What can soil-moisture data say about preferential flow?

Many investigations in the last fifteen years have inferred the occurrence of preferential flow from water content data measured at multiple soil depths at frequent time intervals. Because soil moisture does not directly indicate preferential flow, specific interpretive criteria are necessary.

The most widely used method is analysis of sensor response times with a nonsequential criterion: when a deeper sensor registers an increase in water content earlier than a shallower sensor, this is evidence that water has bypassed a layer of matrix material. This nonsequential response criterion has the advantage of being unambiguous and straightforward, but has limitations. Subsurface flow conditions cannot be identified if the soil is saturated. False negatives can occur if sensors have not been installed in a layer that has been bypassed as well as a layer of preferential flow accumulation. Preferential flow episodes may also be missed when the temporal resolution of soil moisture measurements is too low. Moreover, the chosen criteria that are used to identify the beginning of a new rainfall event can affect the outcome of the preferential flow analysis.

Another common method uses a velocity criterion; response rates indicate preferential flow if a wetting front moved faster than would be possible by means of diffuse flow. This method requires specification of a velocity threshold distinguishing preferential from diffuse flow. This threshold might be based on measurements of saturated hydraulic conductivity or model results, but problems such as spatial heterogeneity of soils can create substantial uncertainty in its results.

Ongoing efforts to develop and test new methods make use of soil-moisture datasets in which the nonsequential criterion indicates preferential flow in some of the infiltration events on record. Analysis of particular characteristics of soil moisture responses for these known incidents, for example peak amplitude or rate-of-rise, indicate promising directions for creating new methods, combining methods, or refining existing ones. Method intercomparison contributes further to understanding and to identification of strengths and weaknesses of existing and new criteria.