The Map is not the Territory: Rethinking the Stand

Advances in remote sensing and information processing have given foresters an unprecedented ability to assess their forests. These powerful tools are beginning to challenge the traditional conception of the "stand" and are opening up a new generation of possibilities for silviculture.

The idea of the stand has a long and multifaceted legacy. Foresters have been using the stand concept to manage forests for hundreds of years, but each forester will give a slightly different definition for what a stand actually is. This is not a surprise - different forestry professionals use stands for very different purposes. Biometricians think about stands in the context of sampling theory. Loggers think about stands in terms of operational units. Ecologists think about stands in terms of species habitat. Silviculturists have to think about everything! The SAF Dictionary of Forestry says that a stand is:

An aggregation of trees or other growth occupying a specific area and sufficiently uniform in species composition, size, age, arrangement, and condition as to be distinguished from the forest or other growth on adjoining areas

Ultimately, we use the concept of a stand because it simplifies the complexity of the underlying forest. As Dr. James Scott, the director of Yale's Agrarian Studies Program notes in his magisterial Seeing Like a State, much of history is the story of mankind attempting to impose "legibility" upon complex natural and social systems. Without some level of abstraction, forests would simply be too complex to understand.

Have we oversimplified? Scott claims that no administrative system can represent our complex world "except through a heroic and greatly schematized process of abstraction and simplification." Or, as philosopher Alfred Korzybski famously quipped, "the map is not the territory." How divergent is our map from reality?

Oversimplification can have costly consequences. Scott spends an entire chapter of his book on the birth of scientific forestry in 16th century Saxony and how oversimplified mapping and management led to production losses of 20-30%. Since the 1500's, cartographers and biometricians have gotten much better at ensuring that the "map" stays in sync with the "territory," but we would do well to keep Scott's warning in mind:

At the limit, the forest itself would not even have to be seen; it could be "read" accurately from the tables and maps in the forester's office… This utopian dream of scientific forestry was, of course, only the immanent logic of its techniques. It was not and could not ever be realized in practice.

In modern forestry, the "map" is often located within a digital Geographic Information System (GIS). Forest managers spend a lot of time and money to implement, maintain, and update these systems. Ultimately, the information contained within them is much the same as the "tables and maps" of our Saxon forebears. While digital technology enables us to accumulate vast databases of tables and maps, modern GIS systems suffer from the same weaknesses as their ancient paper predecessors - they are constantly falling out of sync with the reality on the ground. Major landowners spend millions of dollars every year re-typing stands, cruising timber, and analyzing aerial imagery to try to keep up with the never-ending changes in their forests.

Updating inventory information for each stand is a sisyphean task in itself, but it gets even harder when stand bounds themselves change. One of the classic challenges in maintaining a GIS database is the problem of splitting or merging stand bounds to account for partial harvests, growth differences, etc. What portion of the inventory should be allocated to each section? Where should the lines be drawn? What happens to the record of the stand's management history? All of these factors are important for good silvicultural decision-making, but our intentions often run up against the limitations that our stand-based tools have created for us.

One of the most profound limitations is that we are forced to make single delineation of our forest. As society increasingly demands that we manage for values beyond fiber - especially for ecosystem services like carbon, water, and species habitat - the stand paradigm is beginning to show its limits. Why should the stand bounds used for harvesting be the same as those used for carbon assessment? Why should we use the same bounds for measuring timber volume as we do for mapping a multitude of species habitats? Why should these bounds remain mostly fixed over time?

Cost is certainly a factor, but so is the lack of a clear replacement for the stand concept. At SilviaTerra, we're excited about the potential to move from a vector-based stand paradigm to a raster-based "pixel" paradigm. Recent advances in remote sensing and biometric analysis have made it possible to get a reasonable estimate of the trees in each 1/20 acre square "pixel" of a forest.

pixel-paradigm
Creation of three different property delineations for biomass accounting based on an underlying pixel-level inventory

There are many advantages of maintaining inventory information at the pixel level. Updates to inventory can be driven by frequently updated satellite imagery which is also collected at the pixel level. It is possible to maintain and change many different "stand boundaries" - just "cookie cut" the underlying inventory raster layer by the different boundary vectors. Management history can be tracked at the pixel level too. With this switch, many of the classic GIS problems (splitting and merging, updating, etc.) go away and we gain powerful new ways of visualizing and thinking about our forests.

Our map is getting closer to the territory. Storing inventory information at the pixel level is still an abstraction, but it avoids many of the oversimplifications of the stand paradigm. There are still many challenges to overcome and questions to answer - chief among them is how we should rethink silvicultural planning in this new paradigm. We're excited to get this conversation going and are looking forward to discussing further at the National SAF convention in November.