Spatial and Temporal Analysis of Soil-landscape and Water Quality in the East Mahantango Creek Watershed
Open Access
- Author:
- Kang, Shujiang
- Graduate Program:
- Soil Science
- Degree:
- Doctor of Philosophy
- Document Type:
- Dissertation
- Date of Defense:
- September 22, 2006
- Committee Members:
- Hangsheng Lin, Committee Chair/Co-Chair
Gary Walter Petersen, Committee Member
William Gburek, Committee Member
David Russell Dewalle, Committee Member
Patrick M Reed, Committee Member - Keywords:
- water quality
landscape
soil
temporal analysis
spatial analysis
stream order
stream network - Abstract:
- Nonpoint source pollution is a world-wide concern for its exacerbation of water quality. Understanding spatial and temporal distribution of nonpoint source pollutants associated with soil-landscape variability can assist in predicting and monitoring water quality in watersheds. Variability of landscape features, soil properties, and water quality in the stream-order-based sub-watersheds and buffer zones was investigated in the East Mahantango Creek Watershed (EMC) located in the Ridge and Valley Physiographic Province of east-central Pennsylvania. In addition to spatial analysis of soils and landscape features in the EMC, a systematic temporal analysis (traditional time series analysis, spectral analysis, and wavelet analysis) was conducted for the WE-38 Watershed, a third-order sub-watershed within the EMC using 15 years of monitoring data of hydrology and water quality. Variability of elevation, slope, and land uses along different order sub-watersheds is influenced by the watershed topography (ridges and valleys). The coarse data of surface geology and the STATSGO soils database do not reflect actual variations among different order sub-watersheds. With more detailed SSURGO soils database, the area-weighted method recommended by the USDA-NRCS tends to smear out soil variations. A series of soil property trends were observed in buffer zones of different order streams along the stream network. Within the buffer distance of 150 m, the top two soil layers in lower order stream buffer zones have higher clay content and higher bulk density than those of higher order streams, but have shallower depth to bedrock, lower available water capacity, and lower organic matter content compared to higher order streams. Beyond the buffer zone of 150 m, variations of soil properties appear to stabilize. The buffer distance of 70-100 m seems to be a threshold value that reflects significant changes of soil properties in the buffer zones along the stream network. These trends of soil properties in the buffer zones of different order streams can be explained by the effects of local topographic features and soil transport processes such as colluvial and alluvial process. The soil variability analysis along stream buffer zones provides useful information regarding soil distribution along stream network that would benefit buffer zone management and environmental modeling. The three zones (hillslope zones, transitional zones and floodplain zones) reflecting soil formation and transport can be applicable to generate representative soil properties for environmental modeling and management at watershed or catchment scales. Baseflow water quality in different order streams was investigated using one year of monitoring dataset together with two historical sampling datasets. Overall, a negative correlation between baseflow nitrate concentration and stream order was observed; however, significant spatial-temporal variations existed. Nitrate concentration was also linearly correlated to agricultural land use. The decreasing trend of nitrate concentration with increasing stream order can be explained by the reducing nitrate sources related to decreasing agricultural land use and possible increased in-stream processes along the stream network (such as denitrification and dilution). A positive power function relationship was identified between nitrate loads and stream order. This was mainly due to the power function between discharge and stream order. Overall, it appears that stream order is a helpful index of predicting and monitoring nitrate concentration and loads during baseflow in this agricultural watershed. Stream flow and concentrations of nitrate, sodium, and chloride collected in the past 15 years showed different weak seasonal trends. Both sodium and chloride had a reverse sine-curve trend to flow. Autoregressive analysis demonstrated that high autocorrelations of all variables were within six days (coefficient > 0.5). Nitrate and flow appeared to be with high cross-correlations. Negative cross-correlation between sodium and flow was also found. Sodium was damped compared to chloride. Although no fractal property was found for flow and nitrate with spectral analysis, they showed highly coherence from their higher variations for short wavelengths. A multi-variate temporal model of nitrate was developed to assist in predicting and monitoring nitrate dynamics. Using wavelet analysis, we analyzed temporal characteristics of three hydrological variables (precipitation, stream flow, and well water level) for three periods (15 years, 3 year, and a hydrological year). For three unevenly sampled water quality variables (nitrate, chloride, and sodium), the weighted wavelet Z-transform (WWZ) method was employed. No strong temporal pattern of precipitation was found for all three periods. For the 15 year of continual monitoring datasets, a strong consistent annual temporal pattern of well water level and an intermittent annual temporal pattern of stream flow were observed. For the relative short time periods (three years and a hydrological year), strong seasonal temporal patterns of stream flow and well water level were noticed. Using the WWZ method, seasonal patterns of three stream water quality indicators can be associated with their seasonal shifts. Sodium and chloride peak concentrations in the winter season were well captured by the wavelet analysis. Using wavelet analysis, the degree of temporal patterns was found to decrease in the following general order: well water level > stream flow > precipitation, and nitrate > sodium „d chloride, suggesting an increasing dynamics of processes involved in ground water, stream water, and rain water systems, and a decreasing temporal periodicity associated with nitrate, sodium, and chloride concentrations in the stream with the weighted wavelet Z-transform. Wavelet analysis together with other time series analysis tools can be a potential tool to quantify hydrological and water quality dynamics from a stochastic perspective.