Abstract
In this study, we tried to show the direct and indirect effect of precipitation and temperature on the production of citrus in Tunisia of a governorate i and the neighboring governorates. To do this, we used a new original analysis of spatial econometrics to take into account in an efficient and finely manner the spatial effects, individual and temporal effects of the spatial autocorrelation. This analysis was done on the basis of global spatial autocorrelation test and the spatial autoregressive model (SAR) as well as the spatial Durbin model (SDM). It appears from our results that the available water in the groundwater table of the governorate i can be an effective solution for the farmer who resides there provided that the means are implemented so that he can benefit. Our robustness results based on the cointegration dynamic panel data, also shows the effect of temperature via the hydric resources of the governorate i and that the neighboring governorates represent a negative spillover effect.