MODELLING SOIL TEMPERATURE USING SURFACE AIR TEMPERATURE IN SOUTHWESTERN, NIGERIA

Abstract

The measurement of soil temperature is very important for many purposes such as; Agriculture, Meteorology, Climatology and study of heat storage variation. However, there is an insufficient climatic stations for monitoring soil temperature in Nigeria, hence there is need to carry out a research work on how to predict soil temperature using air temperature. For estimating soil temperature from air temperature, regression analysis was done using a year data (May, 2010 – April, 2011) in Akure and Lagos respectively, with daily mean soil temperature as dependent variables and daily mean air temperature as independent variable and they are validated using the data of the succeeding year (May, 2011 – April, 2012) in Akure (7o 17’N, 5o 18’E) and Lagos (6o 27’N, 3o 24’E) respectively. Correlation coefficients (R) and coefficient of determination () between these two variables were determined. It was observed that the values of R and were different for different seasons in the study area. The highest value of these two variables were experienced in the wet season in Akure and Lagos with R = 0.87 and = 0.75; R = 0.92 and = 0.85 respectively and the dry season with R = 0.84 and = 0.71; R = 0.88 and = 0.77 respectively. The predicted value of soil temperature from the regression equation is found to be well correlated with the measured soil temperature with Mean Bias Errors of 0.64 in Akure and 0.59 in Lagos and a Root Mean Square Errors of 0.82 in Akure and 0.80 in Lagos during the wet season. Also, during the dry season the Mean Bias Errors of -1.83 and -0.49 were observed in Akure and Lagos respectively while the Root Mean Square Errors were 2.13 and 0.87 in Akure and Lagos respectively. The low value of Mean Bias Error and Root Mean Square Error show that there is a strong linear correlation between daily mean air temperature and daily mean soil temperature. The model obtained would enable us to predict soil temperature data for improved modelling of ground wave radio propagation and soil moisture content whenever desired.