Land development impacts on surface runoff using remote sensing and an artificial neural network model
Some types of land development can be associated with increased impervious area causing increase in surface runoff and decrease in ground water recharge. Both of these processes can have large scale ramifications through time. Increased runoff results in higher flows during rain events, which in tur...
| Main Authors: | , , , |
|---|---|
| Format: | Conference or Workshop Item |
| Language: | English |
| Published: |
2005
|
| Online Access: | http://psasir.upm.edu.my/id/eprint/39397/ http://psasir.upm.edu.my/id/eprint/39397/1/39397.pdf |
| Summary: | Some types of land development can be associated with increased impervious area causing increase in surface runoff and decrease in ground water recharge. Both of these processes can have large scale ramifications through time. Increased runoff results in higher flows during rain events, which in turn increases the number of times that a river floods adjacent lands. Likewise, this increase in runoff and channel flows can drastically increase the erosion of river channel beds and banks, potentially destabilizing bridges or local structures. Understanding how the landuse change influence the river basin hydrology may enable planners to formulate policies to minimize the undesirable effects of future landuse changes. Hence there is a growing need to quantify the impacts of landuse changes from the point of minimizing potential environmental degradation. The main objective of this study was to develop a model to assess the impacts of landuse changes on the watershed runoff. While conceptual or physically based models are of importance in the understanding of hydrologic processes, there are many practical situations where the main concern is with making accurate predictions at specific locations. Remote sensing was used in this study because of its ability in providing useful information on landuse dynamics. An Artificial Neural Network (ANN) model was developed. To evaluate the model performance in both training and testing phases, mean absolute error (MAE), mean square error (MSE), U’ Theil and regression analysis were performed. The correlation coefficients were 0.94 and 0.89 for the training and testing, respectively. Most of the data points were within the confidence level of 95%. Most of the scatter values were within 15 % deviation bands. The model was applied to future scenarios using a certain rainfall amount; deforestations of 10%, 50% and 80% will increase the runoff amount by about 60%, 250% and 300% respectively. |
|---|