Browsing by Author "Bakimchandra, Oinam"

Now showing 1 - 1 of 1

Open Access
Integrated fuzzy-GIS approach for assessing regional soil erosion risks
(2011) Bakimchandra, Oinam; Wieprecht, Silke (Dr.-Ing.)
Modelling a dynamic and physical process, such as soil erosion, is prone to errors and problems. The availability of the right kind of data source, quality of data used, scale issues in modelling, measurement errors etc. and the complexity of the model in itself are some of the issues that are explicitly addressed and reported in soil erosion research studies. Existing soil erosion models based on physical processes are very data demanding in both their amount of variables and their temporal and spatial resolution requirements. Hence, data scarcity and lack of reliable data tend to pose a problem for successful application of physical based erosion models. On the other hand, less data demanding empirical based models are developed for a certain environmental set up using erosion plot studies and thus their applicability is restricted to regions where they were developed. Another significant aspect that is overlooked in many past soil erosion risk assessment studies is the nature of the various environmental control parameters involved in modelling, which are fuzzy in reality. When mapping erosion risk, the introduction of fuzzy sets instead of crisp sets to define classes (i.e. degree of hazard or risk) will help to incorporate a degree of fuzziness within each class of the governing parameters. It is found that various existing soil erosion risk models consider each feature and spatial units present on the landscape or catchment as having distinct boundaries. In reality, the existing natural boundaries are much more complex. To cope with such problems of class boundaries and to incorporate the expert knowledge that can represent the processes under investigation, there is a need of fuzzy logic based modelling approach. In this PhD research, a simple and efficient fuzzy logic-based soil erosion risk model for monitoring the soil erosion risk distribution over a regional landscape is developed. The developed model is known as Fuzzy-Water Erosion Risk Classification and Assessment Model (F-WERCAM). As the name indicates, this model is intended for water based soil erosion risk classification and their assessment using a fuzzy logic modelling concept in a GIS platform. The model is designed or set up in such a way that it has minimum input data requirements for model execution, provided the considered input parameters are the main primary governing factors that influence the soil erosion risk of a region. One of the salient features in the F-WERCAM is the multi-stage modelling approach. It consist of 3 stages namely, Stage 1- mapping of the Soil Protection Index (SPI), Stage 2- mapping of the Potential Erosion Risk Index (PERI) and Stage 3- mapping of the Actual Erosion Risk Index (AERI).This set up allows for the simplification of the fuzzy rule bases by reducing the number of input parameters at each stage of the modelling. In addition, this approach allows for a step-by-step evaluation of the intermediate results. For instance, Stage 1 of the modelling approach allows for the evaluation of the SPI of a region, before integrating with the PERI of Stage 2, to obtain the final AERI of a region (Stage 3). The final soil erosion risk map provides qualitative based information on the distribution pattern of the soil erosion risk classes over a region. Apart from the qualitative based spatial information on soil erosion risk obtained from this model, the possibility of transferring the output erosion risk index into quantitative soil loss values (in t/ha/yr) is explored and discussed in this study. The model is successfully tested in Upper Awash Basin in Ethiopia and further used to produce a soil erosion risk map of Italy. The ability of fuzzy logic to describe and transform the knowledge in a descriptive human like manner in the form of simple rules using linguistic variables has provides a new direction and opening to explore and develop a simple and well structured framework for soil erosion risk assessment. Overall, the integration of fuzzy logic within GIS using remotely sensed data in this research tries to address the problems of data scarcity, uncertainty in the input model parameters and handling of large spatial data effectively. From the various assessments and evaluations presented in this research, it is found that such an expert based fuzzy logic model has the potential to be used as a practical tool for assessment of regional soil erosion risk by policy makers and scientists.