Journal of African Earth Sciences 196 (2022) 104681

Contents lists available at ScienceDirect

Journal of African Earth Sciences

journal homepage: www.elsevier.com/locate/jafrearsci

Lateritic processes in Madagascar and the link with agricultural and

socioeconomic conditions
Jonathan D. Paul a, *, Anthony Radimilahy b, Raymond Randrianalijaona b, Tina Mulyakova c
a
Department of Earth Sciences, Royal Holloway, University of London, Egham, TW20 0EX, UK
b
Département des Sciences de La Terre, Université d’Antananarivo, Madagascar
c
Department of Earth, Atmospheric, and Planetary Sciences, Purdue University, West Lafayette, IN, USA

A R T I C L E I N F O A B S T R A C T

Keywords: Madagascar hosts a youthful landscape atop old bedrock geology. Observed rapid surface uplift – possibly a
Erosion consequence of sub-plate mantle plume activity – does not fit the concept of “The Red Island”, i.e. thick piles of
Laterite laterite, formed by slow chemical weathering that requires prolonged periods of tectonic quiescence. Here, we
Madagascar
document the results of a three-year field campaign in which we sought to interrogate and decouple the climatic
Poverty
Precipitation
and tectonic drivers behind laterite formation, and establish a tentative link to agricultural productivity, which
Uplift may reinforce local levels of poverty. We installed 45 rain gauges and merged their data with a satellite pre­
cipitation product to develop a new high-resolution climatology. We also collected 11 sediment samples from
which we established 10Be-derived erosion rates; and measured laterite/saprolite thickness at 60 localities. The
thickest lateritic profiles coincide with a narrow band of annual rainfall, suggesting that laterite evolution
processes co-evolved with climate changes. Laterite thickness correlates negatively with erosion rate, and
strongly positively with the headcount index of poverty. In the absence of agricultural technology, we propose a
direct link between the local prevalence of lateritic soil and poor agricultural productivity. We also suggest that
the spatial distribution of uplift/erosion and precipitation governs the distribution and thickness of laterite.
Uplift and erosion rates vary dramatically over small distances, restricting the development of thick laterite to
two small patches that straddle the central Hauts Plateaux.

1. Introduction lithosphere, driving rapid landscape evolution. A +30 mGal long

wavelength free-air gravity anomaly centred on the island, and thinner
1.1. Regional setting crust, calculated by inverting a suite of teleseismic receiver functions, in
the central Hauts Plateaux relative to the coast (37 km and 44 km,
The island of Madagascar has attracted much recent attention for its respectively), suggest that topography is maintained dynamically at
manifold unique attributes spanning various scientific disciplines. There long wavelengths (>1000 km; Roberts et al., 2012; Paul and Eakin,
are several links. For instance, rapid erosion rates, amongst the highest 2017). Lavas from the north of the island have 39Ar-40Ar ages of 11–12
in the world not solely resulting from human activity, and its isolation Ma; while Neogene-Quaternary alkaline basalts occur at several local­
since rifting from India completed ~84 Ma, have led to unusual speci­ ities, notably in the Ankaratra highlands of central Madagascar (Bar­
ation patterns and the highest proportion of endemic species (~90%) dintzeff et al., 2010). Roberts et al. (2012) inverted a set of 98
globally (e.g. Bourgeat and Petit, 1969; Chand and Subrahmanyam, longitudinal river profiles, which implied that Madagascar was uplifted
2003; Goodman and Benstead, 2005; Kusky et al., 2010). by 1–2 km at rates of 0.2–0.4 km Myr− 1 since 15 Ma. Similarly, Ste­
In common with Africa, whence it rifted in Jurassic times, phenson et al. (2021) presented a combination of geophysical,
Madagascar is surrounded by passive margins and has probably geomorphic, and geochemical observations – including new tempera­
remained relatively stationary since Eocene times (e.g. De Wit, 2003). ture and denudation histories from apatite fission-track and helium
Several lines of evidence have been marshalled to suggest the presence measurements – to argue for 0.5–1.6 km of regional uplift after ~30 Ma.
of an incipient mantle plume that impinges upon the base of the A youthful landscape contrasts dramatically against old bedrock

* Corresponding author.
E-mail address: jonathan.paul@rhul.ac.uk (J.D. Paul).

https://doi.org/10.1016/j.jafrearsci.2022.104681
Received 29 April 2022; Received in revised form 28 July 2022; Accepted 2 August 2022
Available online 6 August 2022
1464-343X/© 2022 The Authors. Published by Elsevier Ltd. This is an open access article under the CC BY license (http://creativecommons.org/licenses/by/4.0/).
J.D. Paul et al. Journal of African Earth Sciences 196 (2022) 104681

geology (Fig. 1a). The latter is dominated by a suite of Late Neo­ dominated by a deeply eroded Precambrian craton, which hosts the
proterozoic and Cambrian metamorphosed terranes that govern later “super-mature” stage of argillitisation, i.e. phyllosilicate assemblages of
fault orientation (e.g. Bertil and Regnoult, 1998). Drainage patterns and modest chemical and mineralogical variability, dominated by laterite
the largely immature, convex-upward shapes of river profiles have been (together with kaolin and bauxite: Dill, 2017). Although Beauvais et al.
affected by the same processes giving rise to deeply weathered pene­ (2008) used Ar–Ar techniques to date K–Mn oxyhydroxides in a west
plains (i.e., flat, pervasive, erosional surfaces), which have been African ore deposit, Madagascar’s laterites lack independent age con­
observed throughout Madagascar (Dixey, 1960; Bourgeat and Petit, straints. Indeed, variations in the structure of weathering profiles across
1969; Burke and Gunnell, 2008). These erosional surfaces are now cut the island have been comparatively poorly studied compared to coun­
by steep cliffs, by inverted teardrop-shaped erosional gullies (known terparts in India and South America (e.g. Gleeson et al., 2004). Estrade
locally as lavaka after the Malagasy for ‘little hole’), and by river et al. (2019) estimated that roughly 85–90% of Malagasy laterite profiles
channels (Wells and Andrimihaja, 1993; Cox et al., 2010). Lavaka usu­ are autochthonous, resulting from in-situ weathering of gneiss and schist
ally develop within the thick (up to 50 m) laterites and saprolites that (basement) bedrock. They identified six well-defined zones to a typical
have developed across the central highlands of the island (Cox et al., weathering profile: a thin, hardened lateritic duricrust layer; a
2009; Kusky et al., 2010). The removal of vegetation, seismicity, cli­ red-brown ferruginous zone, principally composed of secondary Fe ox­
matic aridification, and tectonic uplift probably contribute to the for­ ides and oxyhydroxides; saprolite, weathered bedrock in which the
mation, and growth, of lavaka (Cox et al., 2010). original fabric is largely retained; saprock, compact and physically
strong but slightly weathered rock; and finally the bedrock, which has
not been affected by weathering. By contrast, Berger et al. (2014)
1.2. Laterite studied a 2–3 m-thick lateritic sequence that developed from weathering
of an intrusive igneous rock (tonalite). They did not identify the pres­
Madagascar is colloquially known as "The Red Island" on account of ence of saprolite, rather proposing a two-stage alteration process
the widespread laterite. Laterite is a heavily leached, iron-rich primitive involving initial pervasive loss of sulphides along fissures and cracks,
tropical palaeosol, which typically forms at low latitudes in areas of low followed by a main stage of soil formation (i.e. actual laterization).
relief, tectonic quiescence, and high rainfall (e.g. Tardy, 1997). Lateri­ Beauvais and Chardon (2013) analysed relicts of three lateritic
zation (i.e. prolonged tropical chemical weathering) initially results in palaeo-land surfaces in west Africa, concluding that the region under­
saprolite, i.e. kaolinized rocks; laterite is chemically and mineralogically went low and homogeneous denudation (~2–20 m Myr− 1) over three
different, being more heavily weathered and hence higher in the regolith erosion periods (45–24, 24–11, and 11–0 Ma). They argued that the
profile (e.g. Zauyah et al., 2010). Storey et al. (1997) report that regional cratonic denudational regime was strictly weathering-limited,
Cretaceous basalt has, in some places, been reduced to thick (up to 60 m) and that post-Eocene wet-dry climatic cyclicity shaped successive
piles of saprolite and laterite. Braun et al. (2009) estimated a rate of denudation stages that gave rise to the west African sequence of stepped
saprolite formation (i.e. weathering) of ~15 m Myr− 1 in India, which is lateritic peneplains. Similarly, Dill et al. (2005) detail how pulses of
comparable with previous estimates elsewhere (e.g. ~10 m Myr− 1: uplift and weathering over ~10 Myr created a staircase morphology of
Tardy and Roquin, 1992). Indicative of long periods of tectonic quies­ four peneplains and lithofacies types in southern Malawi.
cence, thick lateritic mantles cap many of the peneplains in Madagascar Cox et al. (2009) analysed 10Be concentrations in lavaka, slope, and
and across sub-equatorial Africa and elsewhere (e.g. India and river sediment in central Madagascar, deriving Quaternary erosion rates
Australia). Dill (2017) notes the importance of parent rock type in of ~12 m Myr− 1, or around three orders of magnitude lower than rates
determining laterite composition and soil type; these rocks are inti­ reported elsewhere (e.g. Grieser, 1994; Ralison et al., 2008). A similar
mately linked to regional geodynamic setting. Central Madagascar is

Fig. 1. (a) Digital Elevation Model (DEM) of

Madagascar from 30 m Shuttle Radar
Topography Mission (SRTM) dataset. Grey
shading = regions of high slope (>10◦ ). Thin
grey line = 1 km elevation contour. Dark red
lines = location of faults interpreted from
SRTM dataset and Landsat imagery. Black
dashed line = border of basement outcrop.
(b) Mean precipitation, 2018–2020. Stars
and circles = rain gauges (this study and
WMO, 2020, respectively). (c) Background
= mean uplift rate since 2 Ma, based on the
surface uplift history of Roberts et al.
(2012). Coloured circles = 10Be-derived
erosion rates (this study and Cox et al., 2009;
Table A2). Thin black lines = drainage.

2
J.D. Paul et al. Journal of African Earth Sciences 196 (2022) 104681

technique was used to estimate erosion rates from quartz cobbles in approach, of active local involvement (under the rubric of “citizen sci­
laterite of ~2 and ~12 m Myr− 1 in the West African Craton and Congo ence”) in the installation of scientific equipment, has been successfully
Basin, respectively (Braucher et al., 1998). demonstrated elsewhere (e.g. Davids et al., 2019; Paul et al., 2020). Data
Lateritic soils (i.e. oxisols) have low nutrient (e.g. N, P, K) content, were recorded every minute on a HOBO RG3-M data logger, and were
with 90–100% of Fe, Al, and Mn oxides; they lack fertility due to this collected at six-month intervals up to April 2021. These data were then
chemical composition and a lower base-exchanging capacity (Dwevedi carefully scrutinised for spikes or NaN errors, before being binned into
et al., 2017). In regions such as Madagascar, the geographic distribution 15-min intervals.
of oxisols is governed by the presence of stable geomorphic surfaces and Satellite estimations of precipitation, such as the various products of
the underlying rock type (itself largely a function of geological setting: the Global Precipitation Mission (GPM: Hou et al., 2014), are useful in
Dill, 2017). Thick oxisol profiles often form over easily weathered basic complementing sparse rain gauge networks. However, these products
rock, and typically are of low bulk density, high permeability, and employ smoothing algorithms that can result in major underestimation
roughly uniform particle size distribution with depth (e.g. Buol and of extremes (Manz et al., 2016). We therefore generate a new map of
Eswaran, 1999). Their friability, low water-holding capacity, and con­ satellite-rain gauge climatologies (i.e. averaged annual precipitation) by
centration of Fe and Al oxides have led to “slash-and-burn” subsistence merging gauge and satellite data using inverse distance weighted
agricultural practices in many areas such as Madagascar that lack interpolation (Manz et al., 2016). The GPM 2ADPR product (raw orbital
modern agronomic technologies (Reardon and Vosti, 1995; Buol and data) is used after attenuation corrections and regridding to obtain
Eswaran, 1999). More advanced economies have allowed farmers to rainfall rate estimations. We augmented our rain gauge data with gauge
mitigate these physiochemical disadvantages via judicious use of lime, time series from the 25 World Meteorological Organisation-maintained
fertiliser, and mechanisation, e.g. in parts of Brazil and eastern India (e. gauges in Madagascar (WMO, 2020), which were extracted and refor­
g. Brahmachari et al., 2019). matted into 15-min bins before the merging procedure. This procedure
Vagen et al. (2006) introduced a visible-near-infrared reflectance was also carried out for >400 rain gauge records in Africa (Paul, 2021).
spectroscopic technique for rapid characterisation of soils (including Where outcrop was present, we measured the thickness of the
oxisols) across Madagascar, in terms of key soil properties such as soil laterite + saprolite profiles directly in three representative locations
organic carbon, cation exchange capacity, and total nitrogen. While spaced ~10 m apart, before averaging. We used an RS-PRO ILDM-150H
research data on Malagasy soil quality are scarce (in spite of the coun­ laser measuring device to triangulate the distance between the surface
try’s significant soil degradation and low agricultural productivity), and base of each profile (i.e. bedrock). In the absence of outcrop, we
they defined 10 key agronomic indicators of soil fertility across recovered three core samples (from surface to bedrock) over an area of
Madagascar, including pH, organic carbon, and silt and clay content. ~100 m2, from 10 cm diameter core cutters suitable for lateritic soil (e.
~79% of Madagascar’s soils were defined as oxisols, defined in this case g. Akpokodje and Hudec, 1992). Table A1 indicates that, where laterite
as having low pH, low contents of exchangeable ions, prominent ab­ was present, weathering profile thickness was evaluated at 32 locations
sorption features associated with a high content of clay minerals (at from outcrop, and six locations from borehole recovery.
1400, 1900, and 2200 nm), and very low organic carbon content Weathering profiles in Madagascar are highly complex, even though
(minimum = 3.3 g kg− 1). Sixteen “land degradation hotspots” correlated the bedrock across the central Hauts Plateaux is typically homogeneous
strongly with the presence of a thick oxisol horizon (Vagen et al., 2006). yet strongly deformed gneiss. The weathering transition is progressive,
The purpose of this paper is therefore to interrogate possible links resulting in rather subtle changes in colour and grain size (relative to
between the spatial extent of laterite, drivers behind its formation, and other regions such as South America: Gleeson et al., 2004). A repre­
agricultural productivity. Madagascar is an excellent natural laboratory sentative profile is shown in Fig. 2. Although our analyses consider the
for this exercise, since (i) uplift, erosion, and precipitation rates exhibit a total height of the entire laterite + saprolite column, we nevertheless
striking spatial variance; (ii) the centre of the island is renowned for its carefully defined their boundary by considering changes in colour,
thick piles of laterite and saprolite; and (iii) its relative isolation and self- structure, and rheology. Fig. 2 shows that a typical lateritic profile be­
containment enhance the robustness of any correlations that might be gins at the top with a hard lateritic duricrust (1–5 cm) followed by red
discovered. We tackled our task over a six-month field campaign in and yellow laterite zones, which are generally homogeneous and
2017, in which specific objectives were to: composed of iron oxyhydroxides (haematite and goethite). Below this
horizon is a saprolitic zone, typically weakly bedded, lighter in colour,
• Obtain an accurate spatial representation of mean annual rainfall. and much softer. This saprolite represents an intermediate stage of
Install a country-wide rain gauge network, such that their data might parent rock alteration. In many localities the saprolite may be divided
be used to correct and improve a remotely sensed precipitation into earthy and rocky units, the latter inheriting structure and colour of
product; the basement bedrock. Samples were collected (Table A2) to charac­
• Obtain a spatially representative distribution of laterite/saprolite terise the in situ-produced 10Be content of surficial material include
regolith thicknesses; and sediment collected from river sand, lavaka, and river terrace colluvium.
• Calculate erosion rates for a greater number of locations beyond We followed the sampling and 10Be analysis procedure of Cox et al.
those of Cox et al. (2009). Collect samples of lavaka and/or river (2009). Briefly, we sampled by digging ~10 cm-deep trenches in the
sediment from which in situ-produced 10Be analyses might be centres of lavaka or river terraces, and collected well-sorted, traction-­
conducted. structured fluvial sand. Isolation of sand-sized quartz followed the
method of Kohl and Nishiizumi (1992). We extracted Be at the Univer­
2. Methods sity of Vermont; data were calibrated using the program CALIB (Stuiver
et al., 2006). Samples were measured at the Purdue Rare Isotope Mea­
We conducted our field campaign during Madagascar’s dry and mild surement Laboratory, and normalised using standards developed by
winter (April to October) 2017. We sought as many localities as possible Nishiizumi et al. (2007), assuming a10Be half-life of 1.39 Myr. We
that would satisfy the co-location of a rain gauge installation, laterite reduced the data by using accepted interpretive models (Granger et al.,
core, and sediment sample for 10Be analysis. We installed Davis 0.2 mm 1996), incorporating a high-latitude, sea-level production rate of 5.3
automatic tipping-bucket rain gauges in secure locations such as atoms 10Be g− 1 yr− 1 (Lal, 1988).
schools, churches, and police stations, away from vegetation or signifi­ We chose one variant of the most widely employed means of
cant slopes. Members of the local community were actively involved in assessing poverty in any society, i.e. a headcount of those considered
the installations and maintenance, while we offered community “poor”, i.e. those whose incomes fall short of a stipulated poverty line (e.
question-and-answer sessions on our work where appropriate. This g. See and Subramanian, 2002). The poverty headcount index, also

3
J.D. Paul et al. Journal of African Earth Sciences 196 (2022) 104681

coincidence between surface uplift and the distribution of lavaka and

Neogene volcanism (black dashed line and red polygons respectively on
Fig. 1c). Uplift rates are greatest in the centre of Madagascar, and over a
patch of the extreme north of the island. 10Be-derived erosion rates are
of the same order of magnitude (between ~3 and 45 m Myr− 1) and are
greatest in the central Hauts Plateaux, decreasing appreciably to the
south.
When a cubic spline is applied to the laterite thickness measure­
ments, two patches of thick lateritic regolith emerge, straddling the
central Hauts Plateau, with greater thicknesses in the southern patch
(Fig. 3a). This pattern mirrors that delineated by detailed on-the-ground
mapping (Wells and Andrimihaja, 1993). Moreover, the most productive
rice-growing areas, as documented by the Food and Agriculture Orga­
nization of the United Nations (2016), lie outside these two patches of
laterite, being mainly distributed along the coast or spanning the
island’s centre (Fig. 3a). Two regions of strikingly high levels of socio­
economic poverty are also observed, notably in the northern Hauts
Plateaux, and also a more diffuse area towards the south. These regions
are separated by a comparatively more prosperous central zone,
although several of the smallest districts here have the highest poverty
level (in the capital city, Antananarivo).
We next analysed the relationship between laterite + saprolite
regolith thickness and (a) mean annual precipitation; (b) local erosion
rate; and (c) poverty levels (Fig. 4). It is apparent that the greatest
thicknesses of laterite occur within a discrete band of annual precipi­
tation (~500–900 mm). A strong negative relationship exists between
erosion rate and laterite thickness less than 10 m. Although the greatest
weathering profile thickness (23.5 m) corresponds to the lowest calcu­
lated erosion rate (2.5 ± 0.5 m Myr− 1, at Ihosy in the barren southern
uplands: Table A1), the relationship between erosion rate and laterite
thickness is weaker for weathering profiles >10 m thick. Finally,
although the poverty headcount index is averaged at the district level,
there is a positive correlation with laterite thickness, albeit slightly
weaker than (b): R2 = 0.88; two-tailed p value = 0.001. There are some
anomalies, however; these relate to predominantly poor and highly
dense urban districts, such as Antananarivo in the centre (laterite
thickness ~0.4 m) and Toliara on the coast (laterite/saprolite regolith
absent).

4. Discussion
Fig. 2. Typical weathering profile of central Malagasy Hauts Plateaux,
demonstrating colour changes and laterite/saprolite interface. Location = The strength of the negative correlation between laterite thickness
Ampandrano (~100 km NW of Antananarivo); locality ID = 13 in Table A1
and 10Be-derived erosion rates underscores the importance of long pe­
(profile thickness = 10.1 m).
riods of tectonic quiescence for lateritization (cf. Beauvais and Chardon,
2013). However, the wide range of 10Be accumulation in our samples
known as FGT(0) to denote its belonging to the Foster-Greer-Thorbecke suggests a similarly wide range of denudation rates over geological
set of poverty indices (Foster et al., 1984), has the advantages of timescales, depending on tectonic setting: the greatest erosion rates
simplicity, additive decomposability, and subgroup consistency, which correspond to rapidly uplifting areas (most notably the central Hauts
allow socioeconomic poverty to be evaluated across all population Plateaux), recording rapid incision of lavaka and other gullies into
sub-groups in a self-consistent manner. We use the most recent and less-dosed material at depth. It must also be recognised that while
smallest (district-) level estimations of headcount index, compiled by the laterite formation is a geologically long process that could reflect varied
World Bank and NASA (SEDAC, 2021). climatic conditions, the current spatial distribution of poverty reflects
far more recent changes (i.e. of a human timescale). Moreover, we
3. Results acknowledge that, with the addition of suitable technologies, agricul­
tural productivity could be enhanced in areas of thick lateritic soil.
We were able to measure laterite + saprolite regolith thickness at 60 Interestingly, for Madagascar at least, there exists a discrete band of
localities across the country, install 45 automatic tipping-bucket rain precipitation in which laterite/saprolite weathering profiles are thick­
gauges, and obtain 11 samples for 10Be analysis to derive erosion rates est. Indeed, many areas of Madagascar that experience the highest
(Table A1). Nine further rain gauges were installed in 2017, but were annual rainfall and erosion, especially the northwest coast, are notable
lost (due to vandalism, theft, poor weather, or the action of livestock) for the complete absence of laterite (as well as oxisols: Vagen et al.,
before data could be recorded. 2006). A comparison of present-day climatological precipitation and
We combined rain gauge data with a satellite precipitation product poverty headcount index (Fig. 5) suggests that Madagascar’s poorest
to generate a high-resolution (~5 km) climatology for Madagascar regions are also some of its driest (although this relationship breaks
(Fig. 1b). Annual rainfall is greatest in the northwest (>1500 mm) and down under ~400 mm yr− 1, as areas that receive this little rainfall are
along the eastern coast, decreasing to a minimum of negligible annual located along the relatively prosperous and well-connected southwest
rainfall along a patch of southwestern coastline. There is a close spatial coastline).

4
J.D. Paul et al. Journal of African Earth Sciences 196 (2022) 104681

Fig. 3. (a) Thickness of laterite and saprolite layer across Madagascar, based on 65 core measurements (locations = small circles). Dashed black line = border of
“lateritic soil area” of Wells and Andrimihaja (1993). Green line delineates most productive rice-growing regions (Food and Agriculture Organization of the United
Nations, 2016). (b) Headcount index of poverty (FGT(0)), i.e. the proportion of the population of a district whose welfare falls below the poverty line (SEDAC, 2021).

Yet there is a mismatch between this present-day snapshot of rainfall several lines of evidence suggest the island is a topographic “swell”
patterns and the relatively long timescales of laterite formation (e.g. dominated by rapid uplift since at least ~15 Ma. These include (i)
1–10 Myr: Beauvais et al., 2008), which implies that rainfall, and the convex-upwards, immature river profiles and K-Ar-dated Neogene
processes that govern the formation of laterite, have evolved together. volcanism in the centre and north of the island (Bardintzeff et al., 2010;
Widdowson (1997) suggests that the periodicity of heavy rainfall (such Roberts et al., 2012); (ii); receiver function estimates of crustal thickness
as the timing of a Monsoon system) could play a critical role in the that suggest sub-plate topographic support (Paul and Eakin, 2017); (iii)
development of laterite profiles, in terms of governing soil strength, flat-lying Upper Cretaceous-Paleogene marine limestones that crop out
friability, and infiltration capacity. Buol and Eswaran (1999) similarly at elevations of 100 s m along the western seaboard, and emergent
note that oxisols preferentially form over polycyclic sediments that have Quaternary coral-rich terraces than rim the coastline (Stephenson et al.,
undergone several repeated stages of weathering prior to deposition. 2019); and (iv) inversions of apatite fission-track and helium measure­
However, other tectonic factors such as changes in surface uplift rate or ments, suggesting that up to 1.6 km of regional uplift occurred during
slope could also be of greater importance, and must be considered Neogene times (Stephenson et al., 2021). So can the presence of thick
together with climatic changes when considering the evolution of lateritic weathering profiles on top of old basement rocks satisfy these
laterite formation processes. Dill (2017) emphasises the importance of observations?
geodynamic setting in governing the distribution of different rock types, On the basis of the distribution of laterite thicknesses and calculated
which control the composition and thickness of laterite weathering erosion rates, we suggest that uplift rates vary significantly across the
profiles (including oxisols). Beauvais and Chardon (2013) refer to the interior of the island, rather than a simple “bullseye” focused on the
primacy of steady-state, weathering-limited denudation regimes in island’s centre, described by long wavelength free-air gravity anomalies
generating thick piles of laterite across west Africa, most of which has (a crude proxy for sub-plate mantle convection). This agrees with a
experienced low rates of uplift in Neogene times (e.g. Paul et al., 2014). notable gradient in the elevation of emergent marine fauna on the
Madagascar appears to be part of the “basin-and-swell” morphology of island’s northern coastline: an ancient (~130 ka) coral reef terrace de­
Africa (e.g. Holmes, 1944; Burke and Gunnell, 2008); specifically, creases in elevation by ~7 m over a distance of 80 km, suggestive of

5
J.D. Paul et al. Journal of African Earth Sciences 196 (2022) 104681

of laterite and oxisols. Contrary to Cox et al. (2009), this spatial coin­
cidence implies that the Malagasy landscape has not been “surprisingly
stable” on timescales of “hundreds of millennia”; at least not in the
central Hauts Plateaux.
Two regions that straddle the island’s centre, where lateritic regolith
is thick, correspond (even to the naked eye: Fig. 3) extremely well to the
distribution of Madagascar’s poorest districts. These are largely upland,
rural areas, where subsistence farming is common; in the southern half
of the island, this mainly takes the form of low-intensity livestock
farming (e.g. von Heland and Folke, 2014). However, it must be rec­
ognised that poverty reflects a convolution of manifold other factors,
including proximity to transport, access to goods from major population
centres and the coast, education levels, and settlement altitude, inter
alia. The most profitable crop in Madagascar is rice: in the centre of the
island, ancestors of the dominant Merina tribe have managed to produce
three harvests per year instead of the traditional two (De Laulanié,
2011). However, rice production is extremely expensive in areas that
lack well-irrigated and nitrogen-rich soils to form paddies. Diversifica­
tion of agricultural practices could improve nitrogen content in soils; for
instance, crops such as cassava and other legumes are
low-water-demand and offer atmospheric nitrogen fixation. Therefore,
across Madagascar, many farmers do not have the means to improve the
quality of oxisols, even by simple techniques such as the addition of lime
(De Laulanié, 2011).

5. Conclusions

We suggest that the spatial distribution of uplift/erosion and pre­

cipitation dictates the distribution and thickness of laterite in
Madagascar. We have also proposed a potential link between the local
prevalence of oxisol and poor agricultural productivity, which, in the
absence of advanced agricultural technologies, may reinforce and
exacerbate one measure of socioeconomic poverty. Laterite formation
only occurs within a relatively narrow band of present-day precipitation
Fig. 4. Relationship between laterite and saprolite thickness and (a) mean (500–900 mm), implying that its governing processes have evolved
annual precipitation from co-located rain gauge or sampled from 5 km clima­ under the influence of geological-timescale changes in climate and
tology grid if gauge not present (N = 65); (b) 10Be-derived erosion rate (N = (especially) tectonics (i.e. epeirogeny). Indeed, alternating wet and dry
18). For laterite thicknesses under 10 m (vertical dashed line), R2 = 0.93 be­ conditions, as well as cyclic pulses of uplift followed by relative tectonic
tween erosion rate and laterite thickness; and (c) District-level headcount index quiescence, are favourable for the formation of oxisols. We have shown
of poverty (SEDAC, 2021; N = 65). how uplift and erosion rates can vary dramatically (by three orders of
magnitude) over small distances (~100 km), restricting the develop­
ment of thick (>10 m) piles of laterite to two much smaller patches of
land, to the north and south of a central belt of recent volcanism and
rapid uplift rates.

Declaration of competing interest

The authors declare that they have no known competing financial

interests or personal relationships that could have appeared to influence
the work reported in this paper.

Data availability

Data will be made available on request.

Fig. 5. Relationship between mean annual precipitation and District-level Acknowledgements

headcount index of poverty (SEDAC, 2021). Points sampled at centre of each
5 km climatology grid square. We acknowledge seed funding from the EU Horizon-2020 scheme.
Figures were prepared using ArcGIS, GMT v6.2.0, and Adobe Illustrator.
large changes in uplift rate over small distances (Stephenson et al., We are grateful to the large number of communities across Madagascar
2019). In the centre of the island, the highest erosion rates (and highest with whom we worked collecting laterite cores, samples for radiometric
predicted uplift rates) occur in a small area marked by the total absence dating analyses, and installing rain gauges.

6
J.D. Paul et al. Journal of African Earth Sciences 196 (2022) 104681

Appendix
Table A1
Coordinates of localities for laterite + saprolite thickness measurements, mean annual rain gauge precipitation, and erosion rates calculated using 10Be data. †Rain
gauge data from WMO (2020). *Thickness evaluated from borehole core recovery. ‡10Be-derived erosion rates from Cox et al. (2009). Table A2 contains additional
information of erosion rate calculations.

ID Name Latitude Longitude Elevation Laterite thickness Mean annual gauged precipitation (2018–20), Erosion rate, m
(dd) (dd) (m) (m) mm Myr− 1

1 Joffreville 49.143 − 12.653 494 0 1383 –

2 Ambodimanga 49.746 − 13.068 302 0 1654 –
3 Ambanja 48.471 − 13.680 71 0 1987 –
4 Ambodimanga 49.683 − 13.765 513 0 1780 –
5 Route Antsohihy 48.749 − 14.911 955 0 – –
6 Amparihy 47.121 − 15.666 28 0 1601 –
7 Fampotabe 50.146 − 15.933 16 0 1535 –
8 Sahondra 45.101 − 16.747 223 0 1225 –
9 Ambodiadaba 47.707 − 16.524 501 1.8 – –
10 Soanierana 49.578 − 16.917 39 0 1552 –
11 Ambodirano* 47.366 − 17.410 685 8.5 755 –
12 Ambatondrazaka 48.417 − 17.850 1018 7.2 718 12.3 ± 1.4
13 Ampandrano 47.311 − 18.036 1083 10.1 790 –
14 Brickaville 49.036 − 18.843 296 0 1343 –
15 Anjiabe 45.887 − 17.561 338 0.6 1201 –
16 Ampanihy 44.413 − 18.111 109 0 1030 –
17 Analavory 46.587 − 18.871 983 2.5 1129 –
18 Ankoririka 46.106 − 18.065 731 4.4 895 –
19 Ambalarano 45.430 − 18.564 683 1.5 1080 32.2 ± 3.5
20 Antananarivo 47.526 − 18.935 1353 0.4 1028 45.0 ± 4.4
21 Moramanga 48.225 − 18.972 958 0.2 1276 –
22 Antsirabe 47.017 − 19.877 1583 0 1005 35.2 ± 3.5
23 Miandrivazo 45.684 − 19.652 394 0 – –
24 Andranolava 44.582 − 19.315 57 0 583 –
25 Mahahoro 48.711 − 19.878 102 0 1079 –
26 Amboasary 46.351 − 20.608 1242 4.0 740 21.8 ± 2.3
27 Tsianaloka 45.270 − 20.397 280 0 694 –
28 Andakana 47.303 − 20.849 1552 0 1008 –
29 Belo sur Mer 43.990 − 20.941 37 0 93 –
30 Route Fianar* 46.595 − 21.621 914 16.0 665 6.8 ± 0.8
31 Manakara 48.069 − 22.019 6 0 1280 –
32 Taolagnaro 46.932 − 25.074 11 0 795 –
33 Besaroa* 45.403 − 21.166 542 11.8 689 –
34 Ihosy 46.052 − 22.422 1033 23.5 540 2.8 ± 0.5
35 Ivohibe 46.894 − 22.489 653 3.8 986 –
36 Morombe 43.386 − 21.776 8 0 112 –
37 Toliara 43.742 − 23.388 16 0 270 –
38 Lavanono 44.377 − 25.220 10 0 352 –
39 Sakaraha 44.789 − 22.744 691 13.2 532 6.0 ± 0.8
40 Benenitra 45.025 − 23.561 263 4.0 – –
41 Ambatofotsy 45.763 − 24.500 347 0.4 725 –
42 Betroka* 46.342 − 23.393 1149 9.9 968 –
43 Mahasoa 45.648 − 23.155 688 13.8 466 –
44 Vohipana 47.486 − 23.561 31 0 1515 –
45 Ejeda 44.596 − 24.257 314 0.5 633 –
46 Befotaka 46.721 − 24.034 836 0 1085 –
47 Mandoto 46.481 − 19.589 1126 0.9 1002 –
48 Anivorano 45.771 − 21.745 642 20.2 679 3.2 ± 0.6
49 Manja 44.427 − 21.542 292 0 444 –
50 Bematso 48.494 − 17.124 973 0.2 – –
51 Kalandy 48.714 − 15.683 643 0 – –
52 Ambohitromby 47.166 − 18.432 1247 6.4 – –
53 Andrasirisa 45.310 − 24.048 466 0.1 – –
54 Ranohira 45.284 − 22.562 985 16.7 – 5.9 ± 0.9
55 Mangorva* 44.632 − 23.200 317 5.2 – –
56 Beraiketa 46.436 − 22.757 903 12.9 – 9.1 ± 0.9
57 Ambohimahamasima 47.259 − 21.906 1309 0.1 – –
58 Begogo 46.761 − 23.498 680 0.2 – –
59 Ambalavao 46.780 − 20.782 1620 1.0 – –
60 Marvombo* 45.718 − 23.656 648 3.8 – –
61 Antsiranana 49.295 − 12.336 171 – 1296† –
62 Vohemar 50.016 − 13.397 8 – 1643† –
63 Sambava 50.183 − 14.281 15 – 1704† –
64 Antalaha 50.277 − 14.925 57 – 1685† –
65 Andapa 49.638 − 14.675 534 – 1620† –
66 Nosy-Be, Hell-Ville 48.275 − 13.411 3 – 1852† –
67 Andamoty 47.959 − 14.361 129 – 1678† –
68 Ambato Boina 46.722 − 16.479 17 – 1267† –
69 Besalampy 44.487 − 16.758 17 – 1442† –
(continued on next page)

7
J.D. Paul et al. Journal of African Earth Sciences 196 (2022) 104681

Table A1 (continued )
ID Name Latitude Longitude Elevation Laterite thickness Mean annual gauged precipitation (2018–20), Erosion rate, m
(dd) (dd) (m) (m) mm Myr− 1

70 Nosy-Boraha 49.923 − 16.918 34 – 1586† –

71 Toamasina 49.406 − 18.143 8 – 1520† –
72 Antananarivo 47.503 − 18.877 1299 – 997† –
73 Antsirabe 47.030 − 19.882 1589 – 1010† –
74 Mahanoro 48.806 − 19.895 2 – 1143† –
75 Mananjary 48.342 − 21.236 9 – 1392† –
76 Farafagana 47.812 − 22.828 11 – 1267† –
77 Taolagnaro 46.980 − 25.018 12 – 932† –
78 Erada 45.946 − 25.298 111 – 650† –
79 Ilakaka 45.226 − 22.701 831 – 421† –
80 Toliara 43.675 − 23.357 12 – 89† –
81 Morombe 43.395 − 21.768 7 – 54† –
82 Morondava 44.345 − 20.322 8 – 69† –
83 Mahajanga 46.357 − 15.735 28 – 1076† –
84 Maintirano 44.074 − 18.101 22 – 830† –
85 Fianarantsoa 47.101 − 21.452 1083 – 945† –
86 Miarinarivo 46.964 − 18.994 1421 4.5 – 5.8 ± 0.8‡
87 Amparafaravola 48.210 − 17.549 850 8.9 – 3.3 ± 0.6‡
88 Onibe watershed 47.125 − 19.005 1303 – – 11.4 ± 1.5‡
89 Ikopa watershed 47.526 − 18.950 1254 5.8 – 5.8 ± 0.8‡
90 Miarinarivo 46.823 − 18.945 1188 3.0 – 15.1 ± 2.0‡
91 Amparafaravola 48.204 − 17.628 778 2.5 – 19.7 ± 2.5‡

Table A2
10
Be sample information for central and southern Madagascar. ‡10Be-derived erosion rates from Cox et al. (2009). aProduction rate multiplier, calculated from Lal
(1988) for neutrons only. bUncertainties on individual samples represent 1σ analytical uncertainty. cErosion rates are calculated using density of 2.7 g cm− 3, atten­
uation coefficient of 165 g cm− 2, and sea-level, high-latitude production rate of 5.3 atoms g− 1 yr− 1. For shallow samples, shielding of 22 g cm− 2 is used, assuming soil
density of 1.45 g cm− 2. Erosion rates are expressed in rock equivalent (r = 2.7 g cm− 2). dFor river samples, slope is the watershed average measured from the 30
m-resolution SRTM DEM data. For lavakas and colluvium, it is the local average, measured at the sample location.
10
Sample ID (per Table A1) Sample type PRM* Beb (atoms g− 1 10
Be) x 105 Erosion ratec (m Myr− 1) Sloped (◦ )

12 Colluvium, 10 cm depth 1.27 5.8 ± 0.15 12.3 ± 1.4 9

19 Colluvium, 10 cm depth 1.25 3.4 ± 0.10 32.2 ± 3.5 10
20 Colluvium, 10 cm depth 1.20 2.3 ± 0.05 45.0 ± 4.4 15
22 Colluvium, 10 cm depth 1.25 2.5 ± 0.09 35.2 ± 3.5 7
26 Colluvium, 10 cm depth 1.78 3.4 ± 0.09 21.8 ± 2.3 7
30 Colluvium, 10 cm depth 2.08 11.2 ± 0.42 6.8 ± 0.8 3
34 Colluvium, 10 cm depth 1.27 13.3 ± 0.55 2.8 ± 0.5 2
39 River sand 1.74 10.5 ± 0.23 6.0 ± 0.8 3
48 River sand 1.28 6.5 ± 0.97 3.2 ± 0.6 2
54 River sand 2.01 11.3 ± 0.54 5.9 ± 0.9 6
56 River sand 1.45 9.7 ± 0.32 9.1 ± 0.9 5
86 River sand 1.92 10.0 ± 0.23 5.8 ± 0.8‡ 9
87 Colluvium, 10–20 cm depth 1.25 6.2 ± 1.05 3.3 ± 0.6‡ 33
88 River sand 2.15 5.9 ± 0.18 11.4 ± 1.5‡ 9
89 Colluvium, 10 cm depth 1.94 10.2 ± 0.31 5.8 ± 0.8‡ –
90 River sand 1.8 3.7 ± 0.14 15.1 ± 2.0‡ 9
91 River sand 1.29 2.1 ± 0.06 19.7 ± 2.5‡ 9

References Brahmachari, K., Sarkar, S., Santra, D.K., Maitra, S., 2019. Millet for food and nutritional
security in drought prone and red laterite region of eastern India. Int. J. Phys. Soc.
Sci. 26 (6), 1–7.
Akpokodje, E.G., Hudec, P., 1992. Properties of concretionary laterite gravel concrete.
Braucher, R., Colin, F., Brown, E.T., Bourles, D.L., Bamba, O., et al., 1998. African laterite
Bull. Int. Assoc. Eng. Geol. 46, 45–50.
dynamics using in situ-produced 10Be. Geochem. Cosmochim. Acta 62 (9),
Bardintzeff, J.-M., Liégeois, J.-P., Bonnin, B., Bello, H., Rasmimanana, G., 2010.
1501–1507.
Madagascar volcanic provinces linked to the Gondwana break-up: geochemical and
Braun, J.-J., Descloitres, M., Riotte, J., Fleury, S., Barbiero, L., et al., 2009. Regolith mass
isotopic evidences for contrasting mantle sources. Gondwana Res. 18, 295–314.
balance inferred from combined mineralogical, geochemical and geophysical
Beauvais, A., Chardon, D., 2013. Modes, tempo and spatial variability of Cenozoic
studies: mule Hole gneissic watershed, South India. Geochem. Cosmochim. Acta 73
cratonic denudation: the West African example. G-cubed 14 (5). https://doi.org/
(4), 935–961.
10.1002/ggge.20093.
Buol, S.W., Eswaran, H., 1999. Oxisols. In: Sparks, D.L. (Ed.), Advances in Agronomy,
Beauvais, A., Ruffet, G., Hénoque, O., Colin, F., 2008. Chemical and physical erosion
151–195, Cambridge, MA: Academic PressCadot O., Dutoit L., Olarreaga M., 2006.
rhythms of the West African Cenozoic morphogenesis: the 39Ar-40Ar dating of
How Costly Is it for Poor Farmers to Lift Themselves Out of Poverty? Policy Research
supergene K-Mn oxides. J. Geophys. Res. 113, F04007.
Working Paper 3881. World Bank, Washington D.C.
Berger, A., Janots, E., Gnos, E., Frei, R., Bernier, F., 2014. Rare earth element mineralogy
Burke, K., Gunnell, Y., 2008. The African Erosion Surface: a continental-scale synthesis of
and geochemistry in a laterite profile from Madagascar. Appl. Geochem. 41,
geomorphology, tectonics, and environmental change over the past 180 million
218–228.
years. Geological Society Memoir 201.
Bertil, D., Regnoult, J.M., 1998. Seismotectonics of Madagascar. Tectonophysics 294
Chand, S., Subrahmanyam, C., 2003. Rifting between India and Madagascar –
(1–2), 57–74.
mechanism and isostasy. Earth Planet Sci. Lett. 210 (1–2), 317–332.
Bourgeat, F., Petit, M., 1969. Contribution à l’étude des surfaces d’aplanissement sur les
Hautes-Terres centrales. Annales Géographiques 426, 158–188.

8
J.D. Paul et al. Journal of African Earth Sciences 196 (2022) 104681

Cox, R., Bierman, P., Jungers, M., Rakotondrazafy, A.F.M., 2009. Erosion rates and Nishiizumi, K., Imamura, M., Caffee, M.W., Southon, J.R., Finkel, R.C., McAninch, J.,
sediment sources in Madagascar inferred from 10Be analaysis of lavaka, slope, and 2007. Absolute calibration of 10Be AMS standards. Nucl. Instrum. Methods Phys. Res.
river sediment. J. Geol. 117, 363–376. B 258, 403–413.
Cox, R., Zentner, D.B., Rakotondrazafy, A.M., Rasoazanamparany, C.F., 2010. Paul, J.D., 2021. Controls on eroded rock volume, a proxy for river incision, in Africa.
Shakedown in Madagascar: occurrence of lavakas (erosional gullies) associated with Geology 49 (4), 422–427.
seismic activity. Geology 38 (2), 179–182. Paul, J.D., Eakin, C.M., 2017. Mantle upwelling beneath Madagascar: evidence from
Davids, J.C., Devkota, N., Pandey, A., Prajapari, R., Ertie, B.A., et al., 2019. Soda bottle receiver function analysis and shear wave splitting. J. Seismol. 21 (4), 825–836.
science – citizen science Monsoon precipitation monitoring in Nepal. Front. Earth Paul, J.D., Roberts, G.G., White, N., 2014. The African landscape through space and time.
Sci. 7, 46. Tectonics 33 (6), 898–935.
De Laulanié, H., 2011. Intensive rice farming in Madagascar. Tropicultura 29 (3), Paul, J.D., Cieslik, K., Sah, N., Shakya, P., Parajuli, B.P., Paudel, S., Dewulf, A.,
183–187. Buytaert, W., 2020. Applying citizen science for sustainable development: rainfall
De Wit, M., 2003. Madagascar: heads it’s a continent, tails it’s an island. Annu. Rev. monitoring in western Nepal. Frontiers in Water 2, 62.
Earth Planet Sci. 31, 213–248. Ralison, O.H., Borges, A.V., Dehairs, F., Middelburg, J.J., Bouillon, S., 2008. Carbon
Dill, H.G., 2017. Residual clay deposits on basement rocks: the impact of climate and the biogeochemistry of the Betsiboka estuary (north-western Madagascar). Org.
geological setting on supergene argillitisation in the Bohemian Massif (Central Geochem. 39, 1649–1658.
Europe) and across the globe. Earth Sci. Rev. 165, 1–58. Reardon, T., Vosti, S.A., 1995. Links between rural poverty and the environment in
Dill, H.G., Ludwig, R.-R., Kathewera, A., Mwenelupembe, J., 2005. A lithofacies terrain developing countries: asset categories and investment poverty. World Dev. 23 (9),
model for the Blantyre Region: implications for the interpretation of palaeosavanna 1495–1506.
depositional systems and for environmental geology and economic geology in Roberts, G.G., Paul, J.D., White, N., Winterbourne, J., 2012. Temporal and spatial
southern Malawi. J. Afr. Earth Sci. 41 (5), 341–393. evolution of dynamic support from river profiles: a framework for Madagascar. G-
Dixey, F., 1960. The geology and geomorphology of Madagascar, and a comparison with cubed 13 (4), Q04004.
eastern Africa. Q. J. Geol. Soc. Lond. 116, 255–268. See, S., Subramanian, S., 2002. Counting the poor: an elementary difficulty in the
Dwevedi, A., Kumar, P., Kumar, P., Kumar, Y., Sharma, Y.K., Kayastak, A.M., 2017. Soil measurement of poverty. Econ. Philos. 18, 1–12.
sensors: detailed insight into research updates, significance, and future prospects. Socioeconomic Data and Applications Center (SEDAC), 2021. Poverty datasets [online].
New Pesticides and Soil Sensors 561–594. Available at: https://sedac.ciesin.columbia.edu/data/sets/browse. (Accessed 20
Estrade, G., Marquis, E., Smith, M., Goodenough, K., Nason, P., 2019. REE concentration September 2021).
processes in ion adsorption deposits: evidence from the Ambohimirahavavy alkaline Stephenson, S.N., White, N.J., Li, T., Robinson, L.F., 2019. Disentangling interglacial sea
complex in Madagascar. Ore Geol. Rev. 112, 103027. level and global dynamic topography: analysis of Madagascar. Earth Planet Sci. Lett.
Food and Agriculture Organization of the United Nations (FAO), 2016. More effective 519, 61–69.
and sustainable investments in water for poverty reduction [online]. Available at: Stephenson, S.N., White, N.J., Carter, A., Seward, D., Ball, P.W., Klöcking, M., 2021.
http://www.fao.org/in-action/water-for-poverty-in-africa/countries/madagasca Cenozoic dynamic topography of Madagascar. G-cubed 22, e2020GC009624.
r/mapping/en/. (Accessed 24 September 2021). Storey, M., Mahoney, J., Saunders, A., Duncan, R., Kelley, S., Coffin, M., 1997. Timing of
Foster, J., Greer, J., Thorbecke, E., 1984. A class of decomposable poverty measures. hot spot-related volcanism and the breakup of Madagascar and India. Science 267,
Econometrica 52, 761–776. 852–855.
Gleeson, S.A., Herrington, R.J., Durango, J., Velasquez, C.A., Koll, G., 2004. The Stuiver, M., Reimer, P.J., Reimer, R., 2006. CALIB radiocarbon calibration. Version 5.0.1.
mineralogy and geochemistry of the Cerro Matoso S.A. Ni laterite deposit, Available at: http://calib.qub.ac.uk/calib. (Accessed 21 May 2021).
Montelibano, Colombia. Econ. Geol. 99 (6), 1197–1213. Tardy, Y., 1997. Petrology of Laterites and Tropical Soils. Balkema, Amsterdam.
Goodman, S.M., Benstead, J.P., 2005. Updated estimates of biotic diversity and Tardy, Y., Roquin, C., 1992. Geochemistry and evolution of lateritic landscapes. In:
endemism for Madagascar. Oryx 39 (1). https://doi.org/10.1017/ Martini, I.P., Chesworth, W. (Eds.), Weathering, Soils and Paleosols. Developments
S0030605305000128. in Earth Surface Processes, vol. 2, pp. 407–471.
Granger, D.E., Kirchner, J.W., Finkel, R., 1996. Spatially averaged long-term erosion Vagen, T.-G., Shepherd, K.D., Walsh, M.G., 2006. Sensing landscape level change in soil
rates measured from in situ produced cosmogenic nuclides in alluvial sediments. fertility following deforestation and conversion in the highlands of Madagascar using
J. Geol. 104, 249–257. Vis-NIR spectroscopy. Geoderma 133 (3–4), 281–294.
Grieser, R., 1994. Environmental Education and Communication (EE&C) Assessment for von Heland, J., Folke, C., 2014. A social contract with the ancestors – culture and
Madagascar. GreenCOM report for U.S. Agency for International Development ecosystem services in southern Madagascar. Global Environ. Change 24, 251–264.
(USAID), Academy for Educational Development, p. 77. Wells, N.A., Andrimihaja, B., 1993. The initiation and growth of gullies in Madagascar:
Holmes, A., 1944. Principles of Physical Geology. Thomas Nelson, Edinburgh, UK. are humans to blame? Geomorphology 8 (1), 1–46.
Hou, A.Y., Kakar, R.K., Neeck, A.A., Azarbarzin, A., Kummerow, C.D., et al., 2014. The Widdowson, M., 1997. Tertiary palaeosurfaces of the SW Deccan, western India:
global precipitation measurement mission. Bull. Am. Meteorol. Soc. 95, 701–722. implications for passive margin uplift. In: Widdowson, M. (Ed.), Palaeosurfaces:
Kohl, C.P., Nishiizumi, 1992. Chemical isolation of quartz for measurement of in situ Recognition, reconstruction and palaeoenvironmental interpretation. Geological
produced cosmogenic nuclides. Geochem. Cosmochim. Acta 56, 3583–3587. Society of London, Special Publications 120, 221–248.
Kusky, T., Toraman, E., Raharimahefa, T., Rasoazanamparany, C., 2010. Active tectonics World Meteorological Organization (WMO), 2020. CMO Survey on National Summaries
of the alaotra-ankay graben system, Madagascar: possible extension of somalian- of Methods and Instruments for Solid Precipitation Measurement at Automatic
african diffusive plate boundary? Gondwana Res. 18, 274–294. Weather Stations: Instruments and Observing Methods Report 102. WMO
Lal, R., 1988. Soil degradation and the future of agriculture in sub-Saharan Africa. J. Soil Publications, Geneva.
Water Conserv. 43, 444–451. Zauyah, S., Schaefer, C.E.G.R., Simas, F.N.B., 2010. Saprolites. In: Marcelino, V., Mees, F.
Manz, B., Buytaert, W., Zulkafli, Z., Lavado, W., Willems, B., et al., 2016. High-resolution (Eds.), Interpretation or Micromorphological Features of Soils and Regoliths.
satellite-gauge merged precipitation climatologies of the Tropical Andes. J. Geophys. Elsevier, Amsterdam, pp. 49–68.
Res. Atmos. 121 (3), 1190–1207.