Rainfall interception and spatial distribution of throughfall in a pine forest planted in an arid zone

Afforestation in arid zones is of increasing importance as it is one of the favored approaches to combat desertification. Rainfall interception by the canopy plays an important role in determining the amount of rainfall reaching the forest floor. The throughfall patterns of coniferous forests planted in arid zones have hitherto not been well documented. The research site in which the measurements were carried out was located within a mature pine forest (Pinus halepensis, Mill) planted in an arid zone (average annual precipitation: 280 mm and annual class-A evaporation pan: 2500 mm). Measurements of precipitation, throughfall, stemflow, soil water content and transpiration were recorded during three years. Canopy cover was estimated from the analysis of hemispherical photographs obtained during the winter. The spatially averaged throughfall, obtained for each event from 20 rainfall gauges installed below the canopy was linearly correlated with gross rainfall and independent of rainfall intensity. We found that the spatial distribution of the throughfall on the forest floor was highly heterogeneous. A fairly consistent distribution of throughfall was evident with some gauges steadily showing a higher percentage of throughfall than others. However, we could not find any relation between the degree of canopy openness (or any other quantifiable canopy characteristic) and the relative throughfall. The spatial pattern of throughfall was however found to be significant when assessed by means of the "K-means Clustering" analysis. The spatial distribution of water in the soil as measured 48 h after rainfall events with a neutron gauge was not correlated to the distribution of throughfall when each of the measuring points was assessed individually. The average throughfall was however well correlated to the average increase in the soil water content and indicates that there is some degree of horizontal water movement (overland or below ground) within the plot. The results presented suggest that point measurements are not the correct tool for water balances of sparse canopy forests and that even for small plots horizontal sampling is of the essence. The practical implication of this study is that it shows that for this particular site a very large fraction of the rainfall reaches the ground and is thus available for plant and/or tree development, a key issue in an area with a rainless period of at least six months during which high pan evaporation rates prevail.