Measurement and modelling of canopy water fluxes in representative forest stands and a matorral community of a small Sierra Madre Oriental watershed, northeastern Mexico

The quantitative importance of canopy water fluxes was evaluated within the four principal vegetation communities of a small watershed in the central Sierra Madre Oriental, northeastern Mexico. The partitioning of cumulative gross precipitation into throughfall during wet season periods in red oak (84.1 ± 2.3%) and cedar-white oak-ash (84.3 ± 3.4%) forests, as well as a matorral subinerme brush community (83.3 ± 3.3%) were not significantly different (α = 0.05). In addition, cumulative throughfall partitioning by a pine-oak forest canopy during wet + dry season periods was not significantly different (83.6 ± 1.8%), suggesting that the rainfall regime of the area may be more important than the vegetation cover present in determining the proportion of cumulative season-long gross precipitation partitioned into throughfall. Although stemflow was assumed to be a negligible component of the canopy water balance of the forest stands in this watershed, this flux was not inconsequential in the matorral subinerme community, accounting for 8.5 ± 1.9% of the cumulative gross precipitation input derived from 25 wet season events. Given the relatively large cumulative stemflow flux observed in the matorral subinerme community, cumulative canopy interception loss was found to be significantly smaller from this community than from the sub-temperate forest stands. The revised Gash analytical canopy water flux model (Valente et al., 1997) using within-stand derived model parameters was found to simulate cumulative canopy interception loss to within 2.9% of the observed flux in the pine-oak stand, and simulated cumulative throughfall to within 2.0% of the observed flux in the cedar-white oak-ash stand. However, poor model performance was found using predetermined parameter values, suggesting that the transferability of this model may be limited. The model was also found to perform poorly at the gross precipitation event scale and at the point spatial scale. Relationships developed between the mean during-event evaporation rate and red oak stand characteristics suggest the need for a multi-layer analytical model. The spatial nature of throughfall, especially in the red oak community, throughfall gauge requirements, and the importance of during-event evaporation losses from these Madrean plant communities are also evaluated.