This resulted in an estimated soil erosion by water in global cropland of Since RUSLE models do not include a description of gully and tillage erosion processes, and also do not represent other geomorphic processes such as landslides and river bank erosion, it is reasonable to assume that their estimates fall into the lower end of the The good correspondence of our results without using constraining factors with regional estimates US and Europe and Doetterl et al.
The estimates reported in this study rest on RUSLE, a deterministic and empirical-based model which was developed based on a statistical analysis of more than 10, plot-years of basic runoff and soil loss data 44 in 49 US locations covering a large variety of landscape conditions. Although RUSLE-based models are derived from the most comprehensive set of measurements available 45 including universally recognized factors that affect soil erosion by water 29 , 33 , they are predominantly built upon parameters that result from experiments conducted in the United States The application to a non-plot-level and in areas outside the range of the original estimates e.
The authors recognize that using an empirical-based prediction tool outside the original range of environmental variables could represent a legitimate concern Considering the proven capacity of RUSLE-based models to overcome their empirical origin 47 , the current lack of better performing models 9 , 31 , and the need for predicting the possible impacts of global change upon soil erosion 27 , the authors argue that at this stage the presented global RUSLE application represents a legitimate approach to narrow the current gap of knowledge and support the targeted soil conservation efforts aiming to mitigate soil erosion.
Given the quantitative and harmonized nature of the data set, there seemed to be no reasons to doubt the consistency between the estimates for the two time periods as well as the reliability of the resulting national trends. The difference obtained from the comparison of the estimates for the two time periods was driven by land use change and was unaffected by the predictive limits of the empirical soil erosion model. Validity, i. As observed by Auerswald et al. Therefore, a cross-comparison of the modelling results to gain insights on the validity of the modelling predictions was performed.
This operation shows that the modelling results are consistent with both empirical observations and other regional soil erosion assessments.
The analysis at meta-data level confirms that our estimates fall in the range of measured data collected by Montgomery 8 , as well that the global model can describe the magnitude of soil erosion incurring between the different land cover types. Adapting the figure he created Fig. In addition, an exiguous deviation was observed from the comparison between our estimates and the ones provided by independent studies of the US Department of Agriculture USDA for cropland in the United States 30 4.
The good agreement between our estimates and the ones provided by independent studies give confidence that the quantitative estimates achieved through the global model are reliable and valid to a level close to these higher resolution regional assessments Supplementary Note 5.
On the basis of a sensitivity analysis Supplementary Note 3 , the authors observed that the soil erosion predictions of the global RUSLE-based model were most sensitive to the cover-management factor C-factor Supplementary Figs.
This supports the hypothesis that a thorough definition in the C-factor of the land uses and their changes, the extent, types and the spatial distribution of the global croplands and cropping systems are the key to improve RUSLE-based global assessments. In addition, the sensitivity analysis allows us to define the influence of the input parameters on the global RUSLE predictions and spatially map their effects on the model output Supplementary Fig.
Comparison of measured and modelled erosion rates. Large parts of the measured data come from the study of Montgomery 8 integrated with data from other meta-analysis studies. The vertical red line indicates average value of soil erosion. The red dots refer to averages soil erosion rates modelled for two country highly susceptible to water erosion Haiti and Rwanda.
The map of uncertainty is presented in Supplementary Fig. The map gives an outline of the geographical distribution of the prediction variance, and it can be used to compare the potential error in different areas of the world. The authors recognise that the modelling based on data-driven assumptions has its limitations, and there is a need for field monitoring and local scale process-based modelling. However, in light of the useful insights gained from the operations of the model performance evaluation, we argue that the presented RUSLE-based global approach constitutes a powerful assessment tool for identifying hotspots and areas of concern at the global scale.
Moreover, in the proposed new form, the global scale RUSLE-based assessment is brought to a new level, as for the first time it links to the key parameters required to assess the effects of global change and support conservation planning and land management.
Hereinafter, we therefore discuss the implications of our global modelling from a multidisciplinary perspective linking the findings of our map to GDP measures to identify potential pressures on food production systems, risks of increased food and feed prices due to phosphorous shortages, the global soil organic carbon SOC pool that forms the basis for emission levels and climate change analyses, the economic costs of soil erosion and the overall implications for policy decision-making and sustainable development goals.
The increasing population places greater pressure on global food production systems. The global spatial coverage of modelled soil erosion enables us to explore the relationship between the average soil erosion in croplands and the GDP of each country based on World Bank figures. The results presented in Fig.
Clearly, the wealthy countries in temperate latitudes have the least erosion with poorest tropical countries being the most susceptible to high levels of soil erosion. The countries that can least afford soil protection measures are the most vulnerable. This emphasises the importance of soil protection in the sustainable development goals if there is any hope of intensifying agriculture in these countries to meet the food needs of the populations.
Soil erosion in cropland areas on a country basis vs. The size of the circle represents the latitude indicating the higher latitude, wealthier countries are the least impacted by soil erosion, either through more favourable climatic conditions, or soil erosion prevention measures.
Along with the loss of fertile soil through erosion as quantified in this study goes the imminent threat of limited nutrient resources. In , clean phosphorous reserves were predicted to run out in only 20—50 years 49 , With continuously rising demand due to the increasing world population and thus higher demand of food in general and livestock products in particular, fertilizer prices are likely to increase.
This may encourage companies to explore new reserves from lower grade rock which is subject to a higher cadmium pollution and exacerbate the conflict with the less and least developed countries due to food and feed shortages. Both developments could impede sustainable soil use and the application of soil conservation management practices even further.
The former development may increase the cadmium pollution of soils and potentially restrict soil usage. The latter, in turn, could lead to an even more intensive land use with negative effects on soil erosion rates.
No substitutes exist for phosphates with the exception of organic farming using manure which is limited by livestock availability in many countries. In this regard, one promising but not yet widely discussed approach could be to protect phosphorous by reducing soil erosion rates.
Cordell et al. Our global soil erosion assessment highlights the areas where agricultural management based on sustainable farming practices with low soil erosion and high phosphorous recycling rates could be most effectively applied to help keep global food and feed prices at reasonable levels. The prediction of the global soil organic carbon SOC pool by Earth-system models is still one of the main sources of uncertainty, undermining the confidence in the carbon C budget and its future projections 52 , Poor representation of different mechanisms driving SOC turnover and low accuracy of soil data inputs are among the primary causes of this uncertainty.
Land-C-atmosphere feedbacks may not be properly disentangled as long as relevant missing processes are not implemented. Among those, soil erosion is certainly a key process as it displaces consistent amounts of C as lateral fluxes, then subjects it to different environmental conditions that control its stabilisation and release. In this respect, the new global soil erosion assessment presented in this paper has the potential to become a reference input for integrating later C fluxes into large-scale model frameworks of different complexity Combining the global soil erosion with a recent SOC map Supplementary Methods , we estimated a gross SOC displacement by soil water erosion on the order of 2.
Geographically Supplementary Fig. Several on-site effects of soil erosion which occur directly at the site where the soil is removed have been cited in the literature Increasingly, scientists also mention offsite effects of soil erosion in the surrounding areas 55 , Given these on- and offsite effects, soil erosion assessments are highly relevant from an economic point of view because erosion is associated with an unequal distribution of economic costs.
It degrades soil and thereby reduces soil productivity and yields of the land due to the loss of fertility and water storage capacity 22 , Land users in economies which can afford it, therefore have an incentive to use fertilizers or water management practices which unfold immediate effects to compensate the negative effects of soil erosion on their short-term yields. In economies that cannot afford these measures, the price of erosion is paid for by reduced food or forest production The on-site costs of soil erosion are thus internalized by pricing them directly into the economic decision making of the land user.
As long as cash flows remain positive, land users, by themselves, thus hardly have an incentive to adopt land conservation practices to contain soil erosion because the costs of adopting these technologies are high and occur immediately reducing early cash flows , while the benefits spread out over a longer time horizon 55 and are discounted more heavily.
From an economic perspective, the high resolution global soil erosion assessment could help to estimate the global costs of these on-site effects especially from a long-term land value perspective.
More precisely, the monetary costs of fertilizers, additional water management, etc. Soil erosion, however, also affects the surrounding areas with off-site costs that are external to the future cash flow calculus of the land user. Here the price mechanism breaks down in that the offsite costs caused by sedimentation, siltation, eutrophication, flooding, etc. While in the short-term, the exploitation of soil resources may be economically sound for land users, this may not be the case from a socio-economic perspective because the society bears the offsite costs e.
In the case of off-site effects, the global assessment of soil erosion is especially beneficial as it provides the long-awaited basis for an economic assessment of the off-site costs of soil erosion on a global scale. An economic assessment, similar to Pimentel et al. Placing a value on the on- and off-site effects may help decision-makers to internalize the off-site costs into the investment calculus of the land user, e.
This is important because fertile soils on the planet are limited and essentially non-renewable, at least on human time scales. It is therefore of crucial importance to protect available soil resources from further degradation if society wants to maintain this precious natural resource for future generations.
In this sense preserving soil quality and achieving a land degradation neutral world have been explicitly recognized in the recently approved sustainable development goals SDG.
Goal 2 explicitly mentions the relevance of maintaining soil quality for achieving food security while Goal 15 calls for a land degradation neutral world by These goals can only be achieved if we are able to limit current soil erosion rates by applying sustainable soil management practices especially in the areas mostly affected by erosion processes. Moreover, the results of this study can also be relevant for Goal 13 aiming at taking action to combat climate change and its impacts 12 , 13 and Goal 6 to ensure availability and sustainable management of water.
The recently endorsed FAO Voluntary Guidelines for Sustainable Soil Management 58 provide the necessary guidance to National governments on the way forward in order to achieve such ambitious goals by The insights of our high resolution global modelling approach can provide a solid starting point to support decision-making in both ex-ante and ex-post policy evaluation, while scientifically, it can enable better estimates of the global SOC pool including the effects of land use change and conservation agriculture.
Our findings can also provide the basis to test the possible effects of the four per mil initiative proposed by the French authorities during the COP21 in Paris. The total modelled area is about million km 2 , providing living space for a population of about 7. The years and form the reference periods to assess the 21st century human-induced soil erosion by water erosion at a global scale. Permanent loss and gain of global croplands, forests and semi-natural grass vegetation are considered in the modelling scheme while the effects of grazing and the establishment of new pasturelands are implicitly reflected.
Climate change and human-induced effects on climate are also not considered. Consistently with the predictive capacity of the model, soil displacement due to processes such as gullying and tillage erosion is not estimated. RUSLE-type models have demonstrated to be able to reduce a very complex system to a quite simple one for the purposes of erosion prediction 9 while maintaining a thorough representation of the main environmental and anthropogenic factors that influence the process Conceptually, these models follow the same principle of complex process-based models, with a driving force erosivity of the climate , a resistance term erodibility of the soil and the other factors representing the farming choice, i.
Field- and catchment-scale experiments that compared the prediction capacity of empirical RUSLE-type models with process-based models did not reveal a substantial difference between the predictive performances of the two modelling approaches 9 , This, together with the moderate data demand of RUSLE-type models has facilitated the process of upscaling. Today, an extensive amount of the literature recommends the use of RUSLE-based models to provide spatially explicit estimations of soil erosion in GIS environments.
In the original version of RUSLE 33 , the region of the model simulation is a specific field plot slope with given size, rainfall pattern, soil conditions, topography, crop system and management practices. Before the introduction of GIS-based computational techniques, the input data employed to run the model were generally directly measured in the field and afterwards imported in a specific software.
For this purpose, the approach paved by previous GIS-based studies dealing with upscaling procedures to extent the applicability of the model as well as regional studies in Europe 59 , Australia 60 and China 61 was followed.
The rates of soil displacement by water erosion are estimated through the GIS raster scheme. Using a GIS raster scheme applied to the USLE model means hypothesizing that each cell is independent from the others with respect to soil erosion. Soil erosion synonym to RUSLE soil loss, Supplementary Note 2 refers to the amount of sediment that reaches the end of a specified area cell on a hillslope that experiences loss of soil by water erosion.
The modelled area does not, in any way, include areas of the slope that experience net deposition over the long term. Conceptually, the global modelling presented in this study is based on the assumption that if a catchment- or a regional-scale application of RUSLE can predict meaningful soil loss estimates, the application of the model at larger-scale will provide meaningful estimates as long as the scale of the input data employed is congruent with the scale used for the estimation of the modelling factors and local applications.
If the scale of the input data meets these requisites, the global scope of the model application represents a consistent repetition in space of the calculation for all cells in the modelled area. For the analysis of the effect of the topography a SRTM 3 arc-seconds ca. This ca. Rainfall and soil characteristics were obtained using the best database available to adequately describe their pattern and dynamics during the elaboration of this study. Both have a spatial resolution of ca.
For processing the main model components, the spatiotemporal variations of the land cover and management C-factor as well as the rainfall erosivity R-factor factors were assessed through new methodologies exclusively designed for this study while the K, LS and P-factor were derived from methods reported in literature Supplementary Methods. The C-factor measures the combined effect of the interrelated cover and management variables on the soil erosion process For this global assessment Supplementary Fig.
We used steps for computing the C-factor in agricultural and non-agricultural land. The IGBP scheme reports seventeen land cover classes Supplementary Table 1 , including three non-vegetated land classes, three developed and mosaicked land classes and eleven natural vegetation classes.
Eleven classes were categorized as non-agricultural land and four classes were precluded from the analysis because they were not relevant for our soil erosion context i.
Friedl personal communication , however, strongly advised against mapping changes by taking simple difference across years due to the overall accuracy across all land classes of ca. Following this advice, the subsequent step of data processing for the MCD12Q1 global land cover maps was modified in order to better represent the forestland area by considering the well-known inability of the MCD12Q1 product to accurately detect forestland in some locations 67 and align the national cropland surface with the more accurate estimations reported by the Food and Agriculture Organization FAO.
The global forestland area for the years and was outlined using the high-resolution global forest change maps derived by Hansen et al. According to Hansen et al. Discrepancies between Hansen et al. The FAO-based cropland surface arable land and permanent crop area for the years and was obtained by calculating the median of the FAOSTAT data for the triennial periods — and — The final map for the year was obtained by substituting Hansen et al.
For the countries that experienced a reduction of the cropland surface during the observed period, the number of cells equal to the cropland decrease which presented the lowest classification confidence in the IGBP class of the MODIS MCD12Q1 were removed from the land cover map of To outline the cropland maps for the two reference periods, twelve-year crop statistical data harvested area were used to statistically describe the national crop conditions and to estimate the C-factor C CROP.
A set of crops were considered 15 and subsequently grouped in 14 crop classes according to their soil cover effectiveness Supplementary Data 1 and Supplementary Table 2. World patterns of rainfall erosivity R-factor Supplementary Fig. The spatially continuous R-factor map at 30 arc-seconds, ca.
We used the algebraic approximation reported by Wischmeier and Smith 63 and Renard et al. The soil properties i. Further soil properties such as soil structure and permeability were derived according to the methodology proposed for the soil erodibility map of Europe We computed the topographical factor LS-factor Supplementary Fig.
Hole-filled SRTM 3 arc-seconds ca. In the main modelling run the P-factor was assumed to remain constant at 1 since suitable spatial information for all the considered countries were not available. The soil erosion rates of the baseline scenarios were predicted for the reference years and without considering conservation cropping and management practices.
We rested this decision on the current lack of adequate standardised and harmonised worldwide spatial information. Besides the baseline scenario, a conservative scenario was estimated for the 54 countries which reported information about the implementation of conservation agriculture to the FAO. These countries regularly submit the proportion of their cropland managed in accordance with the three FAO conservation agriculture standards i. As suggested by Wischmeier and Smith 63 , practices of improved tillage like no-till and cover crops were considered as conservation cropping and management practices and implemented in the C-factor.
This second conservation prediction refers to the maximum technical potential reduction which we use to represent the negative variation in our conservation scenario. In order to evaluate the performance of the global model and lay the groundwork for future studies to better identify areas prone to soil erosion by water and their environmental and socio-economic implications, multiple operations aiming to obtain insights about the validity of the modelling predictions were performed.
These consisted of the following: a comparison of the estimates with the ones provided by independent RUSLE-based assessments in the US and Europe, an analysis at meta-data level to observe if the estimates fell in the range of measured data, and a comparison of the spatial patterns of soil erosion predicted by the global model for the years baseline scenario with the ones reported by the expert-based Global Assessment of Human-induced Soil Degradation GLASOD.
A further comparison was made using the land degradation trends reported by the UN project Global Assessment of Land Degradation and Improvement GLADA , an assessment based on remote sensing time series analyses of the normalised difference vegetation index NDVI for the period — In a final comparison, the measured soil erosion rates from locations across the word were superimposed to the estimates of the global model Supplementary Fig.
The authors declare that all other data supporting the findings of this study are available within the article and its Supplementary Information files, or are available from the corresponding author upon reasonable request. Amundson, R. Soil and human security in the 21st century. Science , — Article Google Scholar. Robinson, D. Soil natural capital in Europe; a framework for state and change assessment.
Keesstra, S. The significance of soils and soil science towards realization of the United Nations sustainable development goals.
Soil 2 , — Oldeman, L. Pimentel, D. The answer is that Antarctica is a unique terrestrial environment: most of the continent is a desert; more specifically, a cold, polar desert. Although Antarctica is covered by a vast amount of ice, it is extremely arid with very little liquid water at the surface. The MDV is often referred to as the coldest, driest, and windiest place on the planet, and thus represents a climatic extreme where we find natural processes not observed anywhere else on Earth.
In fact, the environment is so extreme that it is often used as an analogue for conditions on the surface of Mars. Its high elevation and inland distance from the Ross Sea coast make this location one of the coldest and driest in the already cold and dry MDV. Mullins Glacier, a debris-covered glacier, flows from Mullins Valley into Beacon Valley and is covered with dolerite and to a lesser extent sandstone boulders and sediment.
When soil is left bare, the fine, carbon- and nutrient- rich fractions are removed first, altering physical, chemical and biological soil properties, such as soil albedo, temperature, evapotranspiration, water holding capacity and soil biodiversity at a variety of scales, which disrupt ecosystem functioning. Recent scientific insights, however, place soil degradation, including erosion, in a broader perspective. Soil erosion is not only a biophysical factor but also a feedback component in complex socio-environment systems that disrupts fundamental ecosystem services and the human economic systems that rely on them.
Anthropogenic soil erosion, as caused by improper agricultural practices and overgrazing, accelerates erosion and has repercussions on carbon and phosphorus emissions. Human induced erosion accounts for as much as one third of the carbon emissions that result from land-use change. Moreover, on much of the global farmland, soil is lost at greater rates through erosion than can be replenished through natural soil-forming processes.
By contrast, deposition of eroded sediment can also sequester carbon. Understanding the global feedbacks of soil carbon to climate change is a major challenge. The precise role and impact of soil erosion in terrestrial carbon cycling is uncertain due to lack of harmonised global erosion estimates. Beyond carbon, phosphorus is another essential element in terrestrial ecosystems that is vital for agriculture. It, too, can be dislocated and transported in dust emissions caused by wind erosion.
Recently, attention has been given to the deposition effects that may increase the fertility of weathered soils offsite. Also, dust emissions from cultivated areas may accelerate eutrophication of inland lakes where dust eventually is deposited. Erosion associated with human activities affects land productivity and has economic effects on-site and offsite. On-site costs that directly affect farm or pasture production are mostly absorbed by the land owner.
Off-site costs are mostly borne by society at large through attempts to mitigate their impacts. On-site loss of topsoil is the most severe and affects both short-term and long-term land productivity through losses in fertility, water-holding capacity and changes in soil structure. In smallholder areas, the onsite costs of losing soil nutrients, nitrogen and phosphorous, due to sediment runoff that reduces yields and long-term land productivity can be substantial.
Ultimately, off-site economic effects, such as adverse human and animal health effects, sedimentation of reservoirs, damages to inland infrastructure and ports by dust pollution and eutrophication of surface water, can often be higher than on-site consequences.
Sediment retention in reservoirs not only threatens reservoir longevity, but also may accelerate coastal erosion by reducing sediment flows to the coast.
0コメント