2012 AFE Overview of Past, Current & Future LANDFIRE Data Products and Methods (Archived video)


My name is Matt Rollins and I’ve been working with LANDFIRE, for what, more than 10 years now almost, like a few other people in the room. But now I don’t work for LANDFIRE. I moved on. I’m pretty much 100 percent beaurocat. But I was asked to give an introduction presentation, and I’m happy to do it, excited about it. It’s
good to see a lot of new faces in the room, which is good. A lot of times we see the usual
suspects at these things. And I can guarantee that this’ll be
one of two flow charts in the presentation. You can probably imagine that LANDFIRE has got a flow-chart rich environment and I’m sure that alot of other presentations will have a flow chart. So, stay awake. It occurred to me that many people might not even know
what LANDFIRE is. Just really briefly: LANDFIRE is a national
program. It’s been in place since about 2000 and the goal that program is to develop
geospatial data layers for all 50 states, that are useful for fire management
in terms of decision-support and planning, and wasn’t developed to feed any one specific program and
we went to alot of time and effort to make sure
that the products were useful across the gamut of fire management, and,
it turned out later, for national research questions in
general. So for the talk I was provided at the last minute, witjh some extra slides from one of our business leads, and I stuck those in there. We’re going to start with some thoughts on and fire and ask some questions, go through some of the history really briefly of the program. Talk a little about LANDFIRE 1.0 which is the first national
set of layers that came out for circa 2001, and then the two sets of
refresh data layers that came out later, and then the current phase of updating which will bring layers up to currency as of 2010. I’ll talk a little about the future of LANDFIRE, then we’ll l revisit some of those questions and thoughts on LANDFIRE. I had a bunch of snarky stuff in there and I took it out and then I put some back in and some stuff on that on side
is some commentary that I won’t really address. If you want to talk about that later
during the questions I’m certainly willing to. So, how many people in the room have heard about LANDFIRE before today? We’ve got lots of people. Anyone not heard about LANDFIRE before today? Nah. Good. How many of you have used LANDFIRE directly? — used it for your job? How many
have used it sort of indirectly? (I’m not sure what Henry meant by that.) So, alot of users here in the audience. Are national datasets useful, or are they not needed? Is it more important to have local datasets? The intent of LANDFIRE was to have a seamless, consistent set of data that people could
use that were separate from administrative or
organizational boundaries, much like fires burn across boundaries, etcetera. We don’t need to have a one-on-one answer to that question. How many of you feel that the quality of LANDFIRE is good? How many people think that it’s bad? Or
poor? How many of you are not being honest? Okay. We are in a judgment-free zone here as far as
i’m concerned. so How many of you have actually evaluated
the quality of LANDFIRE in your local areas or for your local projects? That’s good to know. One thing that the program’s trying to do better is to have more of a dialogue with people who have had that experience and try to learn from your experience with evaluating quality for your specific
applications. How many of you have contributed data? Either point data or plot data or some sort
of structured dialogue or feed back on one of our websites? Good. Then okay. Background: starting in about the
late 90s through about 2007, the Government
Accountability Office — I don’t know what they the call themselves now — had these reports that came out that
were driving the federal agencies, and and now currently, the federal and the non-
federal agencies to manage fire with more of a more cohesive cross-boundary cross-organizational
structure perspective. And LANDFIRE — — they pointed out that data don’t
exist for that type of management and so LANDFIRE was conceived as
being a wall-to-wall consistent, comprehensive set of geospatial data layers that can help efforts to manage fire cohesively across organizations and across
the United States. So again, it’s conceived of as a seamless set of comprehensive, consistent geospatial data products that not
only were focused on fire management planning
or fire management tactical activities, suppression or
response, but also fire business –the business of
trying to figure out how to allocate resources ahead of time, how to allocate resources for hazardous fuel reduction, etcetera. Originally FPA & WFDSS were the two national programs that were in place and since then HFPAS, IFT-DSS and other national programs have evolved, and LANDFIRE sort of evolved along with
them, and trying to meet their requirements. That includes the current cohesive strategy
effort that’s going on. There’s a special session on that, I think, tomorrow. Timeline: Originally LANDFIRE was at the Missoula Fire Science Lab pretty much exclusively. In the late 90s the coarse-
scale data layers were put out by the Missoula lab, and Congress and the powers-that-be
wanted to have a finer resolution set of data, and a a more comprehensive set of data layers
to address fire management problems. And so the lab and the EROS data center in Sioux
Falls were asked to put together a proposal. We did a prototype for these two map zones in Utah and Montana and Idaho. That was
completed in 2003. And then the national implementation took place, and then since 2004 took us a
while develop the first set layers for all 50 states and since then, 2011 and 2012, we’ve put out new versions so-called “improved and updated
versions.” So I’ll just touch on what those all meant. We’re looking
at somewhere down the road incorporating a brand new set of imagery and remapping the entire United States. Originally that was conceived in the 2015-2017 time-line. So until then, the refresh activity will continue. But now there have been some discussions
about moving out forward and especially with the new LANDSAT
platform doing a remap sooner than 2015 or 17 but no decisions have been made. These are the geographic areas of LANDFIRE. The smaller units are the MRLC map zones. They were divided up for
different national mapping activity –the National
land cover dataset — which is where we get our LANDSAT data
from. Originally before the LANDSAT data were free, we needed to tie with them we just retained those sites and then regionally we have these six, seven, eight geographic regions. And that’s how we tear up the country and that’s we, progressively, at least for the first
round refreshes, have have updated the data from other geographic
areas, put out the data and then moved on. I wish Olie was here because he is supposed to be mapping all these important islands in the Pacific. So, in one-page, what is LANDFIRE. Again,
objectives were for national assessments for both, so
we deliver both vegetation and fuel layers. All the layers were supposed to be integrated so they made logical sense we drilled down through them, using consistents methods for
comprehensive wall-to-wall. And the methods needed to be repeatable so
that they could be maintained and kept up to currency. There are 24 primary data products describing vegetation fuel and fire regime. The subsequent presentations today will
talk a lot more about those. Again, the original concept was that
it was gonna be useful for national strategic applications, and since the first data came out
people have been exploring and, being inquisitive people, they go out and they use the data for lots of different applications both
for local fire applications, and for all kinds
non-fire applications like habitat assessments. Some of the different examples of that is that there is a bee pollination study. I saw recently a national pica habitat map that was based on LANDFIRE data. We work really closely with the GAP
program whose main goal is to map wildlife habitat. These are the main partners: The Nature Conservancy, the Forest Service, USGS, and our web page is www.landfire.gov. So, the impetus was, back in 2000 this is basically the state-of-art
fuels mapping. This is Nine-Mile Ranger District in Missoula, just north in Missoula. The best they could do for comprehensive
fuel maps was for their administrative units so this is just for the Forest Service land —
everything else is inholdings, or private or other lands — but this is
the best they could come up with as an organization to use to feed fire behavior models during the — this is for the 2000 fire season. LANDFIRE was at the very very early stages
of its prototype in this area and we could show that it was fairly easy to
make a map that was wall-to-wall wasn’t restricted
to any specific units, and was much more applicable to cohesive applications, at least then, and in just this local area. But now nationally. So that’s just to show that basically LANDFIRE filled in the holes. But there were many issues along with that. So over time not only could you use the same data
layers for specific local incidents — and neither of these maps show any LANDFIRE products, but they show analyses that were conducted using LANDFIRE data. This are WFDSS runs and from back in 2003 these are just FARSITE runs for that fire season. But what is important here is that
data for a specific incident, the specific local application, the same exact data can be used for regional applications such as this. It’s the first time fire managers have had access for that sort of equivalent comprehensive and consistent geospatial data. And that’s not only important in a tactical situation when you have fire burning, but it’s also obviously important when you’re trying to allocate
resources for hazardous fuel reduction. So this is the second flow chart, it’s much
simpler – basically production starts with a bunch of foundational geospatial data whether it’s satellite imagery or topography and a bunch of reference data. I think next presentation is all about a reference dataset. Map vegetation: three different types you’ll hear more about that later, and then we map a suite of products that are related fire regimes, and we map a suite of products and are related to fire behavior and fire effects. Everything here is fully dependent on
not just the processes but the quality of the data here, and that’s where users play an
important part because it turns into a feedback. So say
someone uses these and they’re not happy with the way they work, they have the capability of either adding an additional data to the cycle here for future remaps, or for providing some sort of anecdotal
or narrative feedback through one of our LANDFIRE feedback mechanisms so these processes could be changed either locally or systematically across the nation. So. 1.0. The methods from our our prototype
areas in western United States were adapted for the whole nation. It turned out to be much less clean-cut than we had hoped, so our efforts to
map all 50 states were basically constantly in development as we incorporated new
and perplexing mapping issues, for example, mapping in the eastern
United States where we didn’t really know that much about the relationship between potential
vegetation to the extent that it exists there and existing vegetation and how that
translates into fuel. Again those details will be discussed more
in future presentations sessions. Over time we learned that our vegetation
classification system that we had made up ourselves for our
prototype was not going to work for the United States
and so we had to pick a vegetation classification that was in place and consistent for all 50 states and we went with NatureServe’s
ecological systems. I think two talks from now is going to address the vegetation classification system in Don Long’s talk. Later, Wendel will talk about fire regime condition class, not only in this session, but he’s also talking about that in additional sessions, and then we’ll have Tobin talking
about our fire behavior and fire effects models. And, again, these two maps on the side are basically this one is a WFDSS run for very craggy fire in California, and this is an example of a non-fire application of LANDFIRE in Region 4 of the Forest Service, where they did some bighorn sheep habitat assessment with the LANDFIRE layers. So after the National layers were out, we were (aside: are we leaving 5 minutes for questions?) Okay. I’ll hurry through it. Well, we had to figure out how we were going to keep the the layers up to date. How were we going to
bring the layers from the date of the original imagery
that we had for the entire United States up to currency. A this system was developed where instead remapping everything you see
remotely sensed data and brand-new reference data, what we did
was identify areas on the landscape that had changed since the LANDFIRE mapping and
updated those layers — updated those specific areas — and you’ll see
examples of that. So for example where someone out in the United States had done some fuel
treatments, we used polygons and descriptions of
those fuel treatments to change the LANDFIRE fuel data for those areas, and you’ll see lots of examples of that through
today’s talks. So that when we talk about refresh, that’s what we’re talking about — updating local areas or specific types
across the United States that have grown since the last time
of the mapping because it was just unfeasible for
us every time we needed to update, to compile a new set of satellite
data and map from scratch. Hence, this refresh, and then eventually a remap of the United States, because eventually once when we’ve refreshed three or four times you’re sort of going to have a patchwork and it’s going be time to reset with some real consistent set of imagery from one date. So, I’m not going to go on about initially that first refresh was to fix, going back, to fix the problems that we knew about with the data related to places where the data hadn’t been mapped right up to our international borders. We didn’t do a very good job of mapping riparian wetlands. We needed to address the burnability of fuels that were integrated with agricultural and integrated with urban areas. The extent of our barriers or our barren areas were limited. And also we needed to
reevaluate ecological systems as the correct vegetation classification for this application. So, there’ve been two refreshes with one currently under way. There’s suites of data out there that are current as of 2001, current as of 2008, and now in 2013 there’ll going to be a new set of layers that are current as of 2010. There’s that two-year time lag there due to the enormity of the remapping task. Again, this just gets at how we identified areas that had changed, so that (Matt aside: the dates are off here) if the original map is here we could
look at satellite imagery subsequent to that, identify patches that had changed and then
when we remapped, we would just remap for the patches that had changed. Our goal, basically, was to not just
look at areas that had been treated mechanically, or some sort of vegetation management, we used MTBS data layers so we were
incorporating such for wildland fires, we used data that we had for wind
damage, insects and disease damage — these are all existing data layers that we were able to get from the Forest Service and other sources, and we also used LANDSAT-based vegetation change application to look for areas where that had changed that weren’t available to us
as polygon datasets from existing sources. And that turned out to be a very big task for
all 50 states — several thousand LANDSAT scenes. This is the timeline for the most recent refresh
effort. Data compilation and production has been ongoing. The geographic areas are being produced now — the release will
happen sometime next let’s say — let’s shoot this all back based on what I’m learning — and say summer. And then documentation reports, quality
assessments, and close out sometime in 2013. I’m gonna skip this. It’s basically a
summary of all the different refresh efforts. If you stay in the session
all day you’ll hear about each of these. Here are some sites where we accept feedback and the way that the
dialogue with the users — we have entire organization, NIFTT, that is dedicated to developing
tools and working with users on LANDFIRE data — the tech transfer
group. There’s the helpdesk and then this
down here at the bottom, the monthly users call, is something that if you’re not participating in it now, and would
like to participate in it, stop me and we’ll get your name on
the list. It seems to me like it’s a fairly big list
but if you consider the amount of users of LANDFIRE data that are out there, it’s a pretty short list, and so I think
that if you want to be on the list let me know for that monthly call. I am out of time, but Brenda will talk more about how we gather events data and plot data and
how that gets compiled and the timing of that and how that gets
expressed and updated LANDFIRE layers. Here are some take-home messages: you can email our helpdesk and we will get back to you.

Leave a Reply

Your email address will not be published. Required fields are marked *