Sunday, August 26, 2012

Model-Data Integration at the Ground Level

(or… Note to modelers: the tundra is not always warm and sunny)

Several weeks ago, the ORNL Veg team (Colleen Iversen, Joanne Childs, and I) joined Victoria Sloan in Barrow  for the peak-season vegetation harvest and active-layer soil sampling. The work went well and we met all of our objectives in large part because we were joined by UAF vegetation modelers Dave McGuire and Eugenie Euskirchen. The samples we collected from the centers, edges, and troughs of four replicate polygons in four different areas of the BEO have already produced data on leaf area index and depth of organic soil, and as we continue to process the many samples, we will get data on aboveground biomass, specific leaf area, nitrogen and phosphorus content, root distribution, root chemistry, and soil carbon and nitrogen dynamics. 
The plant harvest team at the end of a successful campaign.
 
The reason for collecting these data is not simply intellectual curiosity, although that is a powerful motivation for much or what we do in this unique environment. Our mandate is to gather data that will improve the predictive capacity of earth system models, and one of our specific goals is to improve the definitions of plant functional types (pft) used in models. Dave and Eugenie are leaders in this area, so working side by side with them for several days was a fantastic opportunity for us to learn how our data can be used to improve their pft definitions, and for them to see where the data are coming from. Model-data integration is not simply sharing files over the internet; it must be a personal interaction to be truly successful, and there is probably no better way to achieve this than working together for several days.

Eugenie and Victoria clipping plants from a 25 x 25 cm frame

Dave inspecting the soil core that Joanne and Colleen are processing

Yesterday Victoria and I had another opportunity for close interaction with one of NGEE’s modelers. Dan Hayes joined us on a trip to the Council site on the Seward Peninsula. We wanted to get some initial experience at this site so we can plan a good sampling strategy for next year. As we surveyed the area together and collected samples in an abbreviated version of our Barrow harvest, Dan saw many opportunities for linking our work on the ground with information from remote sensing, which is his primary interest. 
Dan and Victoria recording soil temperature within a thermokarst featur
 
Council is a really beautiful site despite the cold, wet, and windy conditions we had to endure. The plant community is more complex than Barrow’s, and the changes brought on through permafrost degradation are obvious and a great research opportunity. We look forward to working here next year. Dan gets special kudos for his new-found skill as a Mud Road Trucker. Thanks to him for getting us to the site (and back!) safely over a road torn up by storms last November. We also hope Dan will tell his fellow modelers, who visited Council and Barrow two weeks ago under warm and sunny skies, that sometimes model-data integration requires surviving cold, wet, and windy!

The plant community is strikingly different within the thermokarst feature compared to the tundra community above it. (Photo from 2011 -- we saw no blue sky in this week's expedition.)
The tundra community has a fascinating complexity at a very small scale (2011 photo)
Dan navigates our truck through the muddy ruts and past the road construction equipment on the Nome-Council Road.