WP6: Satellite remote sensing and geostatistics (Leader: Dr M. Disney + Lewis, Moncrieff, Huntley, Williams)

WP6 will provide spatial observations (with errors) for assimilation into the ecosystem models of WP7. Provision of spatiotemporal data will be key to WP7 and addressing hypotheses H2, H3 and H7. The main types of EO data that will be used are: (i) moderate spatial, high temporal resolution data (MODIS/MERIS/AVHRR) for studying broad-scale current and historical temporal dynamics; (i) high spatial, low temporal resolution data (current IKONOS, ETM and historical TM data) to populate a database characterising vegetation state over the study areas; (iii) airborne radiometry using an ASD spectroradiometer and digital camera on the flux aircraft (WP5) and airborne imagers and LiDAR data (provisional on ARSF/BAS approval) for detailed spectral and spatial characterisation; (iv) ground measurements to monitor phenology and test models and scaling using the ASD and semi-permanent reflectance and photochemical index monitoring with 3 SKYE sensors.

To deal with the complex nature of the landscape and property scaling (H7), the database will incorporate a 3D scene simulation model (‘landscape model’) with explicit spatial representations of scattering ‘objects’ (trees, layers of grasses, bushes) derived from the high resolution EO data and ground survey (coordinated with WP1) on underlying topography (Disney et al., 2000). This will allow accurate representation of vegetation state in the ecosystem models (WP7) and will be used to understand the scaling and topographic influences of the radiation regime to moderate resolutions using Monte Carlo Ray Tracing (Burgess et al., 1995; Lewis, 1999). Analysis of data will also be used to identify ‘emergent’ scales of ecosystem operation, analyse and investigate topographic controls (H2) and for footprint analysis of tower and aircraft data (WP4,5). The models will also provide a ‘legacy’ product for further studies at the test sites.

The high resolution ‘landscape’ models will provide detailed characterisation of the sites, but these are costly to develop and will be infrequently updated. Higher temporal frequency monitoring will be achieved through moderate resolution EO data. These will be used for scaling analyses (H7) and assimilation (WP7) over the test sites, as well as wider areas at this latitude. The ‘landscape’ models will enable upscaling, and will also be used to develop an ‘EO operator’ linking radiation interception and ecophysiological variables in the ecosystem models, allowing assimilation of satellite-measured reflectance data and simpler characterisation of uncertainties than using higher order EO products (WP7).

Deliverables (addressing H2, H3 and H7)
A. Detailed scene models for Abisko.
B. Repeat of A for Kevo. (A paper in Rem. Sens. Env. using output of A 7 B).
C. Identifying key radiometric processes and spatiotemporal scales of topography, and soil and vegetation distributions at Abisko.
D. Repeat of C for Kevo. (A paper in Glob. Ch. Biol. using outputs of C & D).
E. Assessment of uncertainties in EO data at high latitudes. (A paper in Rem. Sens. Env.).
F. Assimilation of EO data into ecosystem model. (A paper in Glob. Ch. Biol.).
G. Comparison of photosynthetic efficiency derived from aircraft-measured fluxes with those derived from EO data at Abisko.
H. A repeat of G at Kevo. (A paper in Rem. Sens. Env. using outputs of G & H).

Tasks
1. Preparation of EO products for assimilation over test sites, (WP7 tasks 6, 15) (end 2006).
2. Installation of semi-permanent monitoring radiometers to characterise phenology (WP7 task 7) and reflectance (end May 2007 Abisko, end May 2008 Kevo with switchover mid-season each year).
3. Generate DEM and information on vegetation distribution (June 2007 for both sites).
4. Detailed survey of vegetation quantity and structure at test sites (Aug 2007 Abisko, Aug 2008 Kevo).
5. From (3) and (4) develop detailed 3D scene simulation models (end 2007 Abisko, end 2008 Kevo).
6. Validation of 3D models using tower (from 2) and airborne (WP5) radiometric data (as for 5).
7. From (5) identify key spatial and temporal radiometric scales (linked to WP7 tasks 5, 8) (as for 5).
8. Parameterisation of radiation interception in ecosystem model through development of EO operator (WP7, tasks 5, 8) (end 2007 Abisko, end 2008 Kevo).
9. From (2) and (5) generate estimates of photosynthetic efficiency and compare with estimates derived from aircraft flux measurements (WP5).
10. Upscale estimates of photosynthetic efficiency to region using EO data from (3) and compare with flux tower and model estimates (WP7 task 10, WP4 task 8).
11. Provide reflectance data for full assimilation via EO operator (WP7 task 10).