Does forest heterogeneity affect mean throughfall for regenerating secondary forests on Borneo?

Tropical landscape regeneration affects hydrological ecosystem functioning by regulating the amount of water that reaches the soil surface and changing soil infiltration rates. This affects the recharge and storage of water in the soil and streamflow responses. Therefore, it is important to assess how the fraction of rainfall that reaches the forest floor changes as secondary forests mature, and how forest structure affects throughfall via changes in storage capacity and evapotranspiration. Therefore, we monitored throughfall for twelve regenerating, logged-over forest plots in Sabah, Malaysian Borneo over a 7-month period and tested if inclusion of measures of forest heterogeneity improved the prediction of throughfall as a fraction of precipitation. On average across all plots, throughfall was 84% of precipitation, but was lower (as low as 74%) in plots with a longer recovery time since logging. There was a significant relationship between throughfall and tree density and basal area, as well as the Shannon Diversity Index and the coefficient of variation of the diameter at breast height, although species and structural diversity measures (Shannon Index and the coefficient of variation) did not improve model performance substantially. The overall best performing model was a linear regression with tree density. There was no relation between LiDAR-derived Top of Canopy (TCH) and mean throughfall, suggesting that this remotely sensed proxy of canopy height is not needed to estimate throughfall and more in-depth analysis of other LiDAR-products such as point clouds may be required. Our results imply that estimating throughfall in this forest type can be reliably achieved using tree density, and that this is not substantially affected by species diversity or structural heterogeneity variables, at least in the context of logged and regenerating forests in Sabah.