Mathematical modelling of irrigation water requirements for rice farming in Central Thailand

Abstract

Rice farming in Central Thailand relies heavily on controlled irrigation, making efficient water management critical for sustainable production. This study examines how production input costs influence irrigation water demand, measured as irrigation water volume per unit area (m3∙ha−1), using panel data from 215 rice farmers in Sena District, Phra Nakhon Si Ayutthaya Province, covering the period 2016–2023. The dataset comprises 1,720 farm-year observations across approximately 7.58 km2 of irrigated area. A log-linear Cobb–Douglas model estimated using panel regression techniques is applied to quantify the elasticities of irrigation water demand with respect to production input costs. Because all variables are expressed in logarithmic form, the estimated coefficients are directly interpretable as elasticities. The results indicate that fertiliser (β = 0.210), pesticide (β = 0.165), machinery (β = 0.302), and fuel costs (β = 0.128) have positive and statistically significant effects on irrigation water demand per rai, whereas seed and labour costs are not statistically significant. For example, a 1% increase in fertiliser expenditure is associated with a 0.21% increase in irrigation water demand per rai, holding other factors constant. The model explains a substantial share of the variation in irrigation water demand (R2 = 0.63), indicating strong explanatory power. The findings highlight that chemical input use and mechanisation are key drivers of irrigation intensity in rice farming. These results underscore the importance of integrated water–input management strategies, including improved nutrient scheduling, energy- efficient pumping technologies, and data-driven irrigation planning, to enhance water-use efficiency and support sustainable rice production in Thailand.