Submitted to: Hydrological Processes
Publication Type: Peer reviewed journal
Publication Acceptance Date: 2/4/2013
Publication Date: 2/27/2013
Citation: Martinez, G., Pachepsky, Y.A., Vereecken, H. 2013. Temporal stability of soil water content as affected by climate: a simulation study. Hydrological Processes. DOI: 10.1002/hyp9737. Interpretive Summary: When soil water contents are measured across field or watersheds, certain locations have consistently higher or lower soil water contents than the average for the sampling area. When each location persistently deviates from the average by some percentage, the soil water content distribution is called temporally stable. What controls the persistency of those deviations is currently unknown. We undertook a simulation study in which type of soil, type of climate, and the sampling frequency were inputs into a soil water flow model. We found that the robustness of the temporal persistence depends on the interplay of the type of soil and type of climate, and also on the duration of the sampling period. Results of this work can guide the evaluation of a site-specific soil water sampling regimens. It will be useful to multiple groups of specialists who use the temporal persistence to up- and down-scale soil water contents in remote sensing, to infill missing data, to delineate field management zones, to improve hydrologic modeling results, to design adequate treatment experiments with replicated plots and to design sensor networks and optimize the number of sensors.
Technical Abstract: Temporal stability of soil water content (TS SWC) is a natural phenomenon that recently attracts attention and finds multiple applications. Weather and climate are usually mentioned as a factor of TS SWC, but its effect is far from clear. The objective of this work was to use soil water modeling to assess the effect of climate on the TS SWC. We selected four representative climates present in USA and simulated the multiyear soil SWC for sandy loam and loam soils, both having the lognormal spatial distribution of the saturated hydraulic conductivity. The CLIMGEN code was used to generate climate-specific weather patterns, and HYDRUS6 code was used to simulate SWC. Four different methods were applied to select the representative location (RL). High rejection probabilities were found for the hypothesis that the four climates cause the same variability of mean relative differences (MRD) of SWC. The probability that the variance in MRD depended on sampling frequency was always higher than 89% in the sandy loam soil and was above 97% in the loamy soil. Lower probabilities were observed for the robustness of the MRD. The inter-annual difference in MRD variation from short and intensive summer campaigns was highly probable for all climates. There were more coincidences in the RL selected for the sandy loam soil than for the loamy soil. Among the methods used for selection of RL, those that are based on a minimum difference with MRD of zero showed more consistency than others. The TS appears to be the result of the interplay between climate, soil properties, and survey protocols. One implication of this factor interaction effect on TS SWC is that a simulation study can be useful to decide on the feasibility of including a search for the representative location based on TS for a specific site.