In most crops, growth can proceed unimpaired and crop yield can be maximized only when the soil moisture potential remains high (and water remains readily available) continuously throughout the growing season.

Water is critical to crop productivity since crop yields generally increase linearly with water transpired by a crop (Howell, 1990). Excess water (waterlogging) can induce nutrient and aeration stresses and encourage pests that reduce yield and quality (Wesseling, 1974; USCA-ERS, 1987). Water management is also critical to water quality because techniques to optimize water relations for plants can also impact fate and transport of pollutants to surface water and groundwater. Naturally, the adequacy of water for plant growth is primarily related to the amount, frequency and distribution of rainfall, soil properties as they affect processes that regulate soil water availability to plants and landscape properties that regulate the hydrologic cycle within a watershed. Three approaches to precision water management are therefore apparent:

1.     Variable rate irrigation

2.     Matching agronomic inputs to water availability defined by soil and/or landscape properties and

3.     Drainage.

1.     Variable Rate Irrigation

Hillel (1990) defines a well-managed irrigation system as one that optimizes the spatial and temporal distribution of water so as to promote crop growth and yield and to enhance the economic efficiency of crop production (maximum net return). He further states that since the physical circumstances and the socioeconomic conditions for irrigation are site specific (and often season specific) in each case, there can be no single solution to the problem of how best to develop and manage and irrigation management.

Considerable progress has been made with variable rate irrigation systems primarily with sprinkler irrigation provided by center-pivot and linear-move machines. These site-specific irrigation systems require high spatial resolution (currently 10-30 m) achieved by adding more discrete control between contiguous elements of the machine, all at higher costs that those of current systems. Variable irrigation is coupled with precision nutrient and pest management via chemigation, in part because variable irrigation facilitates in creased management precision in space and time and in part because it may not be economically feasible to site-specifically manage only for water. The uniformity of chemical application depends on the uniformity of water application, requiring injection equipment that can vary the amount of chemical injected into the boom in proportion to the fowl rate water in order to achieve the desired chemical application rate.

Success of precision irrigation management has been achieved with regard to application control. The key to the agronomic success of precision irrigation management depends to large extent on how well the water needs of the soil-plant system can be measured or predicted and the accuracy of water application prescriptions. The value of precision irrigation management depends on whether increased profits and the reduction in pollutants more than offset the cost of increased resolution needed in irrigation systems to apply irrigation and chemigate site specifically.

Due to the random variability in water application distributions due to wind, start-stop operations of the self propelled machines, and sprinkler pattern variations combined with the low cost of water and N fertilizers, it is probably not economically feasible to site-specifically manage only for water and/or nitrogen.

2.     Soil-Landscape Water Management

The potential for precision management of agronomic inputs increases with spatial variability in water availability within a field. Differences in water availability within a field are governed by

• The occurrence of dissimilar soil types
• The presence of soil degradation processes (e.g., erosion, compaction and salinization) and
• Variation in landscape

The evidence for spatial variation in water availability is clear. Hanna et al. (1982) reported that north-facing slopes had 20% more available water in soils than south-facing slopes throughout the year. Where as soils on east-facing slopes were the driest. Crop yields are often highest in the lower slope positions where soil water and nutrient contents are higher. Eroded soils often have lower infiltration rates and lower available water than their non-eroded counterparts. Some portion of landscape variability can be attributed to the variation of soil properties with landscape position, whereas some is attributable to redistribution of water within a landscape due to either runoff or subsurface horizontal flow of water. Compacted soils reduce infiltration or restrict plant roots, thereby limiting water availability to plants. Areas of high salinity are known to reduce yields.

Knowledge of the spatial distribution of water availability can be used as a basis for site-specific input recommendations. There are three approaches for mapping soil water variability

• County soil surveys
• Interpolation of a network, usually a grid, of point samples to estimate spatial distribution of soil properties or water content and
• Soil-landscape models to estimate spatial patterns of soil water availability.

The presence of small scale spatial variability in soil physical properties and the high cost of network sampling may limit its use in mapping water availability. Site-specific soil water monitoring is used to some extent as a basis for variable rate irrigation and land in landscape studies. Statistical models of soil-landscape relationships offer opportunities to map spatial patterns of soil properties where relief or some landscape attribute is a primary factor contributing to soil variability. Soil-landscape models are important because terrain modifies the distribution of hydrologic and erosional processes (i.e., soil water content, runoff and sedimentation) and soil temperature in fields, all important is regulating crop productivity of topography wit a regular grid of elevation observations is referred to as a digital terrain model (DTM) when attributes of a landscape are of interest and a digital elevation model (DEM) when merely relief is represented. A DTM allows the estimation of derivatives of elevation including slope, curvature, aspect, catchment area, and surface drainage proximity variables that correlate to soil and land qualities.

The value of DTM is that is increases the resolution of soil maps for use in site-specific management and in environmental modeling by using terrain attributes to spatially distribute estimated soil attribute data. Therefore, terrain modeling efforts have focused on its application to soil survey to model and depict the spatial variability of soil horizons in reference to the topographic surface and spatial application of simulation models to evaluate current and potential management practices regarding their effects on crop production and the environment in space and time. The extent of use of soil-landscape models is currently limited in field applications of precision agriculture. However, high-resolution DEMs can easily be created using DGPs and laser-based systems with high vertical accuracies. As elevation maps become available, soil-landscape modeling techniques such as DTM will be increasingly used in precision agriculture.

3.     Drainage

Poor drainage is often cited by farmers as a source of yield variability within field. Many options for drainage currently exist and can be applied site-specifically. Therefore, there is little need to design site-specific drainage practices. The decision to install drainage is economically, not technically, limited, Regardless of scale, the decision to drain hinges on the expectation of returns on investment that exceed costs of installation. Site specifically, the cost of draining portions of fields or small isolated areas may be higher because the yield depression due to poor drainage can be accurately assessed if sufficient years are included in the calculation. Drainage, therefore, is a site-specific, economic decision based on the conditions at each site and cannot be generalized.