Imagining A Changing Forest

 

A desired condition is not a picture.  It’s a movie.

This is a map of four seral stages for the Pagosa Springs district of the San Juan National Forest.  Young stands of trees (class 1) are very rare.  So are the purple areas representing the oldest stands of trees (class 4).  Most of the map shows middle-aged stands (red and green).  Think about how this information might be used in forest planning.  For instance, the purple areas might be important habitat for late-seral stage wildlife species, they might be mapped as ecological reserves, or they might have some unique social values we want to protect.

Here is a simulation of what could happen to these stands of trees over time due to fire, insects and disease.  Each interval in the movie is a 10-year increment.   It is based on work by Kevin McGarigal of the University of Massachusetts and Bill Romme now at CSU, for the San Juan Forest Plan Revision using a GIS-based simulator called RMLANDS.  It formed an understanding of the historical range of variability of vegetation for the DEIS.

The stand size and distribution is most dependent upon fire interval and fire size, randomly simulated based on historical data.  Over time, the tree conditions seem to float across the landscape like shifting sand.  There are some places where topography seems to influence the disturbances to allow persistence of older trees, but even these areas are eventually affected by the random events.

The smaller the scale, the larger the variation.  If you look at a particular place, there is more change over time in the color of the place.  The larger the scale, there is more likelihood that you’ll find the color you are looking for somewhere.

When planning for forests influenced by disturbance, landscape ecologists advise us that it’s important to think of time and space.   It calls for a discussion beyond static desired conditions.  Instead, a discussion is needed on the disturbance processes, if anything should be or can be done to shape those processes, and what we should do with the conditions that might result.  This is a very different type of forest plan than we have done in the past.

4 thoughts on “Imagining A Changing Forest”

  1. I’m not sure that simulating stochastic processes like fire, insects and disease through computer models and visualizing them is all that helpful. Because the visualization and models makes the assumptions seem more real that they are if you would just say them out loud.

    In addition to the extra funding required to do that, I think is not as good public involvement as just talking about “across the landscape we need a variety of seral stages, right now trees are growing and we are going to try to get more acres of certain stages through letting the trees grow or doing something else, but if big fires and bug outbreaks strike, we may not have enough of certain seral stages to provide for our current desirable array of species, so we probably need to have some extra.”

    How about “we will need to decide how to handle the fact that there will be fires and blowdowns and insect and disease outbreaks; we need to think, in planning that they might occur, and when they occur, we need to decide what we are going to do.” In reality, for many things about large fires and bug outbreaks we know what we generally do; follow fire policy for fires, and based on bug biology either try to protect trees or clean up afterwards. It’s really not all that complicated.

    Which is kind of what you said; but it obviates the need for technical jargon like “disturbance processes” and concepts that are even confusing and questionable to us technical specialists like “historic range of variation” . And probably saves money on modeling and analysis. Again, my view (like Andy) is that simple is usually best.

    Reply
  2. Looks like we’ve got a topic for a future post and discussion thread: the use of models in forest planning. For now, it’s worth noting that models should be no more or no less complex than the questions being asked. Models can be useful to inform a discussion, to aid in understanding of the implications of our value choices, or to clarify what’s really important. Models can be an important tool for scenario planning.

    For a lot of people, I don’t think what Sharon says is obvious to them. They may understand that there are fires and insect and disease outbreaks, but they might not grasp what the resulting pattern of trees might be across the forest over time.

    Plus, when you are dealing with random events (if they are truly random) then what you actually see on the ground is merely one possible outcome of what could have happened. It’s not until you simulate thousands of alternative outcomes that you truly grasp the complexity of managing under uncertainty. These results challenge people’s assumptions.

    Think of a visit to the forest. If you visit a forest only once in the spring, you will have collected “real” data. But if you go back to the spot in the fall, you will see an entirely different condition. So, now to really see the “real” forest, you have to visit at least twice. But what about winter? What about a winter in an el nino year? What about a winter after five years of drought? What about a winter after a pine beetle epidemic after five years of drought? Quite simply, models help us understand a situation without having to collect so much data, especially when the data we’re collecting is merely one possible “draw” in the deck of cards we’ve been given. It’s only after you’ve collected data over a very large period of time – i.e. an “historic” period, can you start to understand if these random events tend to repeat themselves, or stay within a range. This isn’t that technical, but I think it’s confusing to people because our minds sometimes like to seek order in the chaos, and it’s not that simple. Of course, what’s really blowing people’s minds right now is that there might not be historical patterns – there may be places with no equilibrium of forest conditions over time – (another future post) – or the whole game board is shifting. Or the old patterns don’t matter anymore (another knot in the planning problem wickedness). But it’s important to know the old patterns so we know if changes are WITHIN the patterns, or changes TO the patterns.

    But things sometimes aren’t random. I looked at the simulation again. There is a purple spot in the middle of the map that stays purple over a time period equivalent to a person’s life. So throughout all the chaos of the simulation, there is at least one spot that’s a sanctuary from the chaos. It seems there are some influences from topography or the pattern of the vegetation. So, it’s not entirely random, and the general statements that Sharon suggests about a forest influenced by fire don’t hold true in this case. There are dangers of oversimplifying how forests work. That’s why non-spatial vegetation disturbance models (like VDDT that was used for Landfire), that don’t take into account an actual map of topography and the arrangement of stand types, should only be used to answer the most general policy questions. Isn’t this fun?

    Reply
  3. John, you said

    “But it’s important to know the old patterns so we know if changes are WITHIN the patterns, or changes TO the patterns.” I really don’t understand why that IS important.

    I am not oversimplifying how forests work – au contraire, I am clearly with the camp that says “ecosystems are more complex than we think, they are more complex than we can think,” or in its original formulation by J.B.S. Haldane (a population geneticist) “the universe is queerer that we think, it’s queerer than we can think).

    So we don’t know what used to happen, or what is going to happen, other than trees have a lifespan. We don’t know how much or where temperatures will change, nor what species will have advantages due to differences in climate, invasive species and changes in competition and predation. We might like the changes or we might not. We might be able to afford to do something about them or we might not.

    If the simulation shows a purple spot, either you can see why if you go out on the ground, and it makes sense, or it doesn’t and it is a function of the assumptions that are in the model.

    I am leery of models for planning, not for scientific exploration, for three reasons:
    1) they are not always clear to outsiders about their assumptions, frequently biological models do not conduct sensitivity analysis of their assumptions,
    2) the worldview behind the model is often given and not discussed with the public (need I say Forplan?)
    2) that leaves the public from being able to debate and understand the assumptions,
    3) scientists start using a shorthand about model projections as if they had some reality that they don’t (climate science has many difficulties with this).
    4) they cost money, that could be better spent, in my view, collaborating or buying conservation easements for wildlife corridors.

    In post-normal science, trust of scientists and public debate are necessary. I think using models that privilege one group of scientists views, possibly at the expense of other disciplines and the public, have to be scrutinized carefully.

    Reply

Leave a Comment