[ Skip to content]



Arctic Biosphere Atmosphere Coupling at Multiple Scales

You are here: Abacus-IPY >> Work Packages

WP7: Modelling and synthesis (Leader: Dr M. Williams + Harding, Blythe, Disney, Moncrieff)

The aim of the modelling WP is to provide the means to extrapolate and scale information to test H7 by determining (1) whether flux tower measurements are consistent with multiple chamber measurements gathered within the footprint; and (2) whether aircraft measurements of landscape C fluxes are consistent with tower data from within the flight zone. To scale between measurements, we will use two modelling approaches, comparing simple empirical approaches with more complex process-based models. The process models of WP7 will also contribute to testing H2, H3c, H4a,b, and H5a,b.

Simple empirical models (e.g. light response functions) can be fitted to chamber (WP3) and flux tower (WP4) data via maximum likelihood techniques (Van Wijk & Bouten, 2002). We can test whether different chamber sites or tower sites have parameter sets in common (i.e. a similar functional type). For upscaling we grid the land surface, with parameters assigned through surveys and remote sensing and a look-up table. Simple models are computationally cheap and quick.

We will use two process based models, the first focussed largely on C fluxes and the other on water and energy, although there are overlaps between models which we will exploit to confirm predictions.

(1) The coupled SPA (Williams et al., 2000) and DALEC (Williams et al., 2005) models are designed for upscaling C dynamics, and parameter and state estimation through data assimilation. Initial parameter probability density functions (pdfs) are refined with data and Bayesian estimation techniques (Van Oijen et al., 2005). SPA has detailed photosynthesis, transpiration, snow and radiation/energy balance schemes. DALEC tracks the allocation, transfer and turnover of C in plant and soil pools.

(2) We will use JULES (based on Cox et al (1999)) and its aggregation schemes (tiling based on topography and hydrological sub models) to provide the link to global climate models. We will use a detailed landscape-scale analysis to investigate current methodologies used to represent the impact of sub-gridscale topography on water and C stores and fluxes in climate models (e.g. Gedney and Cox (2003), and test simulations of snow melt and soil freezing (Hall et al, 2003).

A. Modelling inter-comparison of chamber data of C fluxes for arctic sites (Abisko, Kevo, Alaska, Greenland, Svalbard) and identifying arctic ecosystem functional types. (A paper in Glob. Ch. Biol.).
B. Upscaling of chamber data and comparison with tower data at Abisko and Kevo. (A paper in Glob. Ch. Biol.).
C. Upscaling of tower data and comparison with aircraft data at Abisko and Kevo. (A paper in J.Geophys.Res.).
D. Data assimilation using tower, chamber and aircraft data to generate detailed C budgets at all field sites. Are flux data consistent with independent measurements of phenology, plant C allocation and SOM turnover? (A synthesis paper in Science or Nature).
E. Predictions of hydrology and energy balance of the land surface (depth of freeze, snow melt and seasonal energy balance), for comparison with tower and aircraft data, addressing H3c.
F. An assessment and further development of the representation of sub-grid hydrology and topography within process models. (A paper in J.Geophys.Res.).

1. Coordinate assessment of measurement errors across all WPs.
2. Set up and manage common data system, in coordination with project manager.
3. Develop, test and implement Bayesian parameter estimation (BPE) scheme.
4. Apply maximum likelihood techniques and BPE using chamber (WP3) to generate parameter pdfs.
5. Apply maximum likelihood techniques and BPE using towers (WP4) to generate parameter pdfs.
6. Use parameter pdfs to separate plant functional types (PFTs), and explore correlations between PFTs and structural parameters such as reflectance, LAI, species composition.
7. Generate meteorological drivers for modelling in both study regions, using WP4 outputs.
8. Field visits to coordinate modelling with field measurements.
9. Develop and implement improved modelling of arctic phenology in DALEC using WP1 outputs.
10. Assign model parameters and initial conditions to identifiable landscape units (use 5, WP2 & WP6).
11. Generate gridded simulations of C and water exchange for landscapes and use footprint modelling from WP4 to produce predictions for comparison with tower and aircraft flux measurements.
12. Use Ensemble Kalman filter to assimilate tower, chamber and EO data at each field site into DALEC model. Identify model-data inconsistencies and analytical uncertainties.
13. Compare model estimates of respiration, turnover and allocation with estimates derived from WP1 and WP2.
14. Feedback results to other WPs to refine their sampling protocols.
15. Test model simulations of snow and soil freezing (using data and analysis from WP4).
16. Test hydrological and topographic tiling routines (using landscape scale analysis and WP5 & 6).

© Abacus-IPY
Last modified: 26 Jan, 2006
The University of Edinburgh Durham University The University of Sheffield University College London The University of York Centre for Ecology and Hydrology University of Stirling Macauley Institute