Since soil water content controls the fundamental land-atmosphere water exchanges, studies of distributed soil moisture content are critical in understanding the hydrological cycle and the energy fluxes at the landscape scale. Developing a global soil moisture (SM) observation system is of paramount importance for broad scale studies of biophysical processes affected by global climate changes in temperature and precipitation because the importance of the water cycle in food production, weather, fresh water availability, stream and watershed studies, ecological processes, and others.

Landscapes are complex systems formed by spatial units (fragments) in which biophysical factors fluctuate according to soil and landuse characteristics. This research builds on the concept that the integration of remote sensing (RS), hydrology, geostatistics, GIScience and landscape ecology will expand the capabilities of RS research and ground readings to explore the relationships between fragments and process in complex landscapes. Through a GIS-RS analysis of fragmentation, I investigate the contribution of soil and landuse classes to landscape complexity as a way to understand the environmental factors influencing SM point readings of a network of field stations at the Little River Watershed (LRW) south Georgia, US. The effects in the spatial and temporal variation of SM and ground temperature caused by the landscape complexity are analyzed within areas equivalents to the pixel sizes of four environmental satellites. Also through a year long field data collection, I establish the spatial and temporal variation of SM and ground temperature within landscape fragments belonging to five land uses adjacent to the stations of the network at the LRW.

The methods include the creation of a GIS database for soil and land cover at <10m spatial resolution, distance and area spatial analysis for field point data, hydrological modeling of water infiltration processes for five different soil types and a generalize linear model for spatial and temporal variation within and among landscape fragments and land uses. The results are supported by rigorous statistical analysis including analysis of variance, t-test, descriptive statistics, correlation analysis, time stability analysis, and generalize linear model. The most important results include: 1. Criteria to define the appropriate pixel sizes of a satellite instrument to study soil moisture and ground temperature at the LRW. 2. Understanding the contribution to landscape and local complexity added by soil and landuse classes and their weight in point readings. 3. The assessment of SM responds by different soils and land cover combinations. 4. The study of temporal stability of soil moisture in fragments matching the size of a Lansat TM or ETM+ pixel size. 5. The correlation between point readings and adjacent fragments under different landuses. 6. The assessment of temporal and spatial variation of SM and ground temperature within and among homogeneous fields matching the size of a Landsat TM or ETM+ pixel size at the LRW.

Overall, this research contribute to satellite study of the water cycle by proposing a landscape ecology–hydrology approach to the problems of point reading interpolation and RS of soil moisture-soil temperature within complex landscapes. This work contributes to the knowledge of the variation of biophysical variables for the remote sensing study of energy fluxes and the water cycle within complex landscapes.