Monday AM Overview

Neighborhood Dynamics of forest ecosystems: an overview of the theory and the model.

Presented by Charlie Canham.

How did we get here?

There were two precursors to neighborhood dynamics. One was gap phase dynamics, which assumed that one could ignore what was going on in the canopy and focus on the gaps. The other was landscape-scale patches for the role of fire in xeric forests. At both scales, the first big step was the creation of a set of equations which treated the landscape as a patchwork of gaps. The focus is on the population dynamics of the discrete patches created by disturbance.

There are several limitations of discrete patch models. There is no way capture heterogeneity in patches, spatial interactions between patches, the interactions between patches and surrounding matrix, and heterogeneity in disturbance severity. Furthermore, nice clean gaps almost never occur.

Thus the need for a spatially-explicit, neighborhood theory of forest ecosystem dynamics. This theory addresses the strong horizontal integration of the ecosystem (tree-tree interactions). Neighborhood phenomena include spatial variation in understory light levels and seed rain and seedling establishment.

Neighborhood phenomenon: Soil resource availability and nutrient cycling.

Charlie is getting good data on calcium cycling - but the picture is complicated. There are multiple nutrient limitations.

The good news: If we have the parent material calcium concentration, we can predict well mid-summer calcium availability. You can predict species composition from parent material well, as long as you're working within a specific species suite with limited range of parent material.

Neighborhood phenomenon: Abundance and activity of small mammals

Spatial variation in occurrence of small mammals is strongly influenced by the spatial distribution of large-seeded tree species that represent important food resources. Spatial variation in rates of seed and seedling predation varies accordingly.

Implications for ecosystem science

There are several implications for ecosystem science. Ecosystem science has barely even embraced gap-phase dynamics, let alone recognized the limitations of that approach. There is widespread recognition that species matter for ecosystem processes, but very little in the way of an explicit framework for incorporating population dynamics in ecosystem models. Spatially-explicit models are especially important when species effects are non-additive, and there is spatial heterogeneity in the environment.

SORTIE

SORTIE is a neighborhood model of forest dyamics. It is individual-tree based and spatially-explicit. The spatial scale of the effective "neighborhood" varies for any given property or process, as needed. Canopy gaps are recognized as heterogeneous entities that emerge as a result of the process of tree mortality. Canopy gaps are perceived differently by different tree species because of differences in light requirements and shade tolerance.

Where are we going?

In various stages of incorporation:

• Substrate dynamics
• Disturbance
• Bath recruitment
• Seed dispersal
• Seed predation
• Seedling dynamics
• Nutrient cycling
• Effects of soil nutrients
• Effects of herbivores
• Adult tree competition
• Harvesting

Basic Goals for Re-engineering

Provide a general platform for the development of neighborhood theory.

• Community dynamics and maintenance of diversity
• Linkages between population dynamics and ecosystem processes
• Incorporation of new processes: pests, pathogens, invasive species, facilitation...
• The role of stochasticity

Provide a flexible tool for the application of neighborhood theory - management.

Discussion

We must be honest about how confident we are in the statements we make about forest management - especially when the data's noisy. Of course, most decisions are currently made with no data or seat of the pants - so the model will be an improvement no matter what.

In forestry, you work with incomplete data - you can't map all the trees. So you must be able to generalize. However, remote sensing software can map locations of trees and crown diameters - plus rough species mix information. So in the future it might be a valuable tool.

Linking herbivores to neighborhood dynamics - difficult.

• Effects on deer are too large-scale - you'd have to model lawns and cornfields.
• The composition of the forest affects deer - lack of shelter, food causes the population to crash.
• The most important aspect of deer is their long-term average impact.
  • One possible way to look at this is to ask: What's the steady-state effect of deer at different densities? What are population-density targets?
  • Problem - we're lacking a body of knowledge as baseline.
• Good first approximation to begin studying deer - set them up as an external phenomenon like weather.