Multi-output ANN Model for Prediction of Seven Meteorological Parameters in a Weather Station
- 185 Downloads
The meteorological parameters plays a vital role for determining various water demand in the water resource systems, planning, management and operation. Thus, accurate prediction of meteorological variables at different spatial and temporal intervals is the key requirement. Artificial Neural Network (ANN) is one of the most widely used data driven modelling techniques with lots of good features like, easy applications, high accuracy in prediction and to predict the multi-output complex non-linear relationships. In this paper, a Multi-input Multi-output (MIMO) ANN model has been developed and applied to predict seven important meteorological parameters, such as maximum temperature, minimum temperature, relative humidity, wind speed, sunshine hours, dew point temperature and evaporation concurrently. Several types of ANN, such as multilayer perceptron, generalized feedforward neural network, radial basis function and recurrent neural network with multi hidden layer and varying number of neurons at the hidden layer, has been developed, trained, validated and tested. From the results, it is found that the recurrent MIMO-ANN having 28 neurons in a single hidden layer, trained using hyperbolic tangent transfer function with a learning rate of 0.3 and momentum factor of 0.7 performed well over the other types of MIMO-ANN models. The MIMO ANN model performed well for all parameters with higher correlation and other performance indicators except for sunshine hours. Due to erratic nature, the importance of each of the input over the output through sensitivity analysis indicated that relative humidity has highest influence while others have equal influence over the output.
KeywordsTime series model Meteorological parameter prediction Recurrent neural network Multi-input multi-output (MIMO)
The authors would like to thank Indian Academy of Sciences, Bangalore for providing fellowship to carry out this research through Summer Internship Program at Indian Institute of Technology Bombay, Mumbai 400 076.
- 1.V. Jothiprakash, S. Kirty, M.S. Tara, Prediction of meteorological variables using artificial neural networks. Int. J. Hydrol. Sci. Technol. 1(3/4): 192–206 (2011)Google Scholar
- 4.M. Hamdi, A. Bdour, Developing reference crop evapotranspiration time series simulation model using class a pan: a case study for the Jordan Valley/Jordan. Earth Environ. Sci. 1(1), 33–44 (2008)Google Scholar
- 7.M. Tektaş, Weather forecasting using ANFIS and ARIMA MODELS. A case study for Istanbul. Environ. Res. Eng. Manag. 1(51), 5–10 (2010)Google Scholar
- 8.ASCE Task Committee on Application of Artificial Neural Networks in Hydrology, Artificial neural networks in hydrology I: preliminary concepts. J. Hydrol. Eng. 5(2), 115–123 (2000a)Google Scholar
- 9.ASCE Task Committee on Application of Artificial Neural Networks in Hydrology, Artificial neural networks in hydrology II: hydrologic applications. J. Hydrol. Eng. 5(2), 124–137 (2000b)Google Scholar
- 14.S. Roy, Prediction of particulate matter concentrations using artificial neural network. Resour. Environ. 2(2): 30–36 (2012)Google Scholar
- 15.K. Sudheer, A.K. Gosain, K. Ramasastri, Estimating actual evapotranspiration from limited climatic data using neural computing technique. J. Irrig. Drain. (May/June), 214–218 (2003)Google Scholar
- 17.V. Jothiprakash, M.R. RamaChandran, P. Shanmuganathan, Artificial neural network model for estimation of REF-ET. J. Inst. Eng India 83(Jun), 17–20 (2002)Google Scholar
- 21.S. Darbandi, H. Arvanaghi, Air temperature estimation using artificial intelligent methods (Case Study : Maragheh City). Eur. J. Sci. Res. 61(2), 290–298 (2011)Google Scholar
- 25.S. Kirty, Artificial neural network and genetic programming models for meteorological parameter prediction. M.Tech Thesis, Indian Institute of Technology, Bombay, 2008Google Scholar