This work is licensed under a Creative Commons Attribution-NonCommercial 4.0 International License
|
||||||||
|
Paper Details
Paper Title
A Neural Network Approach To Real Time Rainfall Estimation
Authors
  Shubham Singla,  Haramrit Singh Sandhu,  Kamal Kumar
Abstract
Rainfall estimation from a watershed is of utmost importance for various hydraulic and hydrologic purposes like flood discharges, flow depths and other flood chracteristics (Jorge 2000). Operationally, rainfall estimation in real time on a 3-hourly timescale is potentially of great benefit for various hydrological forecasting purposes in a basin. In the present study, the aim is to analyze the temporal variations of rainfall using regression and ANN models. In the present study, rainfall estimation of Sutlej sub-watershed Himachal Pradesh which is located in districts of H.P having average elevation of 1400 m and spread over an area of 15802 km2 approximately is done using ANN and regression models. The basin hydrology features like flow accumulation, stream flow direction, watershed etc. have been extracted using SRTM DEM. An attempt has been made to use Trophical Rainfall Measuring Mission (TRMM) data procured from USGS website, as an input to an Artificial Neural Network (ANN) model during this study. Input parameters (Air temperature, Wind Speed, Air pressure, and Specific humidity) effects on rainfall have also been studied and correlation among these parameters has been analyzed. It has been observed that precipitation is heavily dependent on specific humidity and air temperature for the study area. Further, Multiple Linear Regression (MLR) model and ANN models have also been compared and to check the performance of rainfall prediction. Two different algorithms of ANN i.e. Stochastic Gradient Descent (SGD) and Multiple Layer Perceptron (MLP) have been used for rainfall estimation. The RMSE has been observed as 7.46 for SGD (ANN) whereas 8.87 and 9.8 for MLP (ANN) and MLR model respectively. Thus, the results obtained shows that the rainfall may be predicted better with the ANN model than the regression model.
Keywords- Rainfall, Watershed, Regression, ANN, SGD, MLR, MLP
Publication Details
Unique Identification Number - IJEDR1903031Page Number(s) - 176-185Pubished in - Volume 7 | Issue 3 | July 2019DOI (Digital Object Identifier) -    Publisher - IJEDR (ISSN - 2321-9939)
Cite this Article
  Shubham Singla,  Haramrit Singh Sandhu,  Kamal Kumar,   "A Neural Network Approach To Real Time Rainfall Estimation", International Journal of Engineering Development and Research (IJEDR), ISSN:2321-9939, Volume.7, Issue 3, pp.176-185, July 2019, Available at :http://www.ijedr.org/papers/IJEDR1903031.pdf
Article Preview
|
|
||||||
|