The spatial distribution characteristics of throughfall under Abies faxoniana forest in the Wolong Nature Reserve

Rainfall redistribution by forest canopy has been a study hotspot of ecological hydrology all the while, but studies on the horizontal spatial distribution of the throughfall under the forest canopy have so far been neglected. Examination of the spatial distribution of the throughfall is important to comprehend the eco-hydrological processes and the nutrient cycling in the forest. This study has been conducted in an Abies faxoniana forest in the Wolong nature reserve in the western part of the Sichuan province of China to determine the spatial distribution characteristics of the throughfall and to analyze the factors influencing the differences in spatial distribution. Eight rain gauges for throughfall collection were settled under the forest canopy at fixed positions, and the throughfall was recorded for 35 occasions of rain from July to September in 2001 and from June to August in 2002. The rainfall in this area is mainly comprised of light and moderate rains with a total rainfall of less than 31 mm, with rainfall intensities in most of the cases (about 98.2%) of less than 5 mm/h. The results showed that: the relationship between the throughfall percentage and total rainfall at each sampling site can be modeled more accurately by the Logistic equation than by the often used logarithmic equation, which employs ecologically less relevant parameters than the Logistic equation. The Logistic equation model is also fit for the whole forest. The throughfall percentages at the eight sites under the forest canopy were significantly different. The throughfall percentage at the fourth as well as the eighth sampling site showed different characteristics with the other six sites according to Q-clustering analysis. The throughfall under the fourth site had evident rainfall concentration (average throughfall 103.2%), which might be caused by the lower canopy coverage. The throughfall under the eighth site had a bigger variability than at the other seven sites, which might be caused by a more simple canopy form increasing the possible impact of wind. To assess the impact factors of the spatial distribution characteristics of the throughfall, canopy coverage, canopy depth, texture of leaves and branches above the sampling sites and the distance from the sampling site to the nearest tree trunk were analyzed. The results showed that: the throughfall percentage and the canopy coverage at each site seem to be negatively correlated, but the correlation is not significant (R=0.6996); throughfall percentage and canopy depth seem to be negatively correlated as well, the correlation is however not significant (R=0.4523). The relationship between the throughfall and the distance from the sampling site to the nearest tree trunk could be fitted with a polynomial equation with a high correlation coefficient (R=0.9491). The texture of the leaves and the branches above the sampling sites seemed to have influence on the throughfall percentages, but the influence of this factor is complex and needs to be examined further.