growing conditions. ASCE J Irrig Drain Eng
131:85–93
Casa R, Rossi M, Sappa G, Trotta A (2008) Assessing crop water demand by remote sensing and
GIS for the Pontina Plain Central Italy. Water Resour Manag 23(9):1685–1712
Chowdary VM, Ramakrishnan D, Srivastava YK, Chandran V, Jeyaram A (2008) Integrated water
resource development plan for sustainable management of Mayurakshi watershed, India using
Remote Sensing and GIS. Water Resour Manag 23(8):1581–1602
Courault D, Seguin B, Olioso A (2005) Review on estimation of evapotranspiration from remote
sensing data: from empirical to numerical modeling approaches. Irrig Drain Syst 19:223–249
Droogers P, Malik RS, Kroes JG, Bastiaanssen WGM, van Dam JC (2003) Future water management
in Sirsa district: options to improve water productivity. In: van Dam JC, Malik RS (eds) Water
productivity of irrigated crops in Sirsa district, India. WUR and CCS Haryana Agricultural
University, Haryana
Dye P, Versfeld D (2007) Managing the hydrological impacts of South African plantation forests: an
overview. For Ecol Manag 251:121–128
Hellegers PJGJ (2006) The role of economics in irrigation water management. Irrig Drain 55:157–163
Hellegers PJGJ, Perry CJ (2006) Can irrigation water use be guided by market forces? Theory and
practice. Int J Water Resour Dev 22(1):79–86 (special issue)
Hellegers PJGJ, Soppe R, Perry CJ, Bastiaanssen WGM (2009) Combining remote sensing and
economic analysis to support policy decisions that affect water productivity. Irrig Sci 27(3):
243–251
Huffman GJ, Adler RF, Bolvin DT, Gu G, Nelkin EJ, Bowman KP, Hong Y, Stocker EF, Wolff
DB (2007) The TRMM multi-satellite precipitation analysis: quasi-global, multi-year, combined-
sensor precipitation estimates at fine scale. J Hydrometeorol 8:38–55
Kalma JD, McVicar TR, McCabe MF (2008) Estimating land surface evaporation: a review of
methods using remotely sensed surface temperature data. Surv Geophys 29:421–469
Kidd C (2001) Satellite rainfall climatology: a review. Int J Climatol 21(9):1041–1066
Kijne JW, Barker R, Molden D (2003) Water productivity in agriculture: limits and opportunities
for improvement (Comprehensive Assessment of Water Management in Agriculture Series 1).
CABI, Wallingford
Kummerow C, Olson WS, Giglio L (1996) A simplified scheme for obtaining precipitation and
vertical hydrometeor profiles from passive microwave sensors. IEEE Trans Geosci Remote Sens
34:1213–1232
2436 P.J.G.J. Hellegers et al.
Kustas WP, Norman JM (1996) Use of remote sensing for evapotranspiration monitoring over land
surfaces. Hydrol Sci J 41(4):495–516
Molden D, Oweis T, Steduto P, Kijne JW, Hanjra MA, Bindraban PS (2007) Pathways for in-
creasing agricultural water productivity. In: Water for food, water for life: a comprehensive
assessment of water management in agriculture, chapter 7. Earthscan. http://www.iwmi.cgiar.org/
assessment/Publications/books.htm
Moran MS, Jackson RD (1991) Assessing the spatial distribution of evapotranspiration using re-
motely sensed inputs. J Environ Qual 20:725–737
Perry CJ (2007) Efficient irrigation; inefficient communication; flawed recommendations. Irrig Drain
56(4):367–378
Perry CJ, Steduto P, Allen RG, Burt CM (2009) Increasing productivity in irrigated agriculture:
agronomic constraints and hydrological realities. Agric Water Manag 96:1517–1524
Petty GW (1995) The status of satellite-based rainfall estimation over land. Remote Sensing Environ
51:125–137
Petty GW, Krajewski W (1996) Satellite rainfall estimation over land. Hydrol Sci J 41:433–451
Schuurmans JM, Bierkens MFP, Pebesma EJ Uijlenhoet R (2007) Automatic prediction of high
resolution daily rainfall fields for multiple extents: the potential of operational radar. J Hydrom-
eteorol 8(6):1204–1224
Shilpaker RL, Bastiaanssen WGM, Molden D (2009) Geoinformation procedures to assist water
accounting: a case study of the East Rapti river basin, Nepal. Kluwer, Dordrecht (in press)
Smith DM, Kniveton DR, Barrett EC (1998) A statistical modelling approach to passive microwave
rainfall retrieval. J Appl Meteorol 37(2):135–154
Soppe RWO, Hellegers PJGJ, Perry CJ, Boon D, Bastiaanssen WGM, de Wit M, Pelgrum H (2006)
Combining remote sensing and economic analysis to assess water productivity: a demonstration
project in the Inkomati Basin. LEI and WaterWatch Report, Wageningen
Steduto P, Hsiao TC, Fereres E (2007) On the conservative behavior of biomass water productivity.
Irrig Sci 25:189–207
Teixeira AH, Bastiaanssen WGM, Ahmad MD, Bos MG (2009) Reviewing SEBAL input parame-
ters for assessing evapotranspiration and water productivity for the Low-Middle Sao Francisco
River basin, Brazil. Part A: calibration and validation. Agric For Meteorol 149(3–4):462–476.
doi:10.1016/j.agrfor-met.2008.09.016
Verstraeten WW, Veroustraete F, Feyen J (2008) Assessment of evapotranspiration and soil mois-
ture content across different scales of observation. Sensors 8(1):70–117
Young R (2005) Determining the economic value of water: concepts and methods. Resource for the
Future, Washington.
Zwart SJ, Bastiaanssen WGM (2007) SEBAL for detecting spatial variation of water productivity
and scope for improvement in eight irrigated wheat systems. Agric Water Manag 89:287–296