Soil water deficit as a tool to measure water stress and inform silvicultural management in the Cape Forest Regions, South Africa

Abstract: An understanding of variations in water availability to plantation forests on a spatial and temporal scale is essential when designing risk averse and site-specific silvicultural management regimes. Various indices of site water availability were compared to each other and to an independent, unbiased estimate of stand productivity potential, namely site index, across the Tsitsikamma, Knysna and Boland forestry regions of South Africa. This was done to find the balance between water availability indices requiring intensive data inputs (that may be very accurate) and indices with lower input data requirements (but may sacrifice some accuracy). The following indices of water availability (in order from low to higher input data requirements) were tested: Mean Annual Precipitation (MAP); Aridity Index (AI), i.e., MAP as a fraction of mean annual potential evapotranspiration (Ep); Moisture Growing Season (MGS), i.e., the Julian days where long-term MAP exceeds 0.3 times Ep; Water Deficit (WD), an estimate based on a rudimentary water balance with relatively low data inputs. The first three estimates use only climatic variables while the WD incorporates soil water storage capacity to run a water balance calculation. Results showed that both regional climatic variability and soil properties significantly affected the level of water availability, and hence also the potential productivity of pine stands. The shallow and sandy soils from the Knysna and Boland regions exhibited rapid water depletion during periods of decreased precipitation and seasonal shifts, however, the large WD’s (up to 345 mm year-1) observed in several of these sites rapidly changed to surplus values following only one month of high precipitation. Sites from the Tsitsikamma region had significantly larger water retention capabilities and this was attributed to the regional soil properties and climatic conditions. Temporal variations in the WD were also quantified. The WD estimates correlated significantly (r = -0.80, p<0.001) to the respective site indices from sites across all regions. These results underscore the importance of soil water availability on plantation productivity, especially in moderately dry regions or in areas with either shallow soils or a seasonal rainfall pattern. We conclude that the WD is a fairly accurate estimate of site-specific water availability with relatively low data requirements. The WD estimates are far superior to currently used indices of water availability in Southern Africa and has data input requirements that are currently readily available for most plantation forest sites.