Lesson 7

Project 7:  Delineating Vineyards Using Raster GIS Analysis

 
This was a really interesting lesson and challenge.  Using the capabilities of ARCmap, we analyzed a set of environmental variables to determine the most suitable location(s) for the location of vineyards for a small portion of the Napa Valley in California.

 

The first steps taken were to convert a set of vector data to raster data.  The Land Use polygon was reduced to a minimum number of unique feature types through the use of a 'dissolve'.  Next this 'dissolved' layer and a hydro layer were converted from a vector format to a raster format.  This conversion resulted in a map grid of squares, each 10 meters by 10 meters, for each area represented.  These 10 x 10 squares will form the basis for the rest of the analyses performed in this lesson.  By layering different criteria on top of each other, we will arrive at a map of squares that meet all the environmental requirements necessary for a vineyard.  Partial results of the vector to raster conversion are shown in Figure 1 below.  This map shows the flood plain as represented by a raster grid, along with a raster grid showing a set of rivers and streams.  These layers kept partially transparent to show the underlying elevation grid layer.  Also, the points on the map represent the different sampling stations for both soil and weather conditions.  These will be used in later analyses and are shown here for information purposes only.

 

Figure 1:  Results of Initial Vector to Raster Conversion Showing Flood Plain and Sampling Points

 
The next steps were to utilize the the soil and weather data to create a set of layers representing their unique attribute conditions.  Separate raster layers were created for soil depth and soil drainage conditions as well as for minimum temperatures and maximum wind speed conditions.  In addition, the elevation data was used to mathematically simulate the effects of the sun and shade on the underlying topography.  Using the ability of the software to make some layers partially transparent, a three dimensional effect can be created.

The next major step was to reclassify all the different raster layers into sets of binary conditions, those that meet the proper criteria for a vineyard and those that do not.  In this way, each cell can be evaluated, which when combined, can provide an overall picture of those areas that are suitable for locating vineyards.  Figure 2 below shows the results of this process.  In this figure, those conditions NOT meeting the criteria are colored and made semi-transparent so you can see the different layers.  The pink area shows where the soil is not deep enough, the light purple area shows where there isn't enough soil drainage, the light green area shows where the temperatures are too cold, and the light blue areas show where the winds are too high.  The darker purple area follows a 100 meter stream buffer.

Figure 2:  Initial Analysis of Areas Not Suitable for Vineyard Locations

 

Finally, all the criteria are combined into a single "go/no-go" analysis resulting in a listing of the areas that are suitable for locating a vineyard.  In Figure 3 below, the greyed areas are the combined "no-go" areas that exhibit criteria not suitable for planting.  The blue areas mark the limitations of the flood plain and the green areas are the resulting areas that are suitable for planting vineyards.

Figure 3: Areas Suitable for Vineyards

 
Using the results achieved above, a separate layout was created, as is shown in Figure 4 below.  This layout was prepared for printing and could be included in a report to a client on the sites found.

Figure 4: Report Page of Vineyard Suitability Analysis

 

 Try This

 
The Try This segment added a new wrinkle, that of a political dimension.  In this case, we were asked to consider whether our suitable vineyard sites were available on public vs. private land.  The implicit assumption is that only sites that are available on public land would be suitable, however, with the budget crisis currently being faced by the State of California, leasing public lands might be an option for locating revenue (and tax) producing vineyards.  For this reason, I chose to show in Figure 4 both the private as well as the public lands that could make suitable vineyard sites with the yellow areas representing the private land sites and the green areas representing the public land sites.

Figure 5: Public vs. Private Site Analysis

From previous computations, it was determined that 1295 acres of land met the criteria originally specified for locating vineyards.  When the effect of public vs. private land was incorporated into the analysis, I found that approximately 739 acres could be found on private land and approximately 557 acres could be found on public lands.  This is documented in the layout report, Figure 6, shown below.

 

Figure 6:  Suitable Vineyard Site Report

 

Data Issues and Considerations

 
While this analysis was interesting and instructive, I don't believe it is particularly useful.  Perhaps as an overall view of potential sites, it might have some validity, but to be truly a valuable analysis additional, more detailed data should be added including plot maps, detailed property lines, and information at a much more detailed level.

Additionally, the sampling points, while simplified for the purposes of this exercise, are quite spread out over the map area.  In some cases, inferences about soil and weather conditions are being made based on data points well over a mile apart.  Inferences over this wide an area can mask conditions that may occur in areas that might support their own micro-climates.  To get a better analysis of the soil and weather conditions at a particular geographic location, it would be better to draw on data from sampling points that are much closer together.

Overall, this exercise was great to understand how to build information using raster analyses, however, to be truly useful much better and much more specific data would need to be utilized.