Reducing standard errors by incorporating spatial autocorrelation into a measurement scheme for soil carbon credits

Several studies have suggested that geostatistical techniques could be employed to reduce overall transactions costs associated with contracting for soil C credits by increasing the efficacy of sampling protocols used to measure C-credits. In this paper, we show how information about the range of spatial autocorrelation can be used in a measurement scheme to reduce the size of the confidence intervals that bound estimates of the mean number of C-credits generated per hectare. A tighter confidence interval around the mean number of C-credits sequestered could increase producer payments for each hectare enrolled in a contract to supply C-credits. An empirical application to dry land cropping systems in three regions of Montana shows that information about the spatial autocorrelation exhibited by soil C could be extremely valuable for reducing transactions costs associated with contracts for C-credits but the benefits are not uniform across all regions or cropping systems. Accounting for spatial autocorrelation greatly reduced the standard errors and narrowed the confidence intervals associated with sample estimates of the mean number of C-credits produced per hectare. For the payment mechanism considered in this paper, tighter confidence intervals around the mean number of C-credits created per hectare enrolled could increase producer payments by more than 100 percent under a C-contract.