Environmental Prediction Using Thermal Index Due to Statistical Physics of Q-Ising Model

We investigate environmental prediction using thermal index via the temperature-humidity index (THI) due to statistical mechanics of Q-state Ising model with small spins. We find that the environmental prediction realizes optimal THI, if we have perfect knowledges on the assumed true prior and the assumed transition probability from each original state to the observed one. Then, we find that the THI increases/decreases in comparison with the optimum with a decrease/increase in parameters expressing the temperature difference without using the model prior stabilizing the optimal thermal index. Also, we clarify that we can expand a range where the THI approximates its optimum by introducing the model prior.