Data Analysis and Interpretation Using ArcView and Geostatistical Analyst

Collecting information associated with a specific location, georeferenced agricultural data, at the farm is of course only the beginning of the solution to assist in better decision making for improved management and increased productivity. The key to increasing economic yield is using and processing this information, to increase efficiency whether it be by selecting the right crop for the right fields or by calculating the amount of inputs required for a crop throughout a season. The underlying reason that GIS is useful for agriculturalists is that understanding geography is of the utmost importance to people who make a living from the land itself. Analysis of the on-farm data can assist the farmer in making these decisions, thereby earning more income and reducing environmental liability.

A landscape is made up of many interconnected parts, and it is within a GIS that an attempt is made to re-create this complexity within the less complicated structure of a computing environment. Layers of data containing information about a specific subject are the result of this simplification. Layers can include hydrological features, soil characteristics, slope measurements, and indeed anything that the farmer considers to have a bearing on the overall management of a property. Once created, the door to improved efficiency is opened by spatial analysis.

ArcView software provides the new and also more established user with many analytical tools. The ability to compare different layers of information simultaneously often is enough, to the trained eye of the farmer, to spot certain relationships that exist between seemingly unrelated data types. Such an example might be the location of a specific weed within fields and a crop rotation system. Relationships between weed occurrence and rotation type will soon become apparent when historical information regarding field land use is visualized. The ability to extract and highlight certain data features from one layer of information by defining a data feature from a second layer is also very useful. The query builder within ArcView allows the user to investigate information about the mapped information. A well, for example, has many attributes asscociated with it and cannot be defined as simply a location. Knowledge about its depth, the salinity content of its water, and the time required for the aquifer to recharge the well are only some of the different data attributes that can be looked at using this tool. If the data layers exist at the farm level, calculations can be made to show the total land areas under different crops, the value of cropland for insurance purposes, and the degree of risk associated with different crops should various weather events occur.

For more complicated analytical capabilities, Geostatistical Analyst, which is an extension of ArcGIS, provides the ability to more closely interrogate data sets. In addition to enhanced interpolation algorithms, Geostatistical Analyst can predict and measure the statistical probability of the occurrence of a certain feature within a given area. This helps farmers to reduce expensive data collection exercises and instead rely on spatial statistics to provide maps of spatially variable attributes. An example might be soil sampling. The number of samples needed to be collected can be substantially reduced if the statistical variance of field soil types is first calculated. Predictions can be made to indicate potential field areas most at risk from a specific weed by incorporating historical records of infestation and data on soil moisture content throughout a growing season.

Geostatistical Analyst provides a complete set of spatial analytical tools that range from techniques enabling the user to explore the original data to postprocessing the evaluation of uncertainties. Geostatistical Analyst performs two significant tasks: Exploratory Spatial Data Analysis (ESDA) and Surface Creation. The ESDA tools allow you to explore the spatial variance and continuity of the distribution of sample points. By using the ESDA tools, you gain a deeper understanding of your data allowing for better decision making when producing a surface.

The Surface Creation tools use a variety of interpolation methods (inverse distance weighted, trend, radial based, global and local polynomial) and traditional geostatistical techniques (simple kriging, ordinary kriging, universal kriging, indicator cokriging, probability kriging, disjunctive kriging) to produce a variety of output surfaces. There are several spatial analytical tasks that are applied in different combinations for each of the interpolations including variography, declustering, detrending, transformation, cross validation, and validation. Output surface types are prediction, error of prediction, probability map, quantile map, and error of indicators. You can use reliable default parameters for each method to produce the output map.

Geostatistical Analyst allows a statistical insight into the farm data layers themselves. By indicating when a data layer is perhaps less representative of the real-world situation than another, decisions based on results obtained from the nonrepresentative data set can be avoided.