Influence of forest cover on the spatio-temporal heterogeneity of the isotopic composition of precipitation at the small catchment scale

Forest cover influence the isotopic composition of precipitation before it eventually reaches the ground, especially through rainfall interception processes. Many plot scale recent studies focusing on throughfall and stemflow fluxes have demonstrated the role of forest canopy cover in (mostly) enriching their isotopic signature. However, the common approach in small catchments (even forested ones) remains to sample rainfall only at one single location (generally in an open area), assuming that the spatial variability of the isotopic composition of precipitation is small. Only a handful of studies have focused on the spatial variability of the isotopic composition of precipitation, and very few have included the role of forest cover. Nonetheless, a correct characterization of the isotopic composition of the incoming precipitation is essential in isotope-based catchment hydrology, for example to proceed hydrograph separation, as well as for process understanding or models development.

The aim of this study is to investigate the spatio-temporal variability of the isotopic composition of precipitation in a small Mediterranean catchment (0.6 km²) where forest cover roughly 2/3 of the catchment. Precipitation was sampled at the event scale in 31 locations across the catchment with bulk collectors consisting of plastic funnels (130mm diameter) connected to a 0.5-L plastic bin positioned 100cm above ground before each rain event and collected the day after. The sampling locations were distributed ±80m along 5 elevation lines every 50m (from 1150 to 1350m), 15 in open areas and 16 under forest (i.e., collecting throughfall). The percentage of canopy cover above each sampling location was determined using hemispherical photographs. For all events, rainfall was also measured every 5min at 3 locations with tipping bucket rain gauges and meteorological variables at 2 locations (at the ground level and above the forest canopy). Sampled events were analysed both altogether and separating open areas and under forest locations to determine the factors affecting the spatio-temporal variability of the isotopic composition of precipitation at the catchment scale and their relative influence.

Results show that mean δ¹⁸O of the events for the whole catchment varied from -11.96 to -3.6‰ along the year, with a mean coefficient of variation of 39%. Locations under forest were always more enriched than in open areas at the same elevation (+0.67‰ on average). Data analysis using the time stability approach (Vachaud et al., 1985) showed that forest locations had lower persistence of δ¹⁸O spatial patterns than open areas, indicating that spatial variability of isotopic composition was less predictable in forest locations compared to open areas. The elevation effect on δ¹⁸O, often observed in open area locations, was much less apparent in forest locations, confirming that forest introduced additional complexity on the spatial variability of the isotopic signal. Our findings highlight the actual need of taking into account the effect of both elevation and forest cover to assess a catchment scale representative isotopic composition of precipitation.