ADDING VALUE TO SOIL FERTILITY TRIALS WITH A WATER BALANCE
J L Moir, D R Scotter, M A Turner, M J Hedley, R W Tillman, and D J Horne
Institute of Natural Resources , Massey University, Palmerston North
Field trial data relating pasture growth to measures of soil fertility are confounded by many site-specific environmental factors, particularly the weather. One approach to accommodate this is to express fertiliser responses in terms of relative rather than absolute yields. But this approach places constraints on trial design, and is unhelpful when attempting to extrapolate data in order to estimate actual yields at other sites or in other years. We suggest an alternative approach: a simple model that explicitly takes the weather into account. The Excel spreadsheet model assumes that pasture growth is proportional to evapotranspiration, and that the proportionality constant (k) depends largely on the soil's fertility status. The evapotranspiration is found from a simple daily soil water balance.
The potential of this approach is demonstrated using data from two field studies. One measured the response to irrigation near Palmerston North. The other measured production from high and low fertility sites in the Wairarapa. Values for k depend on soil fertility status, and varied from 11 to 19 kg dry matter per hectare for every mm of evaporation. The model can be used with synthetic weather data to generate probability density functions for summer-autumn pasture production. The greatest divergence between the measured and modelled production occurs when rainfall follows a prolonged dry period. Some possible reasons for this are discussed.