Abstract Model simulations are increasingly being used to assess farm performance and its impact on the environment. A crucial input in most of these models is daily rainfall data, which are difficult to obtain for remote areas or if the period required is long. Recently, a new long-term daily climate dataset for all New Zealand, known as Virtual Climate Station (VCS) data, has become available from NIWA (National Institute of Water & Atmospheric Research Ltd). A reasonable agreement between the VCS and observed data has been reported; nonetheless the effect of the uncertainty in the estimated rainfall on model simulations remains an issue to be examined. In this study we investigated the differences in simulated evapotranspiration, drainage and pasture growth by two models of varying complexity when using VCS or observed rainfall data from several locations around New Zealand.
The results showed a general agreement between model outputs whether using observed or VCS rainfall datasets. Model outputs that are closely related to rainfall events, such as drainage, were more sensitive to deviations in the rainfall datasets than the indirect outputs, such as evapotranspiration. Data aggregation and correction of bias improved agreement between observed and VCS data and this was reflected in the model outputs. Although short-term simulations should be analysed with care, overall results were promising. The NIWA VCS rainfall data have the potential to be extensively used in conjunction with simulation models for the assessment of land use and management practices.
Keywords EcoMod; modelling; rainfall; water balance
A08005; Online publication date 14 August 2008; Received 18 January 2008; accepted 13 June 2008
New Zealand Journal of Agricultural Research, 2008, Vol. 51: 317–329
0028–8233/08/5103–0317 © The Royal Society of New Zealand 2008
PDF file of entire paper: Print-quality (1622K) | screen-quality (1597K)
This year's abstracts | Journal home page | All abstracts | Publishing home page