A method for assessing the hydrological impact of afforestation using regional mean annual data and empirical rainfall-runoff curves

Using a case study from Australia, a method was developed to integrate regional-scale rainfall-runoff data with site-scale information to predict the likely magnitude of mean annual runoff reductions due to afforestation. It was hypothesised that large, long-term, mean annual rainfall-runoff datasets can be used to provide statistical limits around the variability of regional hydrology and that data from plantation conversion experiments can be used to identify quantiles within such data that correspond to pre- and post-plantation land-cover, enabling the transparent quantification of regional changes in runoff due to afforestation. Three simple empirical functions with flexible parameter structures and a documented precedence in the assessment of the hydrological impacts of land-cover change were used to perform quantile regression on the regional data. Structural deficiencies were identified in the two parameter versions of all three functions, requiring the use of well-posed, single parameter functions to form coherent quantile curve arrays. Plantation conversion data from a number of Australian sites were used to select regional pre- and post-conversion rainfall-runoff quantile curves for each empirical function which were in turn used to estimate proportional runoff reductions for regional Australia. USDA curves provided the closest agreement with results reported by a number of independent studies compared to a simple tanh function and Zhang curves. The USDA function had a more flexible structure in describing runoff from dry areas than either tanh or Zhang curves. Inflexibility in the Zhang curves was attributed to limitations in the model’s specification.