TOOLS IN DEVELOPMENT

Fuel Bed enhancement of Most Similar Neighbor Analysis

One of the primary weaknesses in the current modeling of fire behavior for FARSITE is poor dead and down fuel bed data.

The INFORMS team along with Roger Ottmar of the PNW Research Station and Rob Seli of the Missoula Fire Lab are working with the FSVeg team to create a lookup table containing all the Fuel Photo Series photos along with the ability to summarize stand fuels transects. Initially these will be included in landscape fuel maps for those stands they were sampled in.

The problem we are trying to address is the fact that default fuel beds used in FVS/FFE can be quite limiting in predicting actual fire behavior for project or going fire planning. This can be improved by inserting data into the FVS Keyword file using the FUELINIT keyword. This sets the fuel bed for that stand and can improve the fire behavior modeling considerably. The reason it will not be tied to imputations with MSN initially is that field reviews indicate stand with similar live tree over story often do not have similar dead and down fuel beds.

Further research will be done to use MSN to impute accurate fuel bed data into polygons. It is currently believed that a secondary MSN run done on initial MSN imputed data may yield accurate fuel bed imputations. This research has yet to be done because further development of this application is required and no area is sufficiently well sampled at this point to use for this type of analysis.

Back to Top

Inclusion of Insect and Disease Modeling in INFORMS

Another area of weakness in fire and fuels as well as other ecological modeling is the inclusion of insect and disease effect in the data.

The INFORMS team is working with the Forest Health Technology Enterprise Team and other forest health and fire applications developer to include pest modeling in Post MSN imputed landscapes. It is well known that trees that are weakened or dead due to pests must be considered differently for many reasons. One of which is that they behave differently during fire or fuels treatments. To data including pest data in any vegetative analysis has required field sampling and data entry of pest sample information.

There are disparate souses of areas that are infested with pests but they have not been brought together in an automated fashion in the past. The FVS application includes pest models but assigning them to the correct stands without each of those stands being sampled has been impossible in the past. With the development of the MSN package there is effectively a stand exam on all forested stands that a pest model could be applied to. The problem is that not all stands have pests and those that do not have the same level of infestation nor the same pest effecting them. The concept we are working on is that of using the pest aerial survey that is done each year identifies these stands spatially and through the INFORMS applications may be able to link the proper pest extension to the proper stand polygons.

This theory is currently being tested on National Forests in Northern Idaho. If initial tests are successful the process will be further automated and deliver as a new package in the INFORMS application for Fire and Fuels planning and other NEPA analysis.

Back to Top

Ability to Integrate Multiple Vegetation Layers with data for use with MSN

One of the current problems is that some sites that want to use the MSN package have data associated with an old vegetation layer then a new vegetation layer (redeliniation) was done and new data is collected in association with the new layer. Selecting either one is not ideal for use with MSN.

The INFORMS team is working with the FSVeg staff, Region 2 and Nick Crookston to develop a method to solve this problem that occurs on a periodic bases in many places. The ideal solution is to redelineate while conserving old polygons that are deemed still accurate and that have a stand exam for those polygons. This process makes redelineation much more expensive and time consuming. For these and many other reasons this is not done.

A situation is then created that makes the old stand layer which may have a lot of stand exam data not desirable to use for analysis. The new stand layer that is more desirable for use of describing current vegetation has no or very few stand exams associated with it for many years after the delineations are completed.

One possible solution is to use the INFORMS MSN methodology to impute stand data for like stands from the old vegetation layer into the polygons of the new layer along with those exams from the new layer.

This technique has not yet been attempted but a design for its implementation has been conceived. Region 2 of the Forest Service has this problem on many Forests because that Region implemented a region wide redelineation and photo interpretation program several years ago and has not had the fund to do large number of stand exams on those polygons since but has may stand exams associated with an old layer.

This concept is currently being tested in that Region and if successful will be included as a new tool in INFORMS in the near future.

Back to Top

Create Multi Cycle Base and Mid Scale Layers

One of the common problems in relating Strategic Planning objectives to Project Scale analysis is the fact that the maps used for both scales are created independently. In other words that is no relationship between them.

The INFORMS team along with Region 3 of the Forest Service is working on developing base scale vegetation mapping with MSN as described in the vegetative mapping (link) section of this website. Then based on algorithms designed by FVS Variant create a methodology of classification and devolving of polygon lines on the base scale to create the mid scale maps for strategic planning.

The power of this method is the relative ease of knowing where specific strategic panning results should be applied because of the tight relationship between the basic and mid scale layers.

The second powerful value of this method is that with the power of INFORMS to grow landscapes and produce base scale maps for future conditions future mid scale maps will be able to be produced just as the base scale maps can.

Back to Top