Assessing the CABALA model for predicting shelterbelt productivity – initial stand volume comparisons

[摘要] Integrated agroforestry systems can provide a wide range of benefits, both to landholders and to society. Shelterbelts are a form of agroforestry typically designed to improve productivity through minimising the effects of damaging winds on livestock and crop/pasture systems, while maximising the land available for agricultural production. In contrast to agroforestry block plantings, shelterbelts are relatively narrow; therefore ‘edge’ trees represent a relatively large proportion these plantings, which is likely to influence their productivity relative to blocks. This study provides an assessment of the capacity of the dynamic forest growth model, CABALA, to predict production potential of agroforestry plantings, particularly shelterbelts, in Tasmania. Although CABALA has previously been used to predict the yield of block plantings in many situations, its capacity to predict productivity of belts has not yet been tested. Four shelterbelts in Tasmania were intensively measured as part of a set of experiments established by Private Forests Tasmania, CSIRO and University of Tasmania to explore the interactions between P. radiata shelterbelts and adjacent pasture. These experiments form the basis of the data used in this study. Measurements of stem diameters at breast height over bark and tree height were used to estimate tree volume from an existing allometric relationship for P. radiata. These tree measures from a complete cross-section (i.e. inner and edge rows) of shelterbelts were used to investigate the ‘edge effect’. Additional data for nearby block plantings were also collated from previous measurement programs. The CABALA model was used to predict standing volume under bark, in m3 ha-1, for each belt and block site. Model inputs were populated using a combination of observations and informed assumptions. Belts were set up in a ‘blocked’ design using the observed number of rows, row spacing and tree spacing, and assuming the maximum allowable between-belt spacing to simulate a contiguous, infinitely long belt. Adjustments were made to the stand volume outputs based on the ratio of stocking from the simulated belt to the observed stocking, to account for the fact that CABALA reports standing volume over the whole area and not just the belt area. Block plantings, and ‘simulated’ block plantings using information from the inner row(s) of the belt, were also set up in an ‘unblocked’ design using the observed row and tree spacings. Sensitivity of the modelled outputs to two uncertain parameters, soil C:N ratio and soil depth, were also tested. For the shelterbelts, there was a significant effect of edge on productivity, with higher stand volume for trees growing on the edges of belts adjacent to the pastures. This agrees with previous results in other agroforestry shelterbelts, including blue gums, mallees and pines. The ratio of tree diameter to height was also significantly higher in edge rows compared with inner rows for all sites, but the edge effect on height was variable.There was reasonable agreement between CABALA predicted stand volume and measured stand volume for both belts and blocks. In general, the model tended to under-predict the productivity of belts and over-predict the productivity of blocks. Variation in agreement for different stand types indicated that CABALA captures stand productivity well for some sites but not others. Testing at two of the shelterbelt sites where soil depth was uncertain indicated that modelled stand volume was relatively sensitive to soil depth. It may be that other site-specific inputs/events, including soils information, management and natural events such as exposure to insect pests or weed competition, were not well captured. Thus adequate soils information and a record of the complete management and disturbance history is important for accurate prediction of the productivity of stands.

[发布日期] 2018-05-28 [发布机构] CSIRO

[学科分类] 地球科学(综合)