Abstract
In Central America, coffee is mainly grown in agroforestry systems. This practice modifies the microclimate, which, in turn, influences coffee growth and development. However, modeling these microclimate modifications is a challenge when trying to predict the development of a disease in the understory crop, based on variables usually monitored in weather stations exposed to full sunlight. Furthermore, critical variables for plant disease development, such as leaf wetness duration and leaf temperatures, are generally not measured by weather stations. In our study, we sought to build models explaining daily minimum and maximum coffee leaf temperatures, daily coffee leaf wetness duration, and minimum and maximum air temperatures in agroforestry systems with a single shade tree species, which are common in Central America, and which were characterized by shade tree height, canopy openness and light gap distribution. The modeled variables were mainly explained by one or more meteorological variables provided by reference weather stations exposed to full sunlight. The presence of shade trees resulted in a buffer effect, reducing daily maximum air and leaf temperatures, and increasing daily minimum air and leaf temperatures. Moreover, except for the daily minimum air temperature under shade, shade tree characteristics affected these microclimatic variables. Indeed, the buffer effect on the daily maximum air temperature increased with shade trees 7 m tall or over, whereas for extreme leaf temperatures, this effect seemed to be further intensified by a dense and homogeneous canopy. The tallest shade trees also tended to provide conditions that reduced coffee leaf wetness duration. The coffee leaf stratum affected the daily maximum leaf temperature, with a top layer intercepting radiation for the lower strata, but had no effect on the daily minimum leaf temperature, detected at night. The models developed were simple equations allowing interpretation of shade tree height, the effects of canopy characteristics on the microclimate and were therefore useful for designing and managing agroforestry system. The more accurate models could be incorporated into an early warning system for coffee pests and diseases in the region.