As I started to describe in class, my project will be dealing with output results from the model software ENVISION. ENVISION is a GIS-based tool for scenario based community and regional planning, and environmental assessments. It combines a spatially explicit polygon-based representation of a landscape (IDUs or Individual Decision Units in my case), a set of application-defined policies, landscape change models, and models of ecological, social, and economic services to simulate land use change and provide decision-makers, planners, and the public with information about resulting effects.
The ENVISION project I am involved with is centered on Tillamook County and its coastal communities. Through a series of stakeholder meetings (which have included a range of people such as private landowners and state land use commissioners) our group identified several land use policies to implement in the ENVISION model. The policies were then grouped into three types of management responses: the baseline (or status quo), ReAlign, and two types of Hold the Line (high vs. low management) scenarios. These policy scenarios have been combined with current physical parameters of the coastline such as dune height and beach width, and will be also linked with climate change conditions at low, medium, and high levels for the next 30, 50, and 100 years.
Since ENVISION is GIS-based already, I am having a tough time coming up with a problem that complements the project in ArcGIS. ENVISION does a great job of visualizing the changes expected for each location along the coast via videos, graphs (see below), and can even include economic estimations.
Therefore, it may be best to explore the capabilities of software like R to analyze the output data. One idea would be to calculate the probability of occurrence for these different events and total number of occurrence. I need to take a deeper look into how these events are calculated to begin with, and determine the inherent estimates of probability and uncertainty. This type of analysis would help determine whether this type of exercise is beneficial for stakeholders and would help answer their own questions of trust in the results.
Another idea would be to focus on specific results from ENVISION and try to determine exactly how one policy is affecting the coastline and creating such disparate results. For instance, the graph below shows Numbers of Flooded and Eroded Structures in Pacific City under three types of scenarios. What is causing the large number of eroded/ flooded structures between 2035 and 2040? Why is there such a small difference between ReAlign and Hold the Line strategies if they are employing such different options? Some of these questions may be answered with a greater understanding of ENVISION, however, these are the types of questions that may be asked by stakeholders and it would be prudent to provide more quantitative answers that ArcGIS or R could glean.
It seems like most of the outputs you have presented are time series estimates, either at the county or city level. I think a great way to complement these graphs would be to create maps of high impact areas along the coast. You could do this for the 30, 50, and 100 year estimates you mention, and also for the 3 different policies. Or a panel of time series hotspot maps.
Visually, I think these might be the most beneficial products for stakeholders. Have you spent much time creating final map products in Arc? This may be a great opportunity for you to use some of the analyses done in ENVISION and create some professional products.
ENVISION uses shps just like Arc does, so any information that has been included in the ENVISION model can be pulled into an Arc map. The graphs were just a quick way to show what kind of information was available to me. Anyways, Julia also mentioned using hotspot analysis to look at the spread of data. I’m thinking I can compare historical to future predictions as they are generated by the rest of my group. Luckily I have a few years worth of map-making under my belt, so at least I can make snazzy figures!
Have you had to create maps for stakeholders before? Any words of wisdom you could pass on?
I can provide you with some assistance using R since I have a good deal of experience with it.
I have a clarifying question regarding the type of data output from ENVISION. You had mentioned that you would like to look at probabilities of events occurring, in order to help provide some advice I need to know what type of data the ENVISION software provides for you?
Thanks for the help.
So ENVISION works on a system similar to Arc, as in it uses shps as the main source of data. These data are then manipulated via plug-ins to generate future predictions based on predetermined parameters including the stakeholder generated policies. My colleague working on the ENVISION side of things has informed me that they can generate probabilities in ENVISION. Therefore I’m trying to come up with a more useful project for this class.
Julia, and Alex above, have suggested using hotspot analysis to generate figures on where the greatest coastal flooding and erosion has occurred and would be occurring in the future. I can then look at the associated property values an run another hotspot analysis. Can you think of any other figures stakeholders may want to see?
I think a hotspot analysis would be an interesting approach. In addition to looking at property values, you could also looks at schools or population densities. In regards to creating map products for stakeholders, I found a couple of pointers to be useful (not sure how much cartography you have had, so sorry if any of this is repeat for you):
1. Use a meaningful color ramp. In hotspot analysis, using a red to blue color ramp is useful for conveying hot to cold, yet beyond this familiar color association, it is better to use a single color ramp that progresses from dark to light.
2. Be very explicit about what the map is conveying and design your maps for your audience. I usually like to run my maps by a person unrelated to the project to be sure the message being conveyed in the map is clear.
3. Find some way of communicating the uncertainty in your maps. If you have probabilities of events occurring on a spatial extent, you can create a probability map to accompany your hotspot maps.
The earlier comment about schools… I am not sure how many schools there are on the beach. But the idea of comparing damage to private investment vs public investment is a good idea. Specifically, roads, culverts/bridges that may be damaged by erosion would be of strong interest to city managers. It may be that constructing a retention berm would be justifiable if the road was in danger of being washed out. You could conceivably add 30ft buffers to the roads and determine if any roads were in a “danger zone”.