The researchers compared different methods to estimate how much water plants need in a dry area. They used artificial neural networks to predict this water requirement with just two factors: average temperature and sunlight. The best network had 2–21 parameters. The Penman-Monteith method calculated an average of 1447.4 mm of water needed annually in Varanasi. Some methods underestimated, while others overestimated, but the artificial neural network gave a close estimate of the water needed by plants.