bei Google ScholarTitelangaben
Adiku, Samuel ; Reichstein, Markus ; Lohila, Annalea ; Dinh, Nguyen Quoc ; Aurela, Mika ; Laurila, T. ; Lüers, Johannes ; Tenhunen, John:
PIXGRO: A Model for Simulating the Ecosystem CO₂ Exchange and Growth of Spring Barley.
In: Ecological Modelling.
Bd. 190
(2006)
Heft 3-4
.
- S. 260-276.
ISSN 0304-3800
DOI: https://doi.org/10.1016/j.ecolmodel.2005.04.024
Abstract
A model, PIXGRO, developed by coupling a canopy flux sub-model (PROXELNEE; PROcess-based piXEL Net EcosystemCO2 Exchange) to a vegetation structure submodel (CGRO), for simulating both net ecosystem CO2 exchange (NEE) and growthof spring barley is described. PIXGRO is an extension of the stand-level CO2 and H2O-flux model PROXELNEE, that simulatesthe NEE on a process basis, but goes further to include the dry matter production, partitioning, and crop development for springbarley. Dry matter partitioned to the leaf was converted to leaf area index (LAI) using relationships for the specific leaf area(SLA). The canopy flux component, PROXELNEE was calibrated using information from the literature on C3 plants and wastested using CO2 flux data from an eddy-covariance (EC) method in Finland with long-term observations. The growth component(CGRO) was calibrated using data from the literature on spring barley as well as data from the Finland site. It was then validatedagainst field data from two sites in Germany and partly via the use of MODIS remotely sensed LAI from the Finland site.Both the diurnal and the seasonal patterns of gross CO2 uptake were very well simulated (R2 = 0.92). A slight seasonal biasmay be attributed to leaf ageing. Crop growth was also well simulated; simulated dry matter agreed with field observed data fromGermany (R2 = 0.90). For LAI, the agreement between the simulated and observed was good (R2 = 0.80), giving an indicationthat functions describing the conversion of fixed CO2 to dry matter and the subsequent partitioning leaf dry matter and LAIsimulation were robust and provided reliable estimates.The MODIS LAI at a resolution of 1000m agreed poorly (R2 = 0.45) with the PIXGRO simulated LAI and the observed LAIat the Finland site in 2001.We attributed this to the coarse resolution of the image and/or the small size of the barley field (about17 ha or 0.25 km2) at the Finland site. By deriving a regression relation between the observed LAI and NDVI from a higherresolution MODIS (500m resolution), the MODIS-recalculated LAI agreed better with the PIXGRO-simulated LAI (R2 = 0.86).