This section describes several issues that may emerge in working with the COAST software. It is not nearly an exhaustive list of possible challenges; neither is there an exhaustive set of possible responses to any of the listed challenges. It is intended to illustrate the problem-solving mindset required to use COAST. Many more issues may emerge in every use of the tool, presenting users with often thorny questions about:
- data collection and manipulation prior to running the tool;
- data analysis and interpretation after running the tool; and
- integration of results with public processes to begin taking action about sea level rise and storm surge.
Users can certainly derive their own answers to these questions and learn how to implement the solutions things through experience, but may also wish to recruit outside assistance (as, for example, via the consulting firm Catalysis Adaptation Partners). It is almost always the case that a professional GIS specialist will need to prepare the input data, before the COAST tool can be run successfully.
- I can’t get the software to run, even after I have input
all the data. One issue might be that your projections are
mismatched. The vulnerable asset layer and the underlying
elevation layer need to be in the same projection. You can
convert data layer projections in a GIS prior to running COAST.
- The software still won’t run, even though everything else
seems correct. Another issue might be that the lidar file
you are you are using has holes where the building structure
footprints are, so that pixels in the lidar file within these
holes have no elevation data associated with them. However to
run properly, COAST requires elevation data throughout the
elevation layer. One solution to this is to digitally
interpolate elevation of the building footprints using
elevations of ground immediately surrounding the footprint, via
any of a number of digital correction approaches.
- My assessed values don’t line up with the parcel maps.
Some jurisdictions have very tricky issues here, such as
multiple structures per parcel, each with different or the same
assessed values; unclear divisions among commercial,
residential, and industrial assessed values; and many others.
Some communities keep track of condominium units outside the GIS
system, which may need to be joined to the parcel layer.
Judgment and analytic capability is required to determine the
most effective way forward. It may be to use GIS tools such as
spatial joins and/or relates, dissolves, centroid weighting, or
other technical means. Most data set-up problems are unique and
will require innovation and expertise. When calculating damage
to real estate, the most attractive results can be obtained by
using building footprints to create the extrusions viewable in
Google Earth. However, many jurisdictions do not have building
values associated with the building footprint GIS layer, and
only have such building value data as an attribute in the tax
parcel map layer. Again, creative technical problem-solving will
be required.
- I have completed a run, but the damage results seem way too
high or low. Be sure you have checked the proper vertical
units when you load your lidar image into COAST. You may have
specified feet instead of meters, or vice versa.
- I have other questions about the results. Again, every
situation is unique and will present idiosyncratic challenges of
interpretation. For example, you may find the counter-intuitive
result that for a scenario through say 2050, results show less
cumulative expected damage in the high sea level rise scenario
than in the low sea level rise scenario. This can occur because
of the annualized increments in the cumulative calculations.
When a parcel is permanently inundated by sea level rise, in say
2020, its value is removed from the vulnerable asset data layer.
It is then not available to be compromised by storm surge from
2021 – 2050. For storm surge, the assumption in the model is
that structures are rebuilt after each event, and are therefore
available to be compromised in future years, resulting in this
counterintuitive cumulative damage result. This situation
similar to incurred medical expenses in extended care facilities
(nursing homes): if someone enters such a facility at age 70 and
dies at 71 (equivalent to a high sea level rise situation, where
the “asset” is removed earlier), the facility will not incur
long term medical expenses for at person, whereas if they live
to be 96 and incur medical expenses each year (equivalent to a
sea level rise scenario, where storm surge damages would
continue through the period), cumulative damages will be higher.
With most COAST output, there are many subtleties of this
nature, because storm surge and sea level rise operate on
different time scales and intensities. Evaluating COAST results
demands superior understanding of the phenomena, the way the
tool operates mathematically, and the nature of cumulative
expected damage costing. Working with the results in public
process also requires the ability to effectively communicate all
these things with stakeholders.
- Problems with Lidar Resolution. One of the major contributing factors on how long it takes for the COAST software to finish calculating is the resolution of the lidar data that has been used. If the lidar covers a large geographic area, and/or has a fine grid cell size (such as 1m or 1 ft), COAST may be slowed to a point where calculation will take many hours or even days, which may be unworkable. Using a GIS software package outside of COAST may be necessary, to crop or resample your lidar data and help COAST calculate in a reasonable amount of time. And note that some resolution sensitivity is already supported in the program, so that if the program is slowing down too much, you can rerun the scenario at a lower resolution. This functionality is located on the "General" tab of the Coast Model Parameters dialog box, where there is a slider bar with options for prioritizing processing speed vs. higher resolution. If a specific resolution value is desired, however, the lidar data may need be resampled beforehand, because the slider presets may not include the particular resolution of interest. This may require the services of a GIS professional.