Soil Erosion Mapping of Vulnerable Agricultural Areas
Development of an ArcGIS Toolbox for Soil Erosion Estimation to map Vulnerable Agricultural Land
Introduction
Soil erosion is a multifaceted and predominant global land degradation process which leads to decline in ecosystem services and functions. Globally, the main cause of soil deterioration is water-induced erosion. Yesuoh, A. & Dagnew, A. (2019), reported that farmland degradation is highly due to soil erosion. Thus, researches on soil erosion and its impact on agricultural lands is needed.
Soil erosion due to trampling and overgrazing. (Kulikov, Maksim & Schickhoff, Udo & Borchardt, Peter. 2016).
Despite knowing that the soil erosion is significant, there is only limited information on soil erosion in Kyrgyzstan (Duulatov, E. et.al. 2021). This is primarily because soil erosion is problematic and costly to measure. Additionally, the vast and diverse landscape, including variations in soil and vegetation cover, climate, landscape features, and land use practices, further complicate the study of soil erosion.
One of the most sensitive types of soil degradation in the country is water erosion. Over two-thirds of arable land faces wind and water erosion, worsened by irrigation erosion (Dzhunushbaev, 1990) .
Protecting soil and practicing sustainability are crucial for the country's economic and social progress.
Objectives
This study mainly aims to developed an ArcGIS Toolbox tailored for soil erosion estimation, specifically RUSLE, to assess risk of agricultural lands. Specifically,
- To spatially predict an annual soil erosion rate across watershed in Kyrgyzstan using the widely recognized RUSLE approach;
- To determine the agricultural areas that are at high risk level of soil erosion; and
- To develop a RUSLE toolbox using various geoprocessing tools in a model builder accessible within ArcGIS Pro.
Significance
This GIS-based toolkit is developed to do spatial analysis and to find erosion-prone agricultural areas and understand erosion risks better.
By developing a user-friendly toolkit, researchers can access and utilize erosion data easily, facilitating informed decision-making and sustainable land management practices.
The toolkit was designed as an ArcGIS add-in that enhances our ability to understand, manage, and mitigate the impacts of erosion on the environment.
RUSLE Model
Revised Universal Soil Loss Equation (RUSLE) is an updated form of the USLE model that predicts long-term, average-annual erosion by water for a broad range of farming, conservation, mining, construction, and forestry uses.
RUSLE is developed and are maintained principally by the USDA-Agricultural Research Service (ARS), the USDA-Natural Resources Conservation Service (NRCS), and the University of Tennessee in the early 1990's.
RUSLE uses 5 main parameters to determine the estimated average annual erosion (rainfall pattern, type of soil, topography, crop management, and conservation practices) (Renard et al. 1997) . The parameters are independently prepared using Geographic Information Systems (GIS) and Remote Sensing (RS) and combined to make a yearly soil loss map, which calculates erosion using the Equation by Wischmeier and Smith, 1965:
A = R × K × LS × C × P
where A = average annual soil loss (ton/ha/year), R = rainfall erosivity parameter (MJ/ha/mm/year), K = soil erodibility parameter (ton/MJ/mm), LS = slope length and topographic parameter, C = crop management parameter, and P = conservation practices parameter.
RUSLE Workflow Diagram (Hagras, A. 2022)
Materials
Data | Factor | Resolution | Source | Web |
|---|---|---|---|---|
Rainfall | R | 30 m | Mean Annual Precipitation from Weather Stations | Kyrgyz Hydro-Meteorological Agency (Ryskal 2020) |
Soil | K | 30 m | Global Soil Data from FAO | |
DEM (ASTER) | LS | 30 m | Earthdata NASA | |
Vegetative cover (NDVI) | C | 10 m | Sentinel 2 | |
Land cover | P | 30 m | ESA Worldcover | |
Cropland | 30 m | Global Food Security-Support Analysis Data at 30 m (GFSAD30). |
Methodology
Watershed Delineation
Another GIS toolkit was developed for watershed delineation for easy processing following the diagram below:
Delineation of watershed in Kyrgyzstan using the DEM (SRTM) with 30m resolution.
- Download the ASTER DEM
2. MOSAIC the DEM using the tool "Mosaic to New Raster".
Make sure that the properties of images didn’t change
3. Prepare the hydrologically conditioned DEM.
Use the "Fill" tool.
4. Compute for the Flow Direction using the "Flow Direction" tool .
5. Compute for the "Flow Accumulation" and "CON" for conditional raster to .
6. Determine/Designate the Pour Point based from stream outlets.
Make sure that the pour point snaps with the flow direction and accumulation. Use the "Snap Pour Points" tool.
7. By running the "Watershed" Tool, it will delineate the watershed based from the elevation, water flow and pour points.
¨The Watershed Delineation Toolbox Kit developed in Model Builder with all the properties.
The interface of Watershed Delineation Tool when you use it.
Preparation of Data Inputs
RAINFALL EROSIVITY FACTOR (R)
Rainfall erosivity was computed for all 35 weather stations, and then the R factor values were estimated through inverse distance weighted (IDW) interpolation using the Spatial Analyst Tools available in ArcGIS Pro.
Formula used for Rainfall erosivity:
R = 0.04830 × P1.61 (2) , where P < 850mm,
R = 587.8 − 1.219 × P2 (3) , whereP ≥ 850mm
From the mean annual rainfall excel file, the Rainfall erosivity is computed using the formula given above.
Using the ArcGIS Pro, weather stations are plotted.
The R factor are mapped using the "IDW" tool - a type of geostatistical interpolation to determine representative rainfall values for areas that it doesn't have values.
SOIL ERODIBILITY (K)
Soil erodibility refers to the soil's susceptibility to being displaced by raindrops and surface water flow. The soil data used in this study were obtained from the FAO World Soil Data and the erodibility factor is determined based from USDA Soil Texture Triangle.
The soil data in shapefile format can be downloaded in the FAO world soil database.
To minimize the processing time, the soil is clipped based from the watershed delineated.
Use the "clip" tool to do this.
With the given soil type, determine the soil texture with its corresponding erodibility factor using the Soil texture pyramid.
Input the K values in the attribute table.
Convert the vector file to Raster using the "Polygon to Raster" tool.
Slope length and steepness (LS) factor
The LS factor represents how topography affects soil erosion (Morgan et al., 1984) . This factor was calculated using data from the ASTER DEM with a 30-meter resolution. The calculation of the slope length factor followed the approach by Moore & Burch (1986).
LS = (A s / 22.13) 0.4 x (sinB / 0.0896) 1.4
where A s is the specific flow accumulation and B is the slope in radians.
Since the Flow accumulation is already produced while delineating the watershed, the slope map in radiance is prepared.
Use the "Slope" Tool to produce slope map. Then use the "Raster Calculator" tool to convert to radiance using the formula in the photo.
Using the Flow Accumulation and Slope, the slope length and steepness factor is calculated as in the equation by Moore & Burch.
Cover management (C) factor
The C factor is derived from Remote Sensing (RS) data is based from the Normalized Difference Vegetation Index (NDVI). This evaluates soil loss specific to different types of land cover. The NDVI is defined in equation:
However, an equation was used to obtain the cover management factor based from Almagro et al. 2019).
Using the Google Earth Engine (GEE), the NDVI is extracted from the Sentinel 2 images dated 2023.
The NDVI is downloaded and further processed in ArcGIS Pro.
The cover factor is computed using the formula given and it was done in "Raster Calculator" Tool.
Conservation practice (P) factor
The conservation practice factor is the relation of soil loss with accurate crop support versus the ratio of straight up and downhill farming (Renard et al. 1997). Below is the reference for the P factor values based from Land Use-Land Cover.
Using the Google Earth Engine, the landcover was downloaded.
Use the "Reclassification" tool to input the P factor values for each land cover type.
The reclassified P factor map.
RUSLE Modeling
Using the Raster Calculator, the Soil Erosion for the watershed is computed following the RUSLE formula.
The Soil Loss Map of the 2 watersheds traversing the Naryn, Jalal-Abad and Osh Region.
The minimum value of Soil Loss is 0 ton/ha/yr, while the maximum value is 150 ton/ha/yr.
The RUSLE Toolbox developed using Model Builder.
The interface of RUSLE Tool when you use it along the properties.
Soil Erosion High Risk Agricultural Areas
After downloading the cropland area, "Reclassification" is done to remove the non-cropland area.
Resampling is also done so the cell size has the same resolution with the soil erosion loss map.
Reclassification is also done to the soil erosion. The value 1 is given to areas with soil loss of above 30 ton/ha/yr.
To determine agricultural areas that has high risk in soil erosion, the Raster Calculator is used. By multiplying the areas with high soil erosion and cropland areas.
Results
Using the generated soil loss raster, a basic spatial analysis is performed to identify the areas within the watershed where agricultural lands are most vulnerable to soil erosion.
The analysis showed that approximately a total of 22,938.42 hectares of lands in Naryn Watershed is prone to high and very high soil loss rates in the watershed.
Soil Loss Rate | No. of Pixels | Area (hectares) |
|---|---|---|
Very Low | 251, 702, 566 | 7,188, 877.0 |
Low | 3, 680, 219 | 105, 110.7 |
Moderate | 1, 248, 623 | 35, 661.9 |
High | 802, 934 | 22, 932.6 |
Very High | 204 | 5.8 |
Table 1. The Soil Loss Rate Area in Naryn Watershed, Kyrgyzstan.
The total area of agricultural land that is affected with soil erosion is 907, 444. 47 hectares or just only about 11% of the total area of the watershed. Although it's just this minimal percent area, it will still affect the food production capacity of the region.
Soil Loss Rate | No. Of Pixels | Area (hectares) | % Area |
|---|---|---|---|
Low | 31, 197, 339 | 894, 900.9 | 10.323 |
Moderate | 429, 751 | 12, 237.48 | 0.1422 |
High | 7, 533 | 216 | 0.0025 |
Table 2. Area of each Soil erosion rate with its % area.
However, this result is affected by the inconsistencies of the temporal data used for each factors.
Conclusion
This study shows that an intricate soil erosion model can be automated using model builder compose of different ArcGIS tools for easier analysis of any watershed. ArcGIS Pro provides all the necessary geoprocessing tools for spatially modeling soil erosion using the RUSLE method. This method depends on various data sources, such as meteorological, geological, topographic, agricultural, and remote sensing data, to have a relevant result.
More over, this study also provided the agricultural areas that is high vulnerability on soil erosion. Using the simple overlay analysis and calculate geometry, an estimation of area (hectares) can be done for each watershed and entity.
REFERENCES
Ahmari, Habib, Matthew Pebworth, Saman Baharvand, Subhas Kandel, and Xinbao Yu. 2022. “Development of an ArcGIS-Pro Toolkit for Assessing the Effects of Bridge Construction on Overland Soil Erosion.” Land 11 (9): 1586. https://doi.org/10.3390/land11091586.
Duulatov, Eldiiar, Quoc Bao Pham, Salamat Alamanov, Rustam Orozbaev, Gulnura Issanova, and Talant Asankulov. 2021. “Assessing the Potential of Soil Erosion in Kyrgyzstan Based on RUSLE, Integrated with Remote Sensing.” Environmental Earth Sciences 80 (18): 658. https://doi.org/10.1007/s12665-021-09943-6.
Kulikov, Maksim & Schickhoff, Udo & Borchardt, Peter. 2016). Spatial and seasonal dynamics of soil loss ratio in mountain rangelands of south-western Kyrgyzstan. Journal of Mountain Science. 13. 316-329. 10.1007/s11629-014-3393-6.
Li, Leilei, Jintao Yang, and Jin Wu. 2019. “A Method of Watershed Delineation for Flat Terrain Using Sentinel-2A Imagery and DEM: A Case Study of the Taihu Basin.” ISPRS International Journal of Geo-Information 8 (12): 528. https://doi.org/10.3390/ijgi8120528.
Rosskopf, Carmen Maria, Erika Di Iorio, Luana Circelli, Claudio Colombo, and Pietro P.C. Aucelli. 2020. “Assessing Spatial Variability and Erosion Susceptibility of Soils in Hilly Agricultural Areas in Southern Italy.” International Soil and Water Conservation Research 8 (4): 354–62. https://doi.org/10.1016/j.iswcr.2020.09.005.
Yesuph, Asnake Yimam, and Amare Bantider Dagnew. 2019. “Soil Erosion Mapping and Severity Analysis Based on RUSLE Model and Local Perception in the Beshillo Catchment of the Blue Nile Basin, Ethiopia.” Environmental Systems Research 8 (1): 17. https://doi.org/10.1186/s40068-019-0145-1.
