Submitted to: Soil Science Society of America Journal
Publication Type: Peer Reviewed Journal
Publication Acceptance Date: 4/22/2010
Publication Date: 9/29/2010
Citation: Nemes, A., Timlin, D.J., Quebedeaux, B. 2010. Ensemble approach to provide uncertainty estimates of soil bulk density in support of simulation based environmental impact and risk assessment studies. Soil Science Society of America Journal. 74:1938-1945. Interpretive Summary: The U.S. Department of Agriculture conducts a national assessment of environmental benefits and effects of the 2002 Farm Bill programs. The assessments which utilize computer simulation models of plant growth and soil processes require knowledge of the density of soil. Since it is difficult to measure soil density over all the locations where the assessment is carried out, it must be estimated from other available soil data. The goal of this study was to develop an improved method to estimate soil density. We developed a novel technique to estimate soil density from readily available soils data. This method also allowed us to account for the uncertainty of the estimated density values. The ability to quantify uncertainty will better support simulation based environmental risk assessment studies. Utilizing the recommended changes, field soil conditions can be simulated more realistically, and as a result, agricultural production and its impact on the environment can be assessed more accurately. Successful completion of this project will equip lawmakers and regulatory agencies with better tools and measures to help promote sustainable agricultural practices, which is in the interest of the general public.
Technical Abstract: Large scale environmental impact studies typically involve the use of simulation models and require a variety of inputs, some of which may need to be estimated in absence of adequate measured data. One important input is bulk density that partially determines conditions for soil aeration, solute transport and storage as well as the outcome of soil carbon stock calculations. Correct representation of Db in simulation studies is essential since any bias or uncertainty will propagate through a variety of processes and time steps. We used a U.S.-wide soil database of point measurements and a ‘k-nearest neighbor’ pattern recognition algorithm combined with a data re-sampling technique to estimate Db and its uncertainty. Soil particle-size distribution and organic carbon content were utilized as input to test the benefit of limiting the development data in terms of (1) Soil Taxonomy classification; (2) sample depth; and (3) soil horizon notation. We obtained an overall root-mean-squared error of 0.17gcm-3 and mean error of 0.01gcm-3. Limiting samples by taxonomic classification proved to be advantageous, while limiting samples by depth helped avoid depth-specific bias in the estimations. Limiting samples by horizon notation did not yield significant improvement due to the great variability of bulk density within horizons. A simple adjustment to the re-sampling utility can be used to generate narrower or wider confidence intervals to the estimates without affecting the mean estimates. Simulation based environmental risk assessment studies can be direct beneficiaries of data with characterized uncertainty. We present a potential application of this approach in which the originally proposed data support of the simulation based U.S. national scale Conservation Effects Assessment Project can be judged and optionally corrected.