Trend analysis of weather data in shrimp farming areas of Nagapattinam district of Tamil Nadu

Abstract

Examining the trends and building the forecasting models for the meteorological parameters has got significance in climate research. An attempt was made to study the trends of temperature and rainfall of Nagapattinam District, a main shrimp farming area in Tamil Nadu. Surface weather data of the study area obtained from Indian Meteorological Department, Pune were analysed using Mann–Kendall trend test (MK test) for the weather parameters viz., maximum, minimum, highest maximum (HMax), lowest minimum (LMin) and mean temperatures, rainfall and number of rainy days. Month-wise, seasonwise (winter, summer and rainy) and year-wise trends were computed for the monthly data from 1961 to 2000. Results of analysis revealed a significant trend for maximum temperature for all the cases of experimentation, whereas the trend was significant for few months or seasons for minimum, mean, highest and lowest temperatures and number of rainy days. Trend does not exist for rainfall of the region. Time series forecasting model was built for monthly maximum temperatures using the data from 1961 to 1980 for building the model and evaluated with the data from 1981 to 2000. Autoregressive integrated moving average model (ARIMA) (2, 0, 1) (1, 1, 1) s was found to be suitable model with highest R2 value of 0.94 and lowest Root Mean Square Error. Model predictions were reassessed with Willmott’s index and found to be reasonably good (0.95) for the study area. The forecasting model for temperature in the study area can be used for shrimp farming crop calendar planning.

Keywords

• Shrimp farming
• trend analysis
• Mann-Kendall test
• ARIMA models