Vulnerability Assessment of Soil Erosion/Deposition in a Himalayan Watershed Using a Remote Sensing and GIS Based Sediment Yield Model

Int.J.Curr.Microbiol.App.Sci (2017) 6(3): 40-56

International Journal of Current Microbiology and Applied Sciences ISSN: 2319-7706 Volume 6 Number 3 (2017) pp. 40-56 Journal homepage: http://www.ijcmas.com

Original Research Article https://doi.org/10.20546/ijcmas.2017.603.004

Vulnerability Assessment of Soil Erosion/Deposition in a Himalayan Watershed using a Remote Sensing and GIS Based Sediment Yield Model

S.S. Rawat1*, M.K. Jain2, K.S. Rawat3, B. Nikam4 and S.K. Mishra5

1National Institute of Hydrology, Roorkee-247 667, Uttarakhand, India 2Department of Hydrology, Indian Institute of Technology-Roorkee, Roorkee-247 667, Uttarakhand, India 3Center for Remote Sensing and Geo-Informatics, Sathyabhama University, Chennai-600 119, India 4Indian Institute of Remote Sensing, ISRO, Dehradun-248 001, Uttarakahnd, India 5Department of Water Resources Development and Management, Indian Institute of Technology-Roorkee, Roorkee-247 667, Uttarakhand, India *Corresponding author

ABSTRACT

Soil erosion strongly affects crop yield, undermines the long term productivity of farm

land and sustainability of farming system, and poses a major threat to the livelihood of the farmers and rural communities. The United Nations Environmental Program reported that the productivity of soil has reduced and resulting in economically unfeasible cultivation on K e yw or ds about 20 million hectare of land each year due to soil erosion and resulting degradation of land. The eroded soil is also a major cause of loss of storage capacity (1 to 2% annual GIS, RS, Soil reduction globally) of multipurpose reservoirs due to sedimentation which affects society erosion, Sediment at large. In developing countries like India, limited data availability constrains the yield, USLE, application of the sophisticated models in proper planning of erosion control measures. In

Transport capacity. the present study, a simple spatially distributed model has been formulated in GIS

Article Info environment for mapping areas vulnerable to soil erosion and deposition in a Himalayan watershed from India. The model discretizes the spatial domain of catchment into

Accepted: homogenous grids/cells to capture the catchment heterogeneity and derived net erosion and 08 February 2017 deposition maps by considering gross soil erosion and transport capacity of each cell. Available Online: Spatial distribution capability of the model has been checked at outlet as well as three 10 March 2017 upstream gauging sites using eight years historical sediment yield data. The model estimated the seasonal sediment yield with less than ±30% errors. Finally, entire watershed

has been classified into six different severity scales of erosion i.e. slight, moderate, high, very high, severe and very sever. Such maps have immense significance to prioritize area specific watershed conservation and management measures.

Introduction

Soil erosion is a serious problem in Asian rivers, and amongst these, Himalayan Himalayas and foothill ecosystem. Eighty rivers are the major contributors (Stoddart, percent of the sediment material delivered to 1969). The Himalayan and Tibetan regions the world’s oceans each year comes from cover only about 5% of the earth’s land

Int.J.Curr.Microbiol.App.Sci (2017) 6(3): 40-56 surface but supply about 25% of the dissolved research due to their complexity and non- load to the world ocean (Raymo and availability of data required for field use. For Ruddiman, 1992). Although, Himalayan example, WEPP model requires as many as region is hydrologically very important as it is 50 input parameters, which can cause the the origin of three world’s major river system problem of equifinality (condition in which viz. the Indus, the Ganges, and the different combination of model parameters Brahmaputra but very few studies have been lead to similar output) (Brazier et al., 2000). reported on rainfall, runoff and its induced Therefore, empirical models are most sediment yield processes in this region. In commonly used in field applications. Himalayan or mountainous regions, the data However, these are based on inductive logic whether on economic, social, or and generally applicable only to those environmental are usually incomplete, when conditions for which the parameters have they exist at all. Therefore, lack of scientific been calibrated. Moreover, there is not a knowledge about this planet’s most complex single model valid for all applications. landscape is main hurdle to plan the developmental agenda for this region. Sediment outflow from a watershed depends on several parameters which are responsible Soil erosion involves the processes of for the soil detachment and the transport detachment, transportation, and accumulation capacity of the path followed by the eroded of soil from land surface due to either impact soil to reach the outlet. However, lack of of raindrop, and shearing force of flowing detailed data with respect to these parameters water. Based on the processes considered, as at river catchment scale is a major constraint well as the complexity, accuracy, scale (space in estimation of sediment yield at catchment and time), and ultimately data requirement, a scale (Van Rompaey et al., 2001). Therefore, wide range of models from simple empirical a simple empirical lumped sediment delivery to complex physically based have been ratio (SDR) approach is used to link the soil developed to model the soil erosion and erosion within a watershed to the sediment subsequent sediment yield (Merritt et al., yield at outlet (Ferro and Minacapilli, 1995). 2003; Aksoy and Kavvas, 2005). Simple and Since, SDR-based sediment yield estimation popular empirical models such as USLE and approach is an empirical lumped approach its derivatives perform very well at the plot (Verstraeten and Poesen, 2001), it performs scale (Risse et al., 1993; Angima et al., 2003). well only if the data of catchment belongs to However, their use at catchment scale could inherent region. Applications of such be challenging. Therefore various physically empirical approach to other catchments are based models such as Water Erosion questionable and require rigorous calibration Prediction Project (WEPP) (Nearing et al., before extension. Also due to lumped nature 1989), Areal Non-point Source Watershed of SDR approach, it is not helpful in Environment Response Simulation prioritization of watershed (ANSWERS) (Beasley et al., 1980), management/treatment activities within a Agricultural Nonpoint Source Pollution river basin/catchment. However, soil erosion Model (AGNPS) (Young et al., 1989), Soil and sediment yield exhibit large spatial and Water Assessment Tool (SWAT) (Arnold variability due to heterogeneity involved in et al., 1993), and many others have been various parameters (catchment physical as developed and these have proved very useful well as climatic) responsible for theirs as research tools. However, the use of occurrence. The technique of Geographical physically based models is mostly limited to Information System (GIS) is well suited for

Int.J.Curr.Microbiol.App.Sci (2017) 6(3): 40-56 quantification of heterogeneity (not only in Kothyari, 2000; Jain and Goel, 2002; Lee, space but also in time) in the topographic, 2004; Jain and Das, 2010). Although USLE is cover type, and drainage features of a a lumped empirical model, this equation has watershed by partitioning the watershed into been a part of several spatially distributed small homogenous grids (Jain et al., 2005; process-based models such as SWAT, Wu et al., 2005; Naik et al., 2009). AGNPS, CREAMS (Knisel, 1980), SEDNET Furthermore, accessibility of data even in (Prosser et al., 2001) and EPIC (Williams, remote areas like Himalayan catchment is the 1983). This is possible due to the major advantage of GIS and remote sensing discretization of heterogeneous catchment (RS) techniques. into small homogeneous grids/cells. In this paper USLE is used for estimation of GSE Keeping in view, limited data availability in within a cell is expressed as: Himalayan catchment and spatial variability and complexity involved in soil erosion and GSEi  RKiLSiCiPi (1) sediment production process, a simple spatially distributed model has been Where, GSEi = gross amount of soil erosion formulated in GIS environment. The model in cell i (MT ha−1 year−1); R = rainfall −1 −1 −1 produces outputs showing spatial variability erosivity factor (MJ mm ha h year ); Ki = due to spatial discretization adopted and, soil erodibility factor in cell i (MT ha h ha−1 −1 −1 therefore, helpful in mapping of sediment MJ mm ); LSi = slope steepness and length source and sink areas and prioritization of factor for cell i (dimensionless); Ci = cover watershed management activities accordingly. management factor (dimensionless) and Pi = However, model structure is kept simple so supporting practice factor for cell i that the application of model can be easily (dimensionless). extended in developing countries like India where availability of data in space and time is Assessment of seasonal sediment transport always a challenge. capacity

Model formulations Sediment transport capacity expresses the erosive power of overland flow or channel The proposed model comprises of three major flow which is responsible for the movement components (1) the assessment of seasonal of eroded soil. In fact, the process of soil gross soil erosion (GSE) for each grid/cell; (2) erosion and deposition is governed by the the assessment of seasonal local transport sediment transport capacity of overland flow capacity (TC) for each grid/cell; and (3) and subsequent channel flow. Therefore, transport limited accumulation algorithm for sediment transport capacity equation is an routing sediment from each of the discretized essential part of all physically based sediment grid/cell to the outlet of the catchment by yield computation models. Several simple to taking into account the local transport complex equations have been developed for capacity of each cell. estimation of transport capacity of different grid/cell within the catchment. One of the Estimation of gross soil erosion most commonly used transport capacity equation as given by Van Rompaey et al., Universal Soil Loss Equation (USLE) has (2001) is as follows: been found to produce realistic estimates of surface erosion over small size areas TC  K RKLS - aS  (2) (Wischmeier and Smith, 1978; Jain and TC IR 42

Int.J.Curr.Microbiol.App.Sci (2017) 6(3): 40-56

Where, TC is the transport capacity (kg m-2 cell. If the local TC is smaller than the -1 year ); KTC is the transport capacity sediment flux, then sediment deposition is coefficient and SIR is the inter-rill slope modeled. If TC is higher than the sediment gradient factor. Van Rompaey et al., (2005) flux, then sediment transport will be supply reported poor performance (R=0.25) while limited. Thus, by introducing the transport using this equation in mountainous part of an capacity, a more realistic representation of Italian catchment. Normally, the topography overland flow in sediment transport can be of hilly areas is such that flow converges at simulated. The model produces different some locations, generally at the junction of maps of erosion, sediment transport, and the steep slope and valley floor (or toe of the sediment deposition rates. For cell-based slope). These locations are generally the end discretization system, transport limited points of the steep slopes or where sudden accumulation can be computed as: flattening in the slopes is observed. As per equation 2 model estimates low values of T  min(GSE  T ,TC ) (4) outi i  ini i transport capacity of theses points due to low values of slopes (S). Hence, most of the D  GSE  T - T (5) eroded sediment will be deposited at theses i i  ini outi points and model underestimates sediment outflow. However, in reality these are intense where Tini = sediment inflow from upstream flow convergent points and hence attributed cells, Touti= sediment outflow from the cell i. to high transport capacity. To overcome this Di = deposition in cell i. problem Verstraeten (2007) suggested to incorporate an upslope contributing area Study area and data used factor in transport capacity equation. After incorporating an upslope contributing area The proposed model has been used to model factor Equation 2 can be re-written as follows: the sediment yield from a hilly watershed named Naula with elevation ranging from 790 TC  K RK S βA γ (3) m to 3088 m and geographically located i TC i i si between 29°44'N and 30°6'20''N latitudes and 79°06'15''E and 79°31'15''E longitudes in the Where, Asi is the upslope contributing area for middle Himalayan region of Uttarakhand state cell i which represents the actual flow of India (Figure 1). Naula watershed is a accumulation value of any cell. Incorporation hydrologically important sub-watershed of of upslope contributing area factor in Ramganga reservoir catchment (3134 square transport capacity equation (Equation 3), km area). Ramganga reservoir, a multipurpose transport capacity of intense flow convergent project (127 meter heigh earth and rockfill areas will be increased significantly even dam) of Government of India built in year sudden fallen in slope (S). 1974, produces approximately 452 MW of electricity and also facilitates irrigation on an Sediment routing using transport limited area of about 5120 square km during non- accumulation monsoon period. Naula watershed covers most of the hilly portion of Ramganga Eroded sediment from each cell follows a catchment (1084 square km drainage area) definite path defined by direction of flowing and is the largest sediment contributor to the water. The amount of sediment outflow from reservoir. The topography of the watershed is one cell to its downstream cell depends on undulating and irregular with slope varying local sediment transport capacity (TC) for a from moderate to steep. Daily rainfall data of 43

Int.J.Curr.Microbiol.App.Sci (2017) 6(3): 40-56 ten raingauge stations located inside/outside Rainfall erosivity (R) of Naula watershed were collected from the Divisional Forest Office (Soil Conservation) Rainfall erosivity or erosion index is one of Ranikhet, Government of Uttarakhand and the important parameters in estimation of used to calculate the weighted rainfall of soil loss from USLE. Ram Babu et al., Naula watershed. There are four sediment (2004) found linear relationship when annual monitoring sites in the Naula watershed. and seasonal erosion index values were Three sediment units i.e., Mehalchauri (166 plotted against annual and seasonal rainfall, square km), Chaukhutia (572 square km) and respectively. On the basis of these linear Naula (1084 square km) are located on the relationships they plotted iso-erodent maps main Ramganga stream. for different zones viz., northern, western, central, eastern, and southern India. In the However, Buda Kedar (295 square km) present study, rainfall erosivity was sediment unit is located on Bino river, a calculated by the relationship developed by tributary of Ramganga river. Daily sediment Ram Babu et al., (2004) for this particular data of four gauging stations, namely Naula, zone of India and presented as: Buda Kedar, Chaukhutia and Mahalchauri were also collected from the same agency. R = 71.9 + 0.361 P (r = 0.91, for 293 ≤ P ≤ 3190) (6)

Generation of input geo-database Where, P is the average seasonal rainfall in mm. In the present study, Equation 6 is used Extraction of drainage network and to compute seasonal values of R-factor by watershed from Digital Elevation Model replacing P with observed seasonal rainfall of (DEM) a particular year.

Conventionally DEM is prepared by Soil erodibility (K) digitization of the contour lines from toposheets of the study area. However, Soil map of the watershed was prepared by digitization of contour lines from toposheets digitization of soil maps available in survey is very painstaking and time consuming report obtained from National Bureau of Soil process, especially when the area of interest is and Landuse Planning (NBSS and LUP, large and hilly (not easy to access). It is better 2004) using ArcGIS. The digitized polygon to use the Shuttle Radar Topographic Mission map of Naula watershed was then rasterized (SRTM) DEMs if cartographic data with scale at 90 m grid cells by using vector to raster above 1:25,000 is not available (Jarvis, 2004). tool. Based on the soil characteristics such as Therefore, in this study, SRTM DEM data of texture, depth, erosion, surface, drainage, 3-arc second (90m X 90m) spatial resolution and slope (NBSS and LUP, 2004), total nine was downloaded from www.landcover. soil mapping units have been identified in org/data/srtm/ and used. the entire study area. Details such as fraction of sand, silt, clay and organic matter and Watershed boundary, drainage pattern and other related parameters information for slope/gradient map, have been extracted from different mapping units are taken from NBSS the DEM with the help of DEM Hydro and LUP (2004) for Naula watershed. K- processing tools available in GIS (ArcGIS values for different soil categories were ver. 9.3). All these maps are required as input calculated according to the procedure given of the model at different stages. by Haan et al., (1994).

Int.J.Curr.Microbiol.App.Sci (2017) 6(3): 40-56

Length-Slope factor (LS) initiation threshold is a numerical value; pixels having flow accumulation value less The combined length-slope (LS) factor in than channel initiation threshold are termed as USLE is a measure of the sediment transport overland flow pixels and those having higher capacity of overland flow and it reflects the flow accumulation value than channel effect of topography on soil erosion. initiation threshold are termed as However, several methods (Wischmeier and channel/stream pixels. The threshold has to be Smith, 1978; Moore and Burch, 1986a, b; chosen in such a way that the total stream Moore and Wilson, 1992; Desmet and length generated using threshold and channel Govers, 1996) have been developed for network seen in satellite data and survey of estimation of LS factor but application of India (SOI) toposheets (digitized in vector these to real landscapes as part of a GIS is a form) are equivalent. Accordingly, a channel major problem. Among these, the one which initiation threshold value of 0.486 square km is best suited for integration with the GIS is is adjudged appropriate to define channel the theoretical relationship proposed by cells for Naula watershed. Moore and Burch (1986a, b) and Moore and Wilson (1992) based on unit stream power Crop management factor (C) theory, given as: To assign the spatial values of crop n m management factor (C), land use map of the  A   sinθ  si i (7) study area was prepared by classification of LSi      22.13 0.0896 LANDSAT TM satellite image using unsupervised and supervised classification in Where, Asi is the specific area at cell i ERDAS image processing software. To defined as the upslope contributing area for discriminate the vegetation from other surface overland grid (Aup) per unit width normal to cover types, the Tasselled Cap flow direction; θi is the slope gradient in Transformations (TCT), Vegetation Index degrees for cell i. (VI), and Water Index (WI) were performed. Overall, six different classes viz., forest, In the original equation of LS factor agriculture, river bed, pasture, water and (Wischmeier and Smith, 1978), slope length settlement are identified in the study area with (λ) is defined as the distance from the point of the help of limited ground truth information, origin of overland flow to the point where Google Earth image and SOI topographic either the slope gradient decreases sufficiently maps. Table 1 represents different land use for deposition to begin or runoff water enters classes that exist in Naula and its a well-defined channel. In the present study subwatersheds. Based on the land cover the slope length (λ) is replaced by the unit categories, the attribute values for the C- upslope contributed area at which formation factor are assigned to individual cells from the of channel is started and taken equal to tabulated values suggested by Wischmeier channel/stream initiation threshold value and Smith (1978), Singh et al., (1992), Haan (ESRI, 1994; Jain and Kothyari, 2000). The et al., (1994). main aim of the replacement of slope length by upslope contributing area is to incorporate Management practice factor (P) the effect of converging and diverging terrain on soil erosion and hydrological aspect of the Based on experimental investigations, values watershed (Moore and Wilson, 1992). The for P-factor have been tabulated for many stream generation threshold or channel management conditions (Haan et al., 1994). 45

Int.J.Curr.Microbiol.App.Sci (2017) 6(3): 40-56

The P-factor is taken equal to 0.7 for However the negative value of ME implies agricultural land as mostly contour cultivation inefficiency of the model in prediction. is followed on agricultural land in the study Relative Root Mean Square Error is estimated area, and unity for other land use/land cover by the following formula: types. P-factor values are added in the attribute field of land use map. 1 n (Y  Y )2 n  obs pred Model applications and discussion of i1 RRMSE  n (9) results 1  Yobs n i1 Generation of gross soil erosion maps Where, n is the number of data points. For The layers of topographic factor (LS), crop wide range of KTC, value of ME (Equation 8) management factor C, soil erodibility factor and RRMSE (Equation 9) have been K, and support practice factor P are overlaid calculated and shown in figure 3. It is evident -5 using ArcGIS. For estimation of gross soil from figure 3 that at KTC=3x10 , ME is the erosion in the watershed for different highest (0.74) and RRMSE is at the lowest years/season, the composite map of KLSCP (0.29). Any change in the value of KTC factor is multiplied by R-factor in raster (=3x10-5) leads to increase and decrease in calculator of ArcGIS. Figure 2 represents RRMSE and ME, respectively. It is worth - gross soil erosion map of Naula watershed for noting that final calibrated KTC value (=3x10 5 the year 1987 as illustration. ) in this case is close to KTC obtained by Verstraeten (2007) for good to moderate Generation of transport capacity maps vegetation. Notably, about 46% area of the study watershed is covered by the forest. Transport capacity of overland flow was Transport capacity maps are generated using calculated for each season and each pixel calibrated KTC value for all years (1979-87). using Equation (3) in ArcGIS. Goodness of fit Transport capacity map for year 1987 is criteria such as Model Efficiency (ME), (Nash presented in figure 4 as illustration. It is and Sutcliffe, 1970) and Relative Root Mean evident from this figure that the ridges and Square Error (RRMSE) have been used to the flattened area near the channel, calibrate the value of KTC appearing in generally cultivated (viz., south-west Equation (3) using five year observed and direction of Chaukhutia gauging site) are the computed sediment outflow data (estimated areas possessing low transport capacity. from Equation 4). Model efficiency (ME) can However, transport capacity is high in be calculated as follows: channel areas and the steep head water areas.

2 Sediment routing considering local  Y  Y  obs pred  transport capacity ME1 2 (8) Yobs  Ymean  Gross soil erosion was estimated using raster Where, Yobs observed seasonal sediment calculator tool of ArcGIS, but the quantity of (tons), Ypred is predicted seasonal sediment eroded soil delivered to the next cell will (tons), Ymean is mean of the observed sediment depend on the transport capacity of the cell. (tons). Value of ME ranges from -∞ to 1. However, at present there is no ready to use Values close to 1 indicate better model fit. tool available in GIS, which estimates the

Int.J.Curr.Microbiol.App.Sci (2017) 6(3): 40-56 quantity of sediment moved in the next cell Investigation of spatial distribution by considering gross soil erosion and capability of model transport capacity of the cell. Therefore using the basic principal of overland flow routing Comparison of sediment yield at outlet of the (Equation 4) a programme was developed in watershed is not sufficient for validation of a Interactive Data Language (IDL). spatially distributed sediment model (Takken et al., 1999) due to highly complex behaviour The developed tool/ programme estimates the of sediment. Therefore, to investigate the sediment moved out from each pixel using spatial distribution capability of the proposed flow direction, flow accumulation, gross soil model, the data of three upstream gauging erosion, and transport capacity maps. Tool sites (Mehalchauri, Buda Kedar, and compares the gross soil erosion (total soil Chaukhutia) are used to check the prediction ready to move out of a particular pixel) and capability of the proposed model. Sediment transport capacity of the flow in that pixel, if yield estimated by the developed model for all transport capacity is equal or greater than these subwatersheds/gauging sites can be read gross soil erosion then entire eroded soil will directly from the corresponding pixel value be transported into the next pixel. from sediment outflow map of Naula watershed. When sediment yields at these The destination of this transported gauging sites were compared with the soil/sediment will be determined using flow corresponding observed values for different direction map. However, if the transport years, it was found that in most cases, the capacity of any pixel is less than total soil model performed very well (74% data points ready to move out of particular pixel, the tool within ± 30% error range, figure 6). However, will assign the difference between transport large error during a particular year may be capacity and the total soil ready to move out, attributed to uncertainties in observations as amount of sediment deposited in that pixel. probably due to temporal variability involved Batch processing option is given in in sediment yield. Furthermore, in the present programme to process temporal data and to study a single vegetation map is used for all save time in repeated operations/processes. years and for entire season, however vegetation has dynamic nature. Finally, Using developed tool, sediment outflow maps considering all data points, the accuracy for different years (1979-87) were produced. obtained is considered well and acceptable Such maps provide the amount of sediment because even the more elaborate process- outflow from the system at every cell and are based soil erosion models are found to useful for determination of sediment flowing produce results with still larger errors (ASCE, out of the watershed at any location. 1975; Foster, 1982; Wu et al., 1993; Wicks and Bathurst, 1996). Figure 5 depicts the sediment outflow map for the year 1987 as illustration. While cell values Generation of vulnerability maps of net of Naula site from theses maps were read for erosion/deposition validation period and compared with observed sediment data, a very low errors have been Using Equation 5, sediment deposition maps observed (26.1%, 9.0%, and 0.2% for 1985, for various years were obtained. Net soil 86, and 87, respectively). Such low errors in erosion/deposition maps of various years were estimation of sediment yield do support the produced by superimposing sediment efficacy of the model. deposition maps over gross erosion map of

Int.J.Curr.Microbiol.App.Sci (2017) 6(3): 40-56 corresponding years. Such maps are helpful in of erosion: Slight (0 to 5 t ha−1 year−1), identifying areas vulnerable to soil erosion Moderate (5 to 10 t ha−1 year−1), High (10 to and sediment deposition in watershed area. 20 t ha−1 year−1), Very High (20 to 40 t ha−1 Figure 7 depicts net soil erosion/sediment year−1), Severe (40 to 80 t ha−1 year−1) and deposition map for year 1987 as illustration. Very Severe (> 80 t ha−1 year−1) as per the As can be seen in figure 7, deposition of guidelines suggested by Singh et al., (1992) sediment resulted near the banks of some of for Indian conditions. The same is also the stream reaches where transport capacity depicted in table 2 and figure 7 for Naula and was low. It is possible to identify the critical its subwatersheds. A significant area (6.4% of areas delivering most of the sediment to the total watershed area) of the Naula watershed river system. Notably, these areas are not is under deposition (Table 2). As can be seen necessarily the same as those producing most from figures 7, most of the deposition of erosion, as most of the eroded sediment is sediment occurred in the sides of the drainage deposited within the catchment, before channels when they reach in valleys and also reaching the river system. The net erosion on flatter land areas found in the cultivated estimated on a cell basis for the watershed is valley lands. grouped into the following scales of severity

Table.1 Land use pattern in Naula and its subwatersheds

% of watershed area Landuse Naula Chaukhutia Buda Kedar Mehalchauri Forest 46.44 59.29 33.93 50.64 Agriculture 23.07 14.65 33.53 20.64 River bed 6.37 5.57 7.06 3.15 Pasture 22.91 19.51 24.05 24.18 Water 0.2 0.11 0.19 0.04 Settlement 1.01 0.86 1.23 1.35

Table.2 Area (%) under different soil erosion/deposition classes in Naula and its sub-watersheds

Erosion/Deposition Classes Watershed Deposition Slight Moderate High Very High Severe Very Severe Naula 6.4 84.7 3.0 1.8 1.4 1.3 1.5 Buda Kedar 7.0 82.4 3.7 2.2 1.4 1.6 1.7 Chaukhutia 5.4 86.2 2.8 1.7 1.3 1.1 1.6 Mehalchauri 4.8 89.1 1.7 1.1 1.4 1.1 0.7

Int.J.Curr.Microbiol.App.Sci (2017) 6(3): 40-56

Figure.1 Location map of Naula watershed

Figure.2 Gross soil erosion map of Naula watershed

Int.J.Curr.Microbiol.App.Sci (2017) 6(3): 40-56

Figure.3 Calibration of KTC for Naula watershed using seasonal rainfall-sediment yield data

1.00 1.40 RR M… 0.00 1.20

-1.00 1.00

-2.00 0.80

RRMSE -3.00 0.60

0.40 -4.00

0.20 -5.00

0.00 -6.00 0.0 1.0 2.0 3.0 4.0 5.0 6.0 7.0 8.0 9.0 10.011.0 -5 KTC (x 10 )

Figure.4 Transport capacity map of Naula watershed

Int.J.Curr.Microbiol.App.Sci (2017) 6(3): 40-56

Figure.5 Sediment outflow map of Naula watershed

Figure.6 Comparison between observed and predicted values of sediment outflow at different gauging sites

Int.J.Curr.Microbiol.App.Sci (2017) 6(3): 40-56

Figure.7 Net erosion/deposition map of Naula watershed

It was estimetd that soil loss in 9.0% area of net soil erosion as illustrated in figure 7 can Naula watershed is above soil loss tolerence be of immense significance in deciding the limit. Notable that soil tolerance limit is priority levels for implementation of the defined as the acceptable rate of soil erosion suitable measures (biological or engineering) at which the quality of a soil as medium for for watershed treatment. plant growth can be maintained. Mandal and Sharda (2011) found soil tolerance limit 5 Usefulness of the study and conclusions ton/ha/year for this particular region. According to table 2 Buda Kedar watershed is On-site erosion reduces the soil quality due to most sensitive to erosion as well as removal of nutrient rich top soil layer and also deposition, and however, Mehalchauri is least reduces the water holding capacity of many sensitive for erosion/deposition. The region is soils. Therefore, in developing countries like clear that Buda kedar consists only 33.9% India, where rural population is more than forest area, however, Mehalchuair comprises 65% and land is the identity of the people, 50.6% forest area. Beside this, a large part of assessment of erosion focuses mainly on the Buda Kedar watershed is under agricuture. on-site effects of erosion. It is evident from the literature (Singh, 2006) that productivity Worth notable that during tillage operations, can be increased by bringing the soil loss near land is exposed and loosened and become or below the soil tolerance limit. Therefore, more suspectible for the erosion. The field comparison between soil tolerance limit and verfication confirmed that erosion and prevailing soil loss can be the best criteria for associated sediment deposition had indeed deciding the priority areas as well as occur in these areas. Such categorizations of treatment activities within the watershed. In

Int.J.Curr.Microbiol.App.Sci (2017) 6(3): 40-56 the present study, the formulated model has within a catchment i.e to identify sources of potential to identify the areas where soil loss sediment erosion. The formulated model exceeding the permissible/tolerance limit (9% involving only elementary processes area of study watershed has been identified in responsible for soil erosion and subsequently the present study) and these are the areas sediment deposition is capable to capture the where immediate soil conservation measures spatial patterns of soil erosion as well as have to be implemented. In the nutshell, our lumped sediment load in the entire watershed target should to bring the soil loss within the up to a cell area (90m x 90m). Formulated soil loss tolerance limit which will certainly model is more rational and pragmatic than the lead to sustainable agricultural production lumped Sediment Delivery Ratio (SDR) system. approach. Worth notable that GIS and remote sensing techniques are capable to access data In the present era of industrialization, more in tough terrain and adverse climatic attention is being paid to the society at large, conditions, since, model requires input data in viz., in flood prevention, water reservoir the form of different thematic GIS layers, preservation, and water pollution control model has potential to be applied for any (Garen et al., 1999). However, it has been watershed, in terms of size and location, observed that sedimentation in the reservoirs throughout the globe. Moreover, the model have occurred at a much faster rate than may be a useful tool in sustainable anticipated. In view of the stupendous task of environmental watershed management construction and the huge cost involved in planning. hydro-power projects, ensuring longevity of the reservoir have become the cause of major References concern to the planners and administrators. Nevertheless, sedimentation data from Aksoy, H., Kavvas, M.L. 2005. A review of number of projects indicated that the actual hillslope and watershed scale erosion sediment production rates (SPR) have been and sediment transport models. Catena, much more than the assumed rates. The Vol. 64, pp. 247– 271. erosion/deposition maps derived by the American Society of Civil Engineers (ASCE), formulated model in the present study exhibit 1975. Sedimentation Engineering. sediment rate at a particular cell in spatial American Society of Civil Engineering, domain, and the value at the outlet cell New York: NY. indicate the sediment outflow from the entire Angima, S.D., Stott, D.E., O’Neill, M.K., watershed. The sediment outflow at any point Ong, C.K., Weesies, G.A, 2003. Soil may be quite helpful in planning to design erosion prediction using RUSLE for sediment production rate (SPR) for a hydro- central Kenyan highland conditions. power project at that location. However, sites Agric. Ecosystems Environ., Vol. 97(1– identified vulnerable to soil deposition from 3), pp. 295–308. such maps could be avoided for planning sites Arnold, J.G., Engel, B.A., Srinivasan, R., for multipurpose projects where huge amount 1993. Continuous-time, grid cell of storage of water is required. watershed model. In: Proceedings of the Conference, Spokane, WA, June 18–19, Whether the main concern of soil and water pp. 267–278. conservation planning is towards prevention Beasley, D.B., Huggins L.F., Monke, E.J. of on-site or off-site effects of erosion, there 1980. ANSWERS - a model for was a growing need for tools that enable to watershed planning. Transactions of define the spatial distribution of erosion American Society of agricultural 53

Int.J.Curr.Microbiol.App.Sci (2017) 6(3): 40-56

Engineers, Vol. 23, pp. 938–944. prioritization using GIS and remote Brazier, R.E., Beven, K.J., Freer, J., Rowan, sensing. Water Resource Management, J.S. 2000. Equifinality and uncertainty Vol. 24, No. 10, pp. 2091-2112. in physically-based soil erosion models: Jain, M.K., Kothyari, U.C., 2000. Estimation application of the GLUE methodology of soil erosion and sediment yield using to WEPP, the Water Erosion Prediction GIS. Hydrol. Sci. J., Vol. 45, No. 5, pp. Project for sites in the UK and USA. 771-786. Earth Surf Process Landform, Vol. 25, Jain, M.K., Kothyari, U.C., Ranga Raju, K.G, pp. 825–845. 2005. Geographic information system Desmet, P.J., Govers, G, 1996. A GIS- based distributed model for soil erosion procedure for the automated calculation and rate of sediment outflow from of the USLE LS-factor on catchments. J. Hydraulic Engineering, topographically complex landscape ASCE, Vol. 131, pp. 755–769. units. J. Soil and Water Conservation. Jain, S.K. Goel, M.K, 2002. Assessing the Vol. 51, pp. 427–433. vulnerability to soil erosion of the Ukai ERDAS (Earth Resources Data Analysis Dam catchments using remote sensing System), 2005. ERDAS Imagine 8.6, and GIS. Hydrol. Sci., Vol. 47, No. 1, Users Manual. ERDAS Inc.: Atlanta, pp. 31-40. GA. Jarvis, A., Rubiano, J., Nelson, A., Farrow, ESRI (Environmental Systems Research A., Mulligan, M, 2004. Practical use of Institute), 1994. Cell based modeling SRTM data in the tropics: comparisons with GRID. Environmental Systems with digital elevation models generated Research Institute Inc., Redlands, from cartographic data. Cali: Centro California, USA. Internacional de Agricultura Tropical Ferro, V., Minacapilli, M. 1995. Sediment (CIAT), Working document, 198, 32. delivery processes at basin scale. Knisel, W.G. 1980. CREAMS: a field-scale Hydrological Sci. J., Vol. 40, No. 5, pp. model for chemical, runoff and erosion 703-717. from agricultural management systems. Foster G.R, 1982. Modelling the erosion Conservation Research Report No. 26, processes. In Hydrological Modelling of South East Area, US Department of Small Watersheds, Haan CT, Johnson Agriculture, Washington, DC. H, Brakensiek DL (eds). ASAE Lee, S. 2004. Soil erosion assessment and its Monograph No. 5, American Society of verification using the universal soil loss Agricultural Engineers, St. Joseph, MI; equation and geographic information 297–380. system: a case study at Boun, Korea. Garen, G., Woodward, D., Geter, F. 1999. A Environ. Geol, Vol. 45, No. 3, pp. 457– user agency’s view of hydrologic, soil 465. erosion and water quality modelling. Mandal, D., Sharda, V.N. 2011. Assessment Catena, Vol. 37, pp: 277-289. of permissible soil loss in India Haan, C.T., Barfield, B.J., Hayes, J.C. 1994. employing a quantitative bio-physical Design Hydrology and Sedimentology model. Curr. Sci., Vol. 100, No. 3, pp: for Small Catchments. Academic Press, 383-390. New York. Merritt, W.S., Latcher, R.A., Jakeman, A.J., Jain, M.K., Das, D, 2010. Estimation of 2003. A review of erosion and sediment sediment yield and areas of soil erosion transport models. Environ. Modelling and deposition for watershed and Software, Vol. 18, pp. 761– 799.

Int.J.Curr.Microbiol.App.Sci (2017) 6(3): 40-56

Moore, I., Burch, G., 1986a. Physical basis of Raymo, M.E., Ruddiman, W.F., 1992. the length-slope factor in the universal Tectonic forcing of late Cenozoic soil loss equation. Soil Sci. Society of climate. Nature (London), Vol. 359, pp: America J., Vol. 50, pp. 1294-1298. 117-122. Moore, I., Burch, G. 1986b. Modeling erosion Risse L.M., Nearing, M.A., Nicks, A.D., and and deposition: topographic effects. Laflen, J.M., 1993. Error measurement Trans. of ASAE, Vol. 29, No. 6, pp in Universal Soil Loss Equation. Soil 1624-1630, 1640. Sci. Society of America J., Vol. 57, pp. Moore, I., Wilson, J.P. 1992. Length slope 825-833. factor for the Revised Universal Soil Singh, G., Babu, R., Narain, P., Bhusan, L.S., Loss Equation: simplified method of Abrol, I.P., 1992. Soil erosion rates in solution. J. Soil Wat. Consrv., Vol. 47, India. J. Soil and water Conservation, No. 5, pp. 423-428. Vol. 47, pp. 97–99. Naik, G.M., Rao, E.P., Eldho T.I. 2009. Finite Singh, K.D. 2006. Participatory watershed Element Method and GIS Based management – a key to sustainable Distributed Model for Soil Erosion and agriculture. J. Indian Soc. Soil Sci., Vol. Sediment Yield in a Watershed. J. 54, pp. 443–451. Water Resources Management, Vol. Stoddart, D.R. 1969. World erosion and 23, pp. 553-579. sedimentation in water. In: Chorley, Nash, J.E., Sutcliffe, J.V., 1970. River flow R.J., (Ed.), Water, Earth, and Man, forecasting through conceptual models. Methuen, London, pp. 43–64. Part I. A discussion of principles. J. Takken, I., Beuselinck, B., Nachtergaele, J., Hydrol., Vol. 10, pp. 282–290. Govers, G., Poesen, J., Degraer, G., NBSS and LUP. 2004. Soils of Uttar Pradesh 1999. Spatial evaluation of a physically- for Optimizing Land Use: Executive based distributed erosion model Summary. National Bureau of Soil (LISEM). Catena,Vol. 37, pp. 431–447. Survey and Land Use Planning: ICAR, United Nation Environmental Program, 1991. Publication No., 68, Nagpur, India. Status of Desertification and Nearing, M.A., Foster, G.R., Lane, L.J., Implementation of the UN Plan of Finkner, S.C. 1989. A process-based Action to Combat Desertification. soil erosion model for USDA-water UNEP, Nairobi, Kenya. erosion prediction project technology. Van Rompaey, A., Bazzoffi, P., Jones, R.J.A., Transactions of the ASAE, Vol. 32, No. Montanarella, L., 2005. Modelling 5, pp. 1587– 1593. sediment yields in Italian catchments. Prosser, I.P., Rustomji, P., Young, B., Moran, Geomorphol., Vol. 65, pp. 157–169. C., Hughes, A., 2001. Constructing Van Rompaey, A., Verstraeten, G., Van Oost, river basin sediment budgets for the K., Govers, G., Poesen, J., 2001. National Land and Water Resources Modeling mean annual sediment yield Audit. CSIRO Land and Water using a distributed approach. Earth Technical Report 15/01, pp. 34, Surface Processes and Landforms, Vol. Canberra. 26, No. 11, pp. 1221-1236. Ram Babu, Dhyani, B. L., Kumar, N., 2004. Verstraeten, G., Poesen, J., 2001. Factors Assessment of erodibility status and controlling sediment yields from small refined Iso- Erodent Map of India. intensively cultivated catchments in a Indian J. Soil Conserv., Vol. 32, No. 2, temperate humid climate. Geomorphol., pp. 171–177. Vol. 40, pp. 123–144.

Int.J.Curr.Microbiol.App.Sci (2017) 6(3): 40-56

Verstraeten, G., Prosser, I.P., Fogarty, P. Agricultural Handbook No. 537, US 2007. Predicting the spatial patterns of Dept of Agriculture, Science and hillslope sediment delivery to river Education Administration. channels in the Murrumbidgee Wu, S., Li, J., Huang, G. 2005. An evaluation catchment, Australia. J. Hydrol., Vol. of grid size uncertainty in empirical soil 334, pp. 440–454. loss modelling with digital elevation Wicks, J.M., Bathurst, J.C., 1996. SHESED: a models. Environ. Modeling and physically based, distributed erosion Assessment, Vol. 10, pp: 33–42. and sediment yield component for the Wu, T.H., Hall, J.A., Bonta, J.V. 1993. SHE hydrological modeling system. J. Evaluation of runoff and erosion Hydrol., Vol. 175, pp. 213– 238. models. J. Irrigation and Drainage Williams, J.R., Dyke, P.T., Jones, C.A. 1983. Engineering, ASCE 119, pp. 364–382. EPIC: a model for assessing the effects Young, R.A., Onstad, C.A., Bosch, D.D., of erosion on soil productivity. In Anderson, W.P., 1989. AGNPS: a non- Analysis of Ecological Systems: State- point source pollution model for of-the-Art in Ecological Modeling. Eds. evaluating agricultural watersheds. J. W.K. Laurenroth et al., Elsevier, Soil and Water Conservation, pp. 168– Amsterdam, pp. 553-572. 173. Wischmeier, W.H., Smith, D.D. 1978. Predicting rainfall erosion losses.

How to cite this article:

Rawat, S.S., M.K. Jain, K.S. Rawat, B. Nikam and Mishra, S.K. 2017. Vulnerability Assessment of Soil Erosion/Deposition in a Himalayan Watershed using a Remote Sensing and GIS Based Sediment Yield Model. Int.J.Curr.Microbiol.App.Sci. 6(3): 40-56. doi: https://doi.org/10.20546/ijcmas.2017.603.004