6180 Geoprocessing with Python
Automating Site Suitability Selection for Hop Farming With Python
Introduction
Figure 1 Hop grower’s reference guide from New South Wales, NZ
Hop cultivation requires very specific environmental parameters to be met to be successful on a commercial level. In the US, states in the pacific northwest and dry climates north of the 35th parallel have proven to be successful regions for hop cultivation. Currently Idaho is the fastest growing producer of hops, quickly rivaling the production from Oregon and Washington, the other two states growing hops on a commercial level. Recently, however, states along the Atlantic and throughout Appalachia have begun growing hops in an attempt to achieve flavors and aromatics indicative to their region. Brewers are cultivating their hops on premise or in the vicinity of their brewery. This trend, while growing, is by no means competing with traditional production, but it does introduce the opportunity to perform site suitability analysis in regions outside of current production areas being tracking by the USDA.
Performing a site suitability analysis over a large area is a computationally intensive procedure. In a previous paper, processing and computing various site suitability parameters proved very time consuming and repetitive. The benefits for the agricultural community of performing such functions however are very impressive, especially in the case of introducing a new crop. Hop cultivation in the US was traditionally in the northeast, with the majority of the population. Rampant spread of downy mildew, a disease that threatens hop production brought on by damp conditions, and the settlement of the Pacific coastal region shifted hop cultivation away from the East Coast. Recently efforts have been made to bring this cultivation back, on a smaller scale, to many states East of the Mississippi. Utilizing site suitability analysis would be tremendously beneficial to these efforts. Automating the process and creating a tool that can consume input from various regions in the US would be tremendously beneficial for quickly analyzing prospects for hop cultivation in these regions.
Methods
Automating this analysis involves a great deal of simple but repetitive raster calculations. There are certain geophysical requirements for optimally growing hops. As a reference for these requirements, the guide book published by New South Wales, NZ “Hops, a guide for new growers” was used as many of the parameters are discussed at length for the benefit of beginner growers (see Figure 1). Data, in the form of raster files, was collected for the following biophysical traits important for the cultivation of hops:
Precipitation
Temperature
Elevation
Slope
Landuse
Utilizing ModelBuilder through ArcGIS Pro to create a workflow, these unprocessed rasters collected from various public sources start the analysis, see Figure 2 below.
Figure 2 Site Suitability Analysis workflow in ArcGIS Pro using Model Builder
Each raster is put through a series of calculations to extract the final Site Suitability raster. A simplified version of the Model can be seen in Figure 3. The first series of calculations processes the raw input data before the fuzzy logic is applied. The second stage applies the calculations to create fuzzy membership on all the processed rasters, except for landuse which is a simple boolean output raster, and scales the output from 0 to 1. The final calculation is averaging the five rasters to generate scores for each pixel across the state of Idaho.
Figure 3 Generalized conceptual workflow for automating Site Suitability Analysis
Figure 4 Output raster from Automated Hop Site Suitability Analysis
Results and Discussion
Automating the site suitability analysis has some obvious benefits. Being able to replicate the process for different locations is preferable to going through the process manually. Having a standard requirement for input will create more standardized output for different locations in the case that there is a need for comparison on a larger scale.
There are many challenges with using input rasters that represent difference values and are produced at varying scales. The end result raster has to be scaled to the coarsest input raster, so the level of detail is directly related to the quality of the input data. Figure 4 offers a great representation of this scale, the output raster is quite coarse, with larger areas of Idaho being generalized with the same score.
As with any any process involving different data inputs, the user of this script will be required to examine each element to make determinations on how to scale the inputs. With the elevation data, the user will have to determine the highest and lowest elevations to input into the raster calculations to create those high and low elevation rasters (see Figure 3).
The final output from the analysis offers a slightly different result from the raster created in the Geocomputation analysis. There are a number of reasons for this, such as small variations made in the processing of rasters for this process, however overall the trend in suitable locations is very similar. Areas in the northern Idaho peninsula and the southern Agriculture belt are still identified as ideal locations for hop cultivation. This coincides with the actual location of many existing hop farms in Idaho, which adds validation to the analysis.
Conclusion
There is a tremendous amount of potential for the application of an automated site suitability model for hops. As a niche agriculture product, hop production can face challenges such as find appropriate conditions for cultivation. An automated tool might provide growers and interested partners a starting block for selecting sites.
This analysis could also be used in conjunction with new endeavors to grow hops in regions where they are not traditionally produced, such as the Great Lakes region and the southeast US. These areas are beginning to see an increase in efforts to grow hops locally, but often ideal conditions are not found in these regions. The ability to find the best location in a less than ideal climate might further these endeavors and lead to a proliferation of local hop varieties for consumers.
To make this application more broadly applicable there are some alterations that need to be made to the model. First modification would be to change the input variables from hard-coded files to a user input selection, with some restrictions on acceptable data types. Using percentages, rather than absolute values gathered from the input data, might eliminate some of the requirements on the user to inspect the raster data before beginning the analysis.
Review the full script here.
MSGIS Program Skills:
GIS Analysis
GIS Workflow
Model Building
Spatial Analysis
Communication Skills
Basic Programming or Scripting