Examining Variability of Methods for Determining Within Plot Soil Moisture Content.

In the Chequmegon National Forest in Northern Wisconsin a spatially intensive sap flux study seeks to find mechanistic controls on canopy transpiration along environmental gradients. It has been found that large variations of transpiration exist along an unevenly distributed moisture gradient extending from the upland forest to the forested wetland. It has been hypothesized that this variability could be explained in part by soil moisture. However, to test this hypothesis we required intensive within-plot soil moisture measurements, to evaluate a large set of soil moisture observations made extensively across the study area. The intensive study consisted of four plots, two from upland plots and two from wetland plots. In previous years soil moisture data was collected by taking 3-5 random samples per plot on 144 plots. To test the validity of this approach for both upland and wetland plots we recorded soil moisture at plot center and also 5 samples in 0.5m increments in the NW, N, NE, SE, S, SW directions. We statistically compared mean soil moisture values from among and within plots. Our results show that the forested wetlands have the greatest within plot soil moisture variation. A 3D GIS model was created to help visualize areas of the study site that may need further validation for soil moisture variability. For example, significantly raised areas on the 3D model may point to inadequate soil moisture measurements. The implications of this study are that 3-5 random samples of soil moisture may adequately capture plot means in upland regions, but in wetland regions a more rigorous sampling procedure is needed to ensure that robust relationships are found between canopy transpiration and soil moisture.