Given the challenges of direct observation of the impacts of planted forests on water yield or quality, and the long timescale between afforestation actions and catchment responses, a variety of analytical tools have been developed to assess the potential impact of afforestation on water. Models currently used to assess the impact of conversion of agricultural land to plantations are typically either directly derived from catchment data or from process simulation models.
Models derived from catchment data are generally based on statistical regression equations or simple concepts that explain part of the observed differences in long-term streamflow from forested and non-forested catchments (e.g. Bosch and Hewlett, 1982; Zhang et al, 2004). These have been integrated into more sophisticated decision support tools to analyse the potential impacts of afforestation on water properties in different parts of the world (McVicar et al, 2007; Van Dijk et al, 2007). While statistically robust, the predictive power of these models for specific applications is limited. Relationships commonly used do not account for differences in species, stocking, plantation age, other management factors (see Chapter 5) or the variety of conditions that might be found within catchments. One obstacle is the large amount of data required to derive a statistically significant relationship due to the often stronger variation caused by (unknown) climate and terrain factors, and sampling or other errors.
Mechanistic approaches (Chappell et al, 2007) may be more promising. These can account for differences in soil properties, rooting depth and vegetation characteristics (e.g. leaf area or surface conductance), and some include dynamic growth models that can simulate the effect of stocking density, soil fertility and management regime (e.g. Gallant et al, 2005). While useful for sensitivity studies, the accuracy and reliability of these models is also constrained by data availability, particularly for soil properties (see Scott et al, 2005; Ilstedt et al, 2007).
Was this article helpful?