Skip to main content
ARS Home » Research » Publications at this Location » Publication #262466
Title: Using measures of information content and complexity of time series as hydrologic metrics

Author
item FENG, PAN - University Of Maryland
item Pachepsky, Yakov
item HILL, ROBERT - University Of Maryland

Submitted to: European Geosciences Union General Assembly Proceedings
Publication Type: Abstract Only
Publication Acceptance Date: 1/28/2011
Publication Date: 4/4/2011
Citation: Feng, P., Pachepsky, Y.A., Hill, R. 2011. Using measures of information content and complexity of time series as hydrologic metrics. [abstract]. European Geosciences Union General Assembly Proceedings. 13(88).

Interpretive Summary:

Technical Abstract: The information theory has been previously used to develop metrics that allowed to characterize temporal patterns in soil moisture dynamics, and to evaluate and to compare performance of soil water flow models. The objective of this study was to apply information and complexity measures to characterize temporal and spatial variations of streamflow. USDA-ARS experimental watersheds in Georgia and in Vermont were selected as the study sites, Both watersheds comprised several monitored sub-watersheds with more than 30-year continuous data records of precipitation and streamflow. Information content measures were the metric entropy and the mean information gain, and complexity measures were the fluctuation complexity and the effective measure complexity. These measures were computed based on the binary encoding of 5-year daily, half-daily and quarter-daily time series of streamflow and precipitation in the watersheds and sub-watersheds. The empirical probabilities of joint or sequential appearances of the binary symbol sequences were input variables to compute the information content and complexity measures. The information content of streamflow time series was much smaller than the one of precipitation data, and had the higher complexity, indicating that watersheds acted as filters of information brought by precipitation. The large variations in information content and complexity measures of streamflow time series with different time scales indicated the significant temporal-scale effects on the measures of streamflow. Moderate spatial scale effects of spatial resolution on the information content and complexity measures of streamflow time series were observed. The latter effects varied depending on sub-watershed location, size, and land use. The information content and complexity measures can provide useful complementary knowledge about the hydrologic system complexity and patterns of its behavior in space and time. The use of these metrics to evaluate and to compare performance of soil water flow models presents an interesting research avenue.