Usefulness of Spatial Landscape Models

Turner et al.’s discussion about the usefulness of spatial models in land management is now a bit of a classic (written in 1995) but it had also been a while since I read it. Re-reading it after coming back from a trip to our study area, many of the paper’s points resonated with what people (many of them natural resource managers) I met with were saying.

Turner et al. suggest that (p.13) “Models that integrate ecological and economic components so that the models can be used to explore both sets of consequences simultaneously are even more valuable [than ecological alone]”. This is the driving rationale for our research project. As it was succinctly put by one potential landowner in the study area, models of this kind will contribute to the development of plans that are based on an ecological approach but backed up with economic justification.

Given the hierarchical nature of landscape ecological processes and the importance of human activity on those processes, Turner et al. highlight (p.15) that “Land ownership has a large impact on management decisions, and a useful contribution of spatially explicit models is the ability to explore the effects of management by various owners within a mosaic of public and private lands.” With a range land owners, including the state and private industrial companies, the UP study area is in this position and the model we are developing will be able to directly consider the impacts of different land owner management strategies for the landscape as a wider region. Thus, one of the driving questions of the research is “how should timber be harvested across space and time in multiple land ownerships to ensure a sustainable landscape?”

One of the most striking things I was told on my trip was that the most useful thing our model would be able to do for land managers would be if it could get people to sit down together to come up with a coherent, sustainable management plan. Again, the links with Turner et al. are clear (p.15); “Communication between land managers and ecologists remains an important challenge, and spatially explicit models have the potential to create a common working framework.”

However, not only is the communication and collaboration side of the research a challenge, but so too is the technical side of things. Turner et al. highlight the issue of data quality; the model will only be as good as the data used and the accurate up-to-date spatial data bases required are expensive to produce. Furthermore, the quality of the data will determine the modeller’s ability to parameterizes the model at a given spatial resolution and extent. I’m currently reviewing the data that has been collected over the past few years by the research group at CSIS regarding the interactions between deer density, tree regeneration and bid habitat, but also the data managed and made available by Michigan’s Department of Natural Resources. Producing an accurate representation of deer population dynamics and movement across the landscape is certainly going to be a challenge. Next, the relationships between deer browse pressure and vegetation regeneration need to be specified and parameterized. The estimates of deer population and location can then be combined with these relationships to dynamically represent the interactions across space.

Once the model is up and running we will be able to examine spatial scenarios of forest management to assess both ecological and economic sustainability. For example, with regard to the appropriate location of mesic confer regeneration “…increasing the [mesic confer] component is expected to increase the number of individuals of conifer-associated bird species. And over time reduce productivity of the summer deer range and expand areas potentially suitable for deer during winter, resulting in a smaller deer herd dispersed over a larger wintering area (Doepker et al, 2001) in turn resulting in less browsing pressure in WUP forests. The eventual size, configuration, contiguousness and/or juxtaposition of restored habitats to existing or historical mesic conifer habitats and winter deer-yards on non-MDNR lands (public and private) may affect the success of these outcomes” (DNR 2004). Right now this confer regeneration is not going well and areas of maple forest are increasing.

Economically, the model should be able to show how different harvest rotations and management plans by private industrial land owners can ensure the most productive use of their land whilst ensuring both ecological and economic sustainability of the landscape. And not only for single landowners. The model should be useful to examine how actions of neighbouring land under differing ownership can work in concert. For example, if the private industrial goal is intensive harvest, maybe the primary objective of the state should be to ensure conifer cover. But the question then is what are the spatial implications of this? Is there any point in confer regeneration (which provides thermal cover for deer in the winter) if the distance between state and corporate land is large and deer cannot move from thermal cover to find food?

These are the sorts of questions and challenges to which spatial landscape models can be applied, and which we are aiming to tackle. Right now though, it’s time to concentrate on the technical development of the model and the representation of the spatio-temporal deer-vegetation interactions.

Turner, M.G., Arthaud, G.J., Engstrom, R.T, Hejl, S.J., Liu, J., Loeb, S. & McKelvey, K. (1995) Usefulness of Spatially Explicit Population Models in Land Management Ecological Applications, 5:1 12-16.

Leave a Reply

Fill in your details below or click an icon to log in: Logo

You are commenting using your account. Log Out /  Change )

Google photo

You are commenting using your Google account. Log Out /  Change )

Twitter picture

You are commenting using your Twitter account. Log Out /  Change )

Facebook photo

You are commenting using your Facebook account. Log Out /  Change )

Connecting to %s

This site uses Akismet to reduce spam. Learn how your comment data is processed.