Supplementary material to “An Alaska Soil Carbon Database”
Published 26 May 2009
Kristofer Johnson, University of Alaska Fairbanks, Fairbanks, Alaska
Jennifer Harden, U.S. Geological Survey, Menlo Park, California
Citation:
Johnson, K., and J. Harden (2009), An Alaska Soil Carbon Database, Eos Trans. AGU, 90(21), 184. [Full Article (pdf)]
Results and Background for the workshop on the Alaska Soil Carbon Database
The Alaska Soil Carbon Database will be a product of the U.S. Geological Survey funded research on “Assessing the Role of Deep Soil Organic Carbon in Interior Alaska: Data, Models, and Spatial/Temporal Dynamics”. The workshop was organized by A. David McGuire (U.S. Geological Survey/University of Alaska Fairbanks), Jennifer Harden (U.S. Geological Survey) and Kristofer Johnson (USGS-funded postdoc at the University of Alaska Fairbanks). Participants who attended the meeting were Norman B. Bliss (ASRC Research and Technology Solutions), Mark Clark (Alaska Natural Resources Conservation Service), Teresa Hollingsworth (USDA Forest Service), Evan Kane (Michigan State University), Torre Jorgensen (Alaska Biological Research, Inc.), Merritt Turetsky (University of Guelph), Dave Valentine (University of Alaska Fairbanks), and Chien Lu Ping (University of Alaska Fairbanks). Other participants in the project that did not attend the workshop are David D'Amore (USDA Forest Service), Michelle Mack (University of Florida), Edward A. G. Schuur (University of Florida), and David Swanson (National Park Service).
Briefly, the primary goal of the database is to compile existing data for the development of soil carbon spatial models and then to incorporate this information into a biogeochemical modeling framework. Other anticipated outcomes of the project include:
- the identification of data gaps important to soil carbon cycling, both spatially and with depth;
- the quantification of the uncertainties of carbon storage in soils;
- the development of a resource for the convergence of spatial models (i.e. calibration and validation), with the idea that using the same and best available data should yield better modeling results;
- the development of an archive for soil carbon data in the future; and
- information to guide future sampling decisions and research activities.
Protocols for contributing data
The group decided on two broad categories of parameters to be included in the database. The first category (“very important”) focuses on parameters that can be more easily compiled and synthesized across datasets and that will contribute the most to research goals (see Table 1). The second category (“desirable if available”) is made of parameters that may not be available across all the contributing datasets, but will support the interpretation of the first category. Defining these categories also helps to encourage timely participation because all the collaborators can focus on filling in the first set of parameters to produce a more complete combined dataset. The database will eventually be housed in the MySQL engine, which interfaces with PHP should the development of a future web-based database be desired.
Complete raw forms of data (i.e. data that are not processed, summed, estimated, or missing) are the most ideal for the database. This is because the database engine, not the contributor, can perform the following “on the fly” calculations of carbon content (g C m-2) from raw data:
- single layers are calculated from measured percent organic carbon and bulk density;
- single layers are summed into horizons;
- horizons are summed into organic, mineral or total profile pools;
- profiles are averaged over a site; and
- sites are averaged for a region.
This has the advantage that the users always know what they are getting in the raw form and, if desired, perform calculations of carbon content with a different protocol. Another advantage is that if the original measurements are corrected, then the “on the fly” calculations will also automatically be updated.
However, there is a large body of data that cannot be used to calculate carbon content without some form of processing (e.g. bulk density estimations from percent carbon) and these procedures can become quite complex. Therefore, in addition to the raw data, the database allows for processed data to be entered if the appropriate metadata is included, e.g. references, regression equations and parameters and notes on interpolation procedures. These details ensure the user will know which carbon content values were processed and under what method.
It was decided that the data would be initially shared among the group and allow for different stages of data “readiness”. Eventually, as data is published, it will also become publicly available through the database. Password protection and other means of regulating data availability (e.g. applying “fuzziness” to plot coordinates) will be incorporated as the database develops. Initially, the database will incorporate data from the collaborators themselves but will eventually accommodate any soil carbon and related parameter data collected in Alaska.
Table 1. The list of parameters identified by collaborators at the workshop as “very important” or “desirable if available” to soil carbon studies in Alaska.
