What does it take to detect a change in soil carbon stock? A regional comparison of minimum detectable difference and experiment duration in the North-Central United States

Authors: NECPÁLOVÁ, MAGDALENA - University Of Wisconsin; ANEX, ROBERT - University Of Wisconsin; KRAVCHENKO, ALEXANDRA - Michigan State University; ABENDROTH, LORI - Iowa State University; Del Grosso, Stephen - Steve; DICK, WARREN - The Ohio State University; HELMERS, MATTHEW - Iowa State University; HERZMANN, DARYL - Iowa State University; KLADIVKO, EILEEN - Purdue University; LAUER, JOSEPH - University Of Wisconsin; NAFZIGER, EMERSON - University Of Illinois; SAWYER, JOHN - Iowa State University; SCARF, PETER - University Of Missouri; STROCK, JEFFREY - University Of Minnesota; VILLAMIL, MARIA - University Of Illinois

Submitted to: Journal of Soil and Water Conservation
Publication Type: Peer Reviewed Journal
Publication Acceptance Date: 10/28/2013
Publication Date: 11/17/2014
Publication URL: http://doi:10.2489/jswc.69.6.517
Citation: Necpálová, M., Anex, R., Kravchenko, A., Abendroth, L., Del Grosso, S.J., Dick, W., Helmers, M., Herzmann, D., Kladivko, E., Lauer, J., Nafziger, E., Sawyer, J., Scarf, P., Strock, J., Villamil, M. 2014. What does it take to detect a change in soil carbon stock? A regional comparison of minimum detectable difference and experiment duration in the North-Central United States. Journal of Soil and Water Conservation. 69:517-531.

Interpretive Summary: Accurate estimation of soil organic carbon (SOC) is crucial to efforts to improve soil fertility and stabilize atmospheric CO2 concentrations by sequestering carbon (C) in soils. Due to high variability among measurements, many SOC experiments have low statistical power, which makes likely erroneous conclusions that a treatment has no effect (Type II error). We evaluated the size of minimum detectable differences (MDD) in SOC given the variability in soils in the North Central USA. The MDD was calculated for a standardized experiment) in randomized complete block design with probability of Type I error (a) of 0.05 and probability of Type II error (ß) of 0.15. The data for the study come from 13 experimental sites located in Iowa, Illinois, Ohio, Michigan, Wisconsin, Indiana, Missouri and Minnesota. When expressed as a proportional difference from the SOC baseline, the MDD at 0-10 ranges from 6 to 31 % and its size increases with soil depth. The explanatory power of the relationships increased with depth. Simulations of SOC dynamics using the DAYCENT model suggest that the duration of a standardized no-till experiment on different soils would have to be between 25 and several hundred years to achieve the MDD in SOC in the top 20 cm. The need to have very large numbers of measurements or very long experiments creates significant challenges to the agronomic study of SOC in these soils.

Technical Abstract: Accurate estimation of soil organic carbon (SOC) is crucial to efforts to improve soil fertility and stabilize atmospheric CO2 concentrations by sequestering carbon (C) in soils. Soil organic C measurements are, however, often highly variable and management practices can take a long time to produce measurable differences in SOC contents. As a result, many SOC experiments have low statistical power, which makes likely erroneous conclusions that a treatment has no effect (Type II error). In this study we evaluate the size of minimum detectable differences (MDD) in SOC given the variability in soils in the North Central USA. The MDD was calculated for a standardized experiment (two by two factorial with 5 replications) in randomized complete block design with probability of Type I error (a) of 0.05 and probability of Type II error (ß) of 0.15. We estimate through simulation the number of years required at different locations for the standardized experiment of no-till management to generate the MDD in SOC. The data for the study come from 13 experimental sites located in Iowa, Illinois, Ohio, Michigan, Wisconsin, Indiana, Missouri and Minnesota. Soil organic C, bulk density, and texture have been measured at depths 0-10, 10-20, 20-40, and 40-60 cm. The calculated MDD in SOC ranges from 1.04 to 7.10 g/kg at 0-10 cm and from 0.84 to 5.35 g/kg at 10-20 cm depth. When expressed as a proportional difference from the SOC baseline, the MDD at 0-10 ranges from 6 to 31 % and its size increases with soil depth. Comparison across the sites shows that the MDD is correlated with SOC concentration and soil texture at corresponding depths. The explanatory power of the relationships increased with depth, i.e. soil texture is more closely related to SOC variability in the deeper layers. Simulations of SOC dynamics using the DAYCENT model suggest that the duration of a standardized no-till experiment on different soils would have to be between 25 and several hundred years to achieve the MDD in SOC in the top 20 cm. The need to have very large numbers of measurements or very long experiments creates significant challenges to the agronomic study of SOC in these soils, as well as to implementing policy to increase C sequestration.