Skip to main content
ARS Home » Research » Publications at this Location » Publication #98669


item Moorman, Thomas
item Smith, Jeffrey
item Karlen, Douglas
item Dao, Thanh

Submitted to: Soil Science Society of America Journal
Publication Type: Peer Reviewed Journal
Publication Acceptance Date: 11/1/1999
Publication Date: N/A
Citation: N/A

Interpretive Summary: Concerns about the effectiveness of conservation programs to address soil degradation and climate change issues have stimulated interest in monitoring the condition and trend of soils on a regional scale. Such monitoring programs must be based on statistically rigorous tests to strengthen the conclusions and to develop sound agricultural policy. This research was conducted to determine the feasibility of using the National Resource Inventory (NRI) of the USDA-NRCS to monitor the state of soils in four different regions of the U.S. We determined that the NRI was an effective framework for this type of monitoring effort. Specific statistical analyses were found to increase the reliability and usefulness of the monitoring data. Researchers and USDA-NRCS program administrators interested in regional soil conservation and soil quality monitoring programs will benefit from these results by applying them to future sampling programs and existing soil quality data.

Technical Abstract: There is increasing interest in assessing the status and trend of soil resources at the regional scale. However, there is little information on the probability distribution and variability of soil properties at this scale. Our objective was to evaluate the distribution and variability of 23 soil properties at a regional scale. Samples were collected irrespective of soil series from two Major Land Resource Areas (MLRA) (9 and 105), and from the Ascalon (Fine-loamy, mixed, superactive, mesic Aridic Argiustoll) and Amarillo (Fine-loamy, mixed, thermic Aridic Paleustalf) soils in MLRA 67 and 77, using the National Resource Inventory sampling design. The distribution of the soil properties was tested for normality, and the variability and magnitude of change that could be detected was estimated using an lsd. Most soil properties were non-normally distributed, with the frequency of non-normality varying between MLRA. Confining sampling to a single soil series did not consistently improve the precision with which soil properties were estimated. Natural log transformation resulted in normal distributions for most soil properties and reduced their variability 2- to 3-fold. However, a few soil properties remained non-normally distributed. A change of <20% of the regional mean could be detected for most soil properties with untransformed data, and <10% of the regional mean for most soil properties following a natural log transformation. Our results indicate that the distribution of soil properties at a regional scale should be examined before any statistical tests are made, and transformed when necessary to improve the power to detect change.