Paustian, K., E.A. Paul, E.T. Elliott
Presented at the All Scientist Meeting (1998-07-21 to 1998-07-22 )
Modeling SOM Dynamics and Management: Decision Making in an Enriched CO2 EnvironmentAs part of an ‘LTER spinoff’ project, funded by DOE, we are conducting field and modeling work to support improved quantification of soil C changes in ecosystems influenced by present and/or past agricultural practices. The focus of the project is: a) to collect and analyze soil C from forest successional systems on abandoned agricultural soils, using adjacent never-tilled forest sites and permanent agricultural sites as references, together with SOM fractionation, natural abundance 13C and 14C dating, and b) to develop key databases for regional extrapolation of soil C dynamics and to improve models of SOM response to management and vegetational changes, using fractionation and isotopic studies from the field sites and from previously investigated agricultural sites. Objectives, progress and outlook are presented for the second objective only.We are estimating historical productivity data for ca. 20 long-term agricultural field sites in the central US by augmenting site records with county-average crop yields obtained from the National Agricultural Statistics Service, US Agricultural Census and state-level records. Data obtained to date go back as far as 1917, but in most cases digital data are available from state sources only from about 1940 and nationally (NASS) since 1971. Data not available in digital form are being collected from printed reports. Initially we are focussing on counties corresponding to the long-term experiments, for which we will compare county-average and experiment station yields to develop correlations and derive ‘best estimates’of long-term trends for the pre-experimental period at each site. To estimate annual C inputs to soil, an extensive analysis of literature data has been used to derive a set of crop specific (which are time dependent to account for genetic improvements) functions relating grain yield to above-ground residue and root production. The yield data and C partitioning functions will first be used to estimate time series of historical C inputs for modeling SOM dynamics at the field sites (see below). Subsequently we will use the NASS yield data (1971-present) and the partitioning functions to estimate the quantify and trends of C inputs, nationally, as part of a larger effort to estimate the C balance for US agricultural soils.The objectives for site specific modeling of SOM dynamics is to evaluate to what extent measured ‘pools’ of SOM, i.e. from aerobic incubation, acid hydrolysis and 14C dating, can be closely approximated by conceptual model pools and how general this correspondence is across differing climates, soil types and management regimes. A modeling framework, TRACE_C, has been developed to test alternative model structures and assumptions and to simulate C isotope dynamics in addition to total C and SOM pools dynamics. The model uses measured/estimated C inputs (see above) and climate data as the primary driving variables. Results so far from several sites show relatively good correspondence between measured and modeled total organic C and 13C signatures for the total soil, in both native and cultivated sites, using the default model structure. Model estimates of quantity and 13C for a ‘resistant’ C were much less consistent with the attributes of the non-acid hydrolyzable fractions.Future work will focus on application of the model across all sites using the composite estimates of historical C inputs, evaluation of alternative model ‘hypotheses’ (i.e. formulations )— with comparisons of total C, analytical ‘pools’,and total soil 13C, C mineralization rates and 13CO2 signatures from long-term aerobic incubations — and testing/adaptation of the model for forest and reforested soils investigated within the project.Return to Contents
Back to meeting | Show |