REMOTE SENSING OF LAND USE/COVER FACTORS INFLUENCING SOIL EROSION IN THE LITTLE CONESTOGA WATERSHED, LANCASTER COUNTY, PENNSYLVANIA

The Little Conestoga basin is the highest producer of suspended sediment per unit area delivered to the Conestoga River, Susquehanna River and Chesapeake Bay. In this study, we used 15 ASTER scenes (15-30 m resolution), as well as co-acquired MISR and MODIS imagery (250 m resolution) spanning the period between April 2000 and October 2005. The ASTER data were analyzed by using the V-I-S remote sensing model {Ridd, 1995, Int. J. Rem. Sens., 16(2): 2165-2185} to monitor land use changes as linear proportions of green vegetation, impervious surfaces and bare soils. Other data sets examined included slope derived from LIDAR data, cumulative rainfall data from the National Weather Service, and high resolution (0.6 m) color infrared (CIR) orthoimagery acquired in 2002.

ASTER was useful for distinguishing dry, bare soils from green vegetation and impervious surfaces such as roofs, roads and parking lots. Soil samples from the area contain varying amounts of quartz, muscovite, illite, kaolinite and hydroxy-interlayered smectite, the latter of which controls soil crusting properties and subsequent erodability. The distribution of these minerals also was mapped using ASTER, despite limited spectral resolution and mixing with dry vegetation material in several fields.

Assuming static impervious cover relative to seasonally changing vegetation, changes in bare ground relative to crop cover were plotted with time. The results show that the proportion of bare ground to stabilizing crop cover is high throughout much (September to May, 75%) of the year, with minimal bare soil exposure occurring during the height of the summer growing period (June to August, 25%). Using the larger-scale MODIS and MISR imagery peaks in bare soil cover coinciding with periods of heavy rainfall (e.g. Spring storms or snow melt and Fall tropical systems), correlate well with periods of high turbidity in the lower Susquehanna and upper bay. Conversely, periods of high bare soil cover coinciding with low rainfall or low bare soil cover coinciding with high rainfall, correlate with periods of lower turbidity in the lower Susquehanna and upper bay. Analysis of LIDAR slope data shows that the Indian run tributary contains the highest percentage of slopes between 8-10 degrees, which are most susceptible to rill and gully erosion, as resolved in the CIR imagery.