Improved estimation of biomass accumulation by environmental plantings and mallee plantings using FullCAM
Paul, K., Roxburgh, S., Raison, J., Larmour, J., England, J., Murphy, S., Norris, J., Ritson, P., Brooksbank, K., Hobbs, M., Neumann, C., Lewis, T., Read, Z., Clifford, D., Kmoch, L., Rooney, M., Freudenberger, D., Jonson, J., Peck, A., Giles, R., Bartle, J., McAurthur, G., Wildy, D., Lindsay, A., Preece, N., Cunningham, S., Powe, T., Carter, J., Bennett, R., Mendham, D., Sudmeyer, R., Rose, B., Butler, D., Cohen, L., Fairman, T., Law, R., Finn, B., Brammar, M., Minchin, G., Van Oosterzee, P., and Lothian, A. (2013) Improved estimation of biomass accumulation by environmental plantings and mallee plantings using FullCAM. Report. CSIRO, Canberra, ACT, Australia.
![[img]](https://researchonline.jcu.edu.au/style/images/fileicons/application_pdf.png) |
PDF (Published version)
- Published Version
Restricted to Repository staff only |
Abstract
Overview: During the last two years, a major nationally-collaborative research program has been lead by CSIRO to improve the estimation of biomass accumulation by mixed-species environmental plantings and mallee eucalypt plantings. It has involved evaluation of the uncertainties associated with using alternative approaches to biomass estimation, and the collation and refinement of new and existing field inventories and biomass estimates for these plantings, growing in various configurations throughout the non-arid (>300 mm mean annual rainfall) regions of Australia. A large database on growth and biomass accumulation across a wide range of planting types has been developed, comprising 1,480 site-based observations, or 884 site-based observations not including repeated measures at the one site, 183,675 stem diameter measures (36% from new work in this project) and 8,288 measures of tree or shrub above- and below-ground biomass (40% from new work in this project). These data have been analysed to identify the key factors affecting the growth of plantings, resulting in 26 statistically-different categories of plantings. Modifiers that account for large variations in growth of these categories of plantings have been developed for use in FullCAM (which underpins the Reforestation Modelling Tool, RMT).
Sampling error: Sampling error was found to be the main factor affecting the accuracy of biomass estimates. Unless a sufficient number of trees/shrubs are sampled in a manner that is representative of the planting, biomass estimates can have high coefficients of variation of >50%. Even within reasonably homogeneous plantings, a large number of trees needed to be sampled to obtain biomass estimates with a 90% chance of being within ±10% of the true mean. For block planting geometries, the number of trees required to be measured based on simple random sampling was 700-1,600 and 130-280 for direct seeded and tubestock plantings, respectively. In linear plantings, the number of trees required was 540-1,030 and 116-180, respectively. Guidance is also provided on sampling strategies to decrease sample error by providing representative plots, and correct definition of the extent (area) of the planting for consistency in estimates of biomass when comparisons between linear and block plantings are required.
Verification of allometrics: Direct field measures of above- and below-ground biomass (through whole-of-plot harvesting) were used to test the reliability of a range of allometric equations. It was shown that uncertainties resulting from the application of allometric equations to estimate above-ground biomass are very low (generally <10% difference between measured and estimated biomass) when using site-based allometrics, or moderate (generally <16% difference between measured and estimated biomass) when generalised non-site allometrics are used. This report contains a comprehensive set of new allometric equations that can be used to estimate biomass of mixed-species environmental and mallee eucalypt plantings.
Root to shoot ratios: Root to shoot ratios can be high in young plantings grown in water and nutrient-limited environments. They ranged between 0.28 and 0.81 across 13 sites studied. Ratios were higher in tree-dominated plantings, where the ratio tended to decline as productivity increased. A set of new root allometric equations were verified with direct measurement from whole plot excavation.
Uncertainty in biomass estimates: Several sources of uncertainty contribute to the challenge of reliably estimating biomass accumulation, with estimates being highly variable within a planting, across the broader landscape, and over time. Substantial errors can result from sampling the tree/shrub population, in measuring trees/shrubs during field inventories, and in applying allometric equations to convert field inventories of measurements of stem diameter into biomass. The magnitude and relative importance of these sources of error was quantified using extensive field measurements combined with simulation modelling.
Key factors influencing growth: Detailed statistical analysis of biomass measurements contained in the database has enabled plantings to be categorised into three planting types: mixed-species environmental plantings in temperate regions, mixed-species environmental plantings in tropical regions, and mallee eucalypts. These three planting types were then further categorised according to; (i) planting geometry (blocks, or linear plantings of varying widths), (ii) stand density, and (iii) species in the case of mallees (E. polybractea, E. loxophleba ssp. lissophloia, or 'other' species), or species-mix in the case of environmental plantings (shrub-dominant or tree-dominant). These categories provided the basis for calibration (estimation of appropriate modifiers) of the Tree Yield Formula in FullCAM.
FullCAM yield curve calibrations: The un-calibrated yield curve for environmental plantings in FullCAM generally lead to underestimation of biomass. To account for differences in growth rates between the 26 categories of plantings, new modifiers have been provided for the Tree Yield Formula within FullCAM. With these new modifiers, the overall model efficiency was only 46 and 63% for mixed-species and mallee eucalypt plantings, respectively. However, there was no apparent bias in model predictions and the model is satisfactory for most individual planting categories or types. Therefore, modelled estimates of biomass accumulation will be reliable on average, but estimates at any particular location will be uncertain, with either significant under- or over-prediction possible. Results indicate that when compared to the un-calibrated yield curves, early growth was likely to be more rapid, and total above-ground biomass may be higher for many plantings at maturity. Some recommendations are provided for how the model might be applied spatially so as to increase its utility without compromising confidence in predictions. One key recommendation is related to temporal applicability; the recommended age domain of the calibrations is less than 24, 19 and 14 years for temperate environmental plantings, tropical environmental plantings, and mallee plantings, respectively.
Significant challenges remaining: There are three key challenges remaining:
• Most field observations were for young stands, and temporal change in longer-term stand dynamics in above- and below-ground biomass remains poorly understood, especially for mixed-species stands where tree/shrub dominance is likely to change with stand age and/or disturbance (e.g. by fire or drought). Repeated temporal measurement of growth in contrasting types of plantings will be extremely valuable for model evaluation of both above- and below-ground biomass, and for further refinement of the calibrations of FullCAM's yield curves in the future.
• To increase confidence in spatial application of the calibrations, additional estimates of biomass carbon are required from a range of climatic regions for 3 of the 26 planting categories (E. polybractea in dense or sparse narrow linear plantings, and wide linear environmental plantings where stocking and fraction of trees is high).
• There are some factors which are known to influence growth but which are yet to be accounted for in calibration of the yield curves within FullCAM. These include access to a watertable (including establishment in riparian areas), whether a planting is growing in saline surface soil, or whether the planting has been coppiced. Further work is required to assess the impacts of these factors for the different types of plantings over the longer-term.
Actions (Repository Staff Only)
 |
Item Control Page |