Search

Soil erosion is a serious issue in highland areas of Sri Lanka that are dominated by extensive agricultural activities. Methods capable of rapidly assessing soil erosion are required to determine suitable conservation methods and to monitor their effectiveness to preserve this non-renewable resource. The widely used method of determining soil erosion rate is the use of field measurements and/or empirical modeling. Although the latter method along with the use of GIS/Remote Sensing has become very popular, its main limitations are the lack of model validation and lack of comparison of modeled rates with acceptable soil loss tolerance values. In this study, soil erosion rates of six sub-catchments were calculated with the revised universal soil loss equation (RUSLE) following five different methods in GIS/RS environment and then compared with field-based measurements for verification. The mean modeled soil erosion rates of the six sub-catchments were within 0.12 and 7.70 t ha−1 y−1, whereas the field-based rates range from 1.14 to 15.83 t ha−1 y−1. Moreover, modeled erosion classes within the sub-catchments were similar to field-based localized erosion rates over different land use types quantified from plot experiments. Hence, we recommend the integration of RUSLE with GIS/RS as an effective tool for soil erosion hazard mapping at the catchment scale. We propose a typical rate of 1 t ha−1 y−1 and an upper limit of 2 t ha−1 y−1 as the soil loss tolerance level for the study area based on experimentally derived soil production rates. Finally, a soil erosion map of the entire Upper Mahaweli River Catchment was developed using the best method, and then its sub-catchments were ranked according to severity of soil erosion. The methodology adapted here can be used as a guide in land management and soil conservation programs.

How flowing water and organisms can shape Earth9s surface, the Critical Zone, depends on how fast this layer is turned over by erosion. To quantify the dependence of rock weathering and the cycling of elements through ecosystems on erosion we have used existing and new metrics that quantify the partitioning and cycling of elements between rock, saprolite, soil, plants, and river dissolved and solid loads. We demonstrate their utility at three sites along a global transect of mountain landscapes that differ in erosion rates – an “erodosequence”. These sites are the Swiss Central Alps, a rapidly-eroding, post-glacial mountain belt; the Southern Sierra Nevada, USA, eroding at moderate rates; and the slowly-eroding tropical Highlands of Sri Lanka. The backbone of this analysis is an extensive data set of rock, saprolite, soil, water, and plant geochemical and isotopic data. This set of material properties is converted into process rates by using regolith production and weathering rates from cosmogenic nuclides and river loads, and estimates of biomass growth. Combined, these metrics allow us to derive elemental fluxes through regolith and vegetation. The main findings are: 1) the rates of weathering are set locally in regolith, and not by the rate at which entire landscapes erode; 2) the degree of weathering is mainly controlled by regolith residence time. This results in supply-limited weathering in Sri Lanka where weathering runs to completion in the regolith, and kinetically-limited weathering in the Alps and Sierra Nevada where soluble primary minerals persist; 3) these weathering characteristics are reflected in the sites9 ecosystem processes, namely in that nutritive elements are intensely recycled in the supply-limited setting, and directly taken up from soil and rock in the kinetically settings; 4) the weathering rates are not controlled by biomass growth; 5) at all sites we find a deficit in river solute export when compared to solute production in regolith, the extent of which differs between elements. Plant uptake followed by litter export might explain this deficit for biologically utilized elements of high solubility, and rare, high-discharge flushing events for colloidal-bound elements of low solubility. Our data and new metrics have begun to serve for calibrating metal isotope systems in the weathering zone, the isotope ratios of which depend on the flux partitioning between the compartments of the Critical Zone. We demonstrate this application in several isotope geochemical companion papers.

The Upper Mahaweli Catchment (UMC) is a very important watershed in the Central Highlands of Sri Lanka since it contains a series of multipurpose reservoirs. Hence, conservation of soil in the UMC should be given the highest priority to reduce sediment yields in streams and to maintain the storage capacities of hydropower reservoirs. The sediment yields measured in the sub-catchments of the UMC before 1995 have revealed that soil erosion is intense in the contributing catchments and siltation in some reservoirs is at an alarming level. The situation in the Upper Uma Oya, a sub-catchment of the UMC, is the worst, reporting the highest sediment yield measured for any catchment in Sri Lanka. Mahaweli Authority of Sri Lanka had embarked on two soil and water conservation projects in the Upper Uma Oya catchment from 1995 to 2005 together with monitoring of sediment yields in the stream. Sediment yield measurements in this catchment revealed that the wash load in the stream had been reduced by a factor of five after the implementation of conservation programmes in 1995. Our temporal analyses showed that the role of the other key factors that control soil erosion on a hillslope such as erodibility, slope, erosivity and land use cover is minimal to generate a five-fold reduction in the wash load of the stream. Hence, we report that the conservation measures adopted in the critical areas of the Upper Uma Oya have been very successful and had greatly contributed to the reduction of soil erosion. However, even after conservation, soil erosion rates estimated by stream loads are about 10 times higher than the natural background rate. Hence, as shown in other catchments elsewhere in the world, the ongoing rate of soil erosion can be further lowered by strengthening the existing soil and water management practices together with restoration of vegetation on bare lands.

Forest ecosystem nutrition and rock weathering are intimately linked through the supply of mineral nutrients and recycling (re-utilisation of nutrients released from plant litter during mineralisation). Here, we report on novel applications of Mg stable isotopes in combination with mass balance calculations to quantify nutrient supply (from bedrock, atmospheric wet and dry deposition) and recycling in a tropical ecosystem, characterised by a supply-limited weathering regime in the Central Highlands of Sri Lanka. We measured Mg stable isotopes, radiogenic Sr isotope ratios, and element concentrations in the different compartments of the Earth's Critical Zone (bedrock, saprolite, soil, soil pore water, stream water and vegetation) as tracers for element mass fluxes and pathways. We find that in the clay-rich saprolite Mg isotopes are fractionated towards isotopically heavy Mg during secondary mineral formation. Soil pore water and stream water are the complementary isotopically light reservoirs. The Mg isotope composition of above-ground vegetation is on average identical to soil pore water. Based on these observations we suggest that a tight near-surface mineral nutrient loop between plants and soil sustains the tropical forest ecosystem. This loop is disconnected from the rock weathering nutrient source in the saprolite beneath soil. External atmospheric wet and dry inputs (mainly Mg and Ca) strongly contribute to the plant-available nutrient budget that the nutrient depleted regolith does not supply. These findings, based on Mg stable isotopes, agree with independently-obtained information on bio-element (P, K, Ca, Mg, Si) recycling. This nutrient supply setting contrasts with that reported for the Southern Sierra Nevada (USA), where a kinetically limited weathering regime prevails. Our study illustrates the application of new flux estimators based on elemental and metal stable isotope measurements across a range of climate and weathering regimes.

The reduction of productivity in arable lands of the tropical highlands through intensified soil erosion is a major reason for food scarcity in many developing countries. Identifying soil erosion hot spots in highland agricultural croplands and monitoring the effectiveness of soil conservation are crucial for improving land management practices. This paper develops a low cost and efficient method to quantify the rate of soil loss caused by splash and sheet erosion, in-situ at the plot scale. This method assumes that textural differences between the topsoil and the underlying soil within the plot are due to erosion of the top layer. Grain size analyses conducted in agricultural soils of the Upper Uma Oya Catchment (UUOC) of Sri Lanka provided data to quantify splash or sheet erosion and deposition. Net soil loss from the plot was calculated based on rates of area weighted soil loss and deposition. The net splash and sheet erosion rate in the plot was 20 t ha−1 yr−1, which is comparable with previous soil erosion results obtained in the area (25 t ha−1 yr−1) at similar temporal and spatial scale and under the same agricultural use. This study suggests that approximately half of the total erosion in the UUOC area results from splash and sheet erosion, whereas the rest is driven by concentrated flow through rill and gully erosion. This method is potentially transferable to other croplands as a robust approach for rapid measurement of splash and sheet erosion.

The serpentinite blocks of Indikolapelessa, located along an identified litho-tectonic boundary between the Highland Complex (HC) and the Vijayan Complex (VC) of Sri Lanka, have undergone extensive lateralization with metal enrichment. Characteristic serpentinite vegetation with some endemic species was recognized in the soils and supergene deposits develop on serpentinite lithology. This type of geological and ecological relationship forms vegetation covers on serpentinite lithologies which are sharply demarcated from the surrounding metamorphic terrains. The aforesaid “geo-ecological phenomenon” can be used as a tool for geo-vegetation mapping in ultramafic terrains to trace the geological boundaries in landscapes where rock outcrops are virtually absent. We successfully applied the concept of geo-vegetation mapping in order to demarcate the boundary of underlain serpentinite rocks from surrounding non-serpentinite metamorphic rocks (e.g. granitic gneiss). The hypothesis was supported by the geochemical variations of soils/supergene deposits found at serpentinite and non-serpentinite sites, especially immobile elements and some trace elements. Based on whole rock chemistry and soil chemical data obtained, we suggest that the Indikolapelessa serpentinite outcrop, together with the other four serpentinite outcrops, is more likely to represent the Mg-rich mantle fragments at the time of overthrusting of the two crustal blocks of HC and VC during the Pan-African event.

Estimation of suspended sediment concentration (SSC) in rivers is a prerequisite to address many issues related to hydrology. Therefore, we make an attempt in this study to introduce a low-cost technique to estimate the SSC. Both surface and depth-and-width-integrated water samples were collected and measured for SSC from eight tributaries in Sri Lanka over a complete hydrological year. A site-specific calibration curve was established between SSCs measured by two methods for each tributary where R2 varied from 0.72 to 0.99. The same relationship is developed in general for all tributaries studied in the hilly terrain of Sri Lanka. This generic model exhibits a strong correlation (R2 = 0.91), which will be useful to calculate an accurate SSC from a simply measured surface SSC. To select the appropriate gauging method, be it surface or depth-and-width-integrated sampling, a new concept of surface sampling threshold factor (SSTF) is introduced. The preliminarily analysis on SSTF using available data for the studied catchments reveals that surface sampling is only adequate for estimating a representative SSC if SSTF varies from 35 to 45. When SSTF deviates from this range, the SSC measured by surface sampling needs to be adjusted by depth-and-width-integrated sampling.

In the Sri Lankan Highlands erosion and chemical weathering rates are among the lowest for global mountain denudation. In this tropical humid setting, highly weathered deep saprolite profiles have developed from high-grade metamorphic charnockite during spheroidal weathering of the bedrock. The spheroidal weathering produces rounded corestones and spalled rindlets at the rock-saprolite interface. We used detailed textural, mineralogical and chemical analyses to reconstruct the sequence of weathering reactions and their causes. The first mineral attacked by weathering was found to be pyroxene initiated by in situ Fe oxidation. Volumetric calculations suggest that this oxidation leads to the generation of porosity due to the formation of micro-fractures allowing for fluid transport and subsequent dissolution of biotite and plagioclase. The rapid ensuing plagioclase weathering leads to formation of high secondary porosity in the corestone over a distance of only a few cm and eventually to the final disaggregation of bedrock to saprolite. The first secondary phases are oxides or amorphous precipitates from which secondary minerals (mainly gibbsite, kaolinite and goethite) form. As oxidation is the first weathering reaction, the supply of O-2 is a rate-limiting factor for chemical weathering. Hence, the supply of O-2 and its consumption at depth connects processes at the weathering front with those at the Earth's surface in a feedback mechanism. The strength of the feedback depends on the relative weight of advective versus diffusive transport of O-2 through the weathering profile. The feedback will be stronger with dominating diffusive transport. The low weathering rate is explained by the nature of this feedback that is ultimately dependent on the transport of O-2 through the whole regolith, and on lithological factors such as low bedrock porosity and the amount of Fe-bearing primary minerals. Tectonic quiescence in this region and low pre-development erosion rate (attributed to a dense vegetation cover) minimize the rejuvenation of the thick and cohesive regolith column, finally leading to low denudation rates. (C) 2015 Elsevier B.V. All rights reserved.

Silicate weathering - initiated by major mineralogical transformations at the base of ten meters of clay-rich saprolite generates the exceptionally low weathering flux found in streams draining the crystalline rocks of the mountainous and humid tropical Highlands of Sri Lanka. This conclusion is reached from a thorough investigation of the mineralogical, chemical, and Sr isotope compositions of samples within a regolith profile extending >10 m from surface soil through the weathering front in charnockite bedrock (a high-grade metamorphic rock), corestones formed at the weathering front, as well as from the chemical composition of the dissolved loads in nearby streams. Weatherable minerals and soluble elements are fully depleted at the top of the profile, showing that the system is supply-limited, such that weathering fluxes are controlled directly by the supply of fresh minerals. We determine the weathering rates using two independent means: (1) in situ-produced cosmogenic nuclides in surface soil and creek sediments in the close vicinity of the regolith combined with immobile element mass balance across the regolith and (2) river dissolved loads. Silicate weathering rates determined from both approaches range from 16 to 36 t km(-2) y(-1), corresponding to a weathering front advance rate of 6-14 mm ky(-1). These rates agree across the 10(1) to 10(4) - y time scales over which our rate metrics integrate, suggesting that the weathering system operates at steady state. Within error these rates are furthermore compatible with those obtained by modeling the advance rate of the weathering front from chemical gradients and mineral dissolution rates. The silicate weathering flux out of the weathering profile, measured on small creeks, amounts to 84% of the profile's export flux; the remaining 16% is contributed by non-silicate, atmospheric-derived input. The silicate weathering flux, as measured by dissolved loads in large catchments, amounts to ca. 50% of the total dissolved flux; the remainder being contributed by dust, rain, and weathering of local marble bands. Spheroidal weathering is the key processes of converting the fresh bedrock into saprolite at the weathering front. The mineralogical composition of weathering rinds shows that the sequence of mineral decomposition is: pyroxene; plagioclase; biotite; K-feldspar. Observable biotite alteration does not appear to initiate spheroidal weathering within corestones; therefore, we infer that other processes than biotite oxidation, like pyroxene oxidation, clay formation from pyroxene and plagioclase decomposition, the development of secondary porosity by plagioclase dissolution, or even microbiologic processes at depth enable the coupling between slow advance of the weathering front and slow erosion at the surface. The comparison to tectonically more active tropical landscapes lets us conclude that the combination of hard rock with tightly interlocked mineral grains and slow erosion in the absence of tectonically-induced landscape rejuvenation lead to these exceptionally low weathering rates. (C) 2013 Elsevier Ltd. All rights reserved.

[1] The brittle structural inventory of southern and southwestern Sri Lanka has been studied by kinematic, mineralogical and thermochronological techniques. Thermochronological analyses of faults comprise apatite and zircon fission track (FT) data and apatite (U-Th-Sm)/He ages from fault planes and undisturbed host rocks, and range between ∼65 and ∼230 Ma. The ages of both settings are undistinguishable for topographic altitudes below ∼100 m, while fault planes from higher elevations are significantly younger than the corresponding host rocks. Thermal history modeling and qualitative interpretation of the thermochronological data identify at least five episodes of thermal overprint associated with faulting activity occurring at 159 ± 18, 144 ± 14, 120 ± 10, 94 ± 8, and 70 ± 10 Ma. The kinematic, mineralogical and thermochronological data collectively show that Sri Lanka was subjected to major N-S oriented extension subsequent to the Gondwana breakup. The resulting brittle structures exert primary control on the island's geomorphology, especially on the Southern Escarpment of the Sri Lankan Highlands.

Escarpments are prominent morphological features along high-elevation passive margins. Recent studies integrating geomorphology, thermochronology, and cosmogenic nuclide-based denudation rate estimates suggest a rapid phase of denudation immediately after the earliest stages of seafloor spreading, and subsequent slow denudation rates since. To constrain the geomorphic evolution of passive margins, we have examined the development of the Sri Lankan escarpment. Cosmogenic nuclide data on river sediment along a north-south transect across the southern escarpment reveal that the landscape is eroding ten times more rapidly in the escarpment zone (26 to 71 mm kyr(-1)) than in the high-elevation plateau above it and in the lowland plain beneath it (2.6 to 6.2 mm kyr(-1)). Unlike these low denudation rate areas, the escarpment denudation is strongly and linearly hill slope-dependent. This shows that denudation and retreat are tightly interlinked within the escarpment, which suggests that the escarpment is evolving by rift-parallel retreat, rather than by escarpment downwearing. Supporting evidence is provided by the morphology of rivers draining the escarpment zone. These have steep bedrock channels which show sharp and prominent knickpoints along their longitudinal profiles. It appears that fluvial processes are driving escarpment retreat, as rivers migrate headwards were they incise into the high-elevation plateau. However, the average catchment-wide denudation rates of the escarpment zone are low compared to the denudation rates that are estimated for constant escarpment retreat since rifting. In common with other escarpments worldwide, causes for this slow down can be tectonic change related to flexural bending of the lithosphere, climate change that would vary the degree of precipitation focused into the escarpment, or the decrease in the contributing catchment area, which would reduce the stream power available for fluvial erosion. (c) 2006 Elsevier B.V. All rights reserved.

[1] Some of the lowest weathering and erosion rates in any mountain range in the world have been measured using cosmogenic nuclides in the steep, humid, tropical highlands of Sri Lanka. The total preanthropogenic denudation rates were measured in creek sediments and soil samples from unperturbed rain forest sites, bedrock from mountain crests, and bedrock from inselbergs. Denudation rates are in the range of 5–30 t km−2 yr−1 (2–11 mm ky−1). These rates average denudation over the last 50–250 ky. Weathering exports in rivers draining the mountainous Central Highlands show that silicate weathering rates are also low, varying from 5 to 20 t km−2 yr−1 today (2–7 mm ky−1), but they represent a significant fraction of the total denudation. All these observations run contrary to the conventional geomorphologic and geochemical wisdom that would predict rapid erosion for highlands of high relief, temperatures, and precipitation. We speculate that the high relief in Sri Lanka represents the remnant of a geomorphic block that was uplifted during rifting at 130 Ma or even earlier and that was reduced to the interior of the island by rapid receding of escarpments after continental breakup. It is possible that throughout this history, hillslopes, where not exposing bare bedrock, were protected by thick weathered profiles. Such clay-rich layers would inhibit silicate weathering by shielding bedrock from weathering agents. In the absence of landscape rejuvenation, physical erosion rates are low, and fresh mineral surfaces are not being supplied. The observation that wet, steep, tropical highlands can have low rates of rock weathering and erosion has some potentially profound implications for the long-term controls of atmospheric CO2 budgets: High temperature and precipitation, which are much invoked though controversial agents for silicate dissolution and CO2 drawdown, become ineffective in promoting weathering in areas that are not tectonically active.

We quantify the difference between the human-caused sediment yield and the natural rates of soil production and bedrock erosion in a now largely deforested tropical highland. The present-day rate of soil loss in the Upper Mahaweli catchment, Sri Lanka, is calculated by using suspended river-load fluxes. These data provide spatially averaged sediment yields of 130-2100 t-km - 2 .yr - 1 . Local rates of soil loss from agricultural plots on hillslopes are as high as 7000 t.km - 2 .yr - 1 . By comparison, natural rates of sediment generation, as determined by measuring cosmogenic 1 0 Be in quartz from sediments and soils, are only 13-30 t.km - 2 .yr - 1 . The natural rates presented here provide a benchmark against which recent erosion rates, determined by various sediment gauging techniques, can be referenced. In the Sri Lankan highlands, these results suggest that soil is now being lost 10-100 times faster from agriculturally utilized areas than it is being produced.

A number of multipurpose reservoirs have been impounded in the Upper Mahaweli Catchment (UMC) mainly to generate hydropower, which contributes to about 40-50% of the total hydropower production in the country. The natural forest cover of the UMC has been gradually decreased during the last two centuries due to the large-scale deforestation for plantation agriculture. Subsequently, the forest cover in the hill country has been reduced further to a few isolated patches in order to meet the demand for agriculture, developmental activities and human settlements. Today, vegetables are grown extensively on the steep slopes of the UMC without proper land management practices. Consequently, this agriculturally active land is exposed to severe soil erosion and landslides, in parallel with the rapid rate of deforestation. Estimates of the rate of soil loss on hill slopes and sediment yields in the fluvial system of the UMC indicate that the human-mediated activities in the UMC have increased rates of ongoing erosion by > 100 times over the background rates of natural erosion. Thus, the river Mahaweli and its tributaries carry enormous amounts of sediments during the rainy seasons, both as bed and suspension loads downstream and large amounts are deposited in the multipurpose reservoirs leading to a reduction of their storage capacities. Siltation in hydropower reservoirs could pose a serious threat, leading to the reduction of hydropower generation in the country and a worsening situation could be anticipated in the future. Therefore, proper management of the land by minimizing the in-situ erosion and preserving the detached soil within the catchments through effective conservation measures need to be implemented in order to ensure a steady and continuous power supply from the hydropower resources.