Keywords
forest hydrology, rainfall-runoff curves, quantile regression
Start Date
1-7-2012 12:00 AM
Abstract
Following the procedure of Greenwood et al. [2011], a re-parameterisedform of the hyperbolic tanh function of McMahon et al. [1992] and the well-knownUSDA model are used to explore the regional hydrological impacts of afforestationin Australia. Quantile regression is used to analyse a large, long-term, meanannual rainfall-runoff dataset. According to Budyko [1974], variability in regionalmean annual runoff is driven by the regional radiation-water balance or potentialevapotranspiration. The mean annual dataset is taken to represent acomprehensive description of the regional hydrology, supported by Australia’sdistinctive regional hydro-climatology. Regional plantation and pre-conversion‘grassland’ runoff relationships are identified by statistical association with high andlow potential evapotranspiration rainfall-runoff quantiles using site-scaleexperimental data. Both models gave similar results, however the standard USDAmodel provided a superior fit to the regional data than the re-parameterised tanh.The result underlines the importance of regional context, that is, therepresentativeness of the regional data and the chosen model, in providingconfidence in an assessment, over the quality of site data. A case studydemonstrates how good regional context can be used to transparently scrutinisethe appropriateness of site data for use in regionalisation. The approach hasapplicability to any distinctive hydro-climatic region that can be reasonablydescribed by rainfall-runoff functions.
Integrating Site- and Regional-scale Data in Assessing the Hydrological Impact of Afforestation Using Rainfall-runoff Curves
Following the procedure of Greenwood et al. [2011], a re-parameterisedform of the hyperbolic tanh function of McMahon et al. [1992] and the well-knownUSDA model are used to explore the regional hydrological impacts of afforestationin Australia. Quantile regression is used to analyse a large, long-term, meanannual rainfall-runoff dataset. According to Budyko [1974], variability in regionalmean annual runoff is driven by the regional radiation-water balance or potentialevapotranspiration. The mean annual dataset is taken to represent acomprehensive description of the regional hydrology, supported by Australia’sdistinctive regional hydro-climatology. Regional plantation and pre-conversion‘grassland’ runoff relationships are identified by statistical association with high andlow potential evapotranspiration rainfall-runoff quantiles using site-scaleexperimental data. Both models gave similar results, however the standard USDAmodel provided a superior fit to the regional data than the re-parameterised tanh.The result underlines the importance of regional context, that is, therepresentativeness of the regional data and the chosen model, in providingconfidence in an assessment, over the quality of site data. A case studydemonstrates how good regional context can be used to transparently scrutinisethe appropriateness of site data for use in regionalisation. The approach hasapplicability to any distinctive hydro-climatic region that can be reasonablydescribed by rainfall-runoff functions.
