The projects below have been conducted by Don Morgan over the last five years

Characterising Natural Landscapes

Table of Contents

• Methods
• Results
• Discussion

Methods

Because this project aims to develop methodology (for analysing landscapes), we followed no specific methods, but rather use a set of general approaches for each objective:

• Data management
• Characterising natural landscapes
• Predicting long term consequences
• Linking scales
• Including stakeholders

We developed and applied methods in a series of pilot projects. We employ standard GIS technology, computer code and landscape modelling languages, using non-proprietary software when possible and developing modules rather than large integrated applications. Several reports describe the methods developed and the results of testing in case studies; these reports are referred to in the results section. Here we provide and overview of the methods used in the various projects.

Our original systems design philosophy was to use non-proprietary software to facilitate the free distribution of all computer models. Initially we relied on proprietary information systems software such as the PAMAP and the ARC/INFO GIS. Currently, the BC MoF does its data management using the ArcINFO and Oracle and we are dependent upon it for basic data management. However, most of the project application development has been based on free systems.

Inventories often lack some of the basic ecological units desired for analysis of habitat and coarse filter biodiversity. To address this problem, we investigated several methods for characterizing landscapes, from simple GIS analysis of current inventories to sophisticated landscape analysis and terrain modelling. We used GIS scripts to extract information on wetland complexes, age class patterns, patch size categories, species groupings and terrain categories using elevation, slope and forest cover data.

We developed methods to predict fine scale ecological variation, creating units similar to groups of BEC site series. BEC site series are defined as a combination of soil moisture and nutrient regimes (Meidinger and Pojar 1991, Banner et al. 1993). Conventional terrestrial ecosystem mapping (used to map site series) is extremely expensive (approx. $5.00/ha), hence it is not available for most areas of BC. By working with ecological experts and applying terrain modelling techniques, we generated the soil moisture component of site series. When soil moisture is combined with floodplain and alluvial fan modelling and forest cover information, a powerful tool emerges, enabled us to delineate ecological units with in landscapes. We applied the floodplain model in the Iskut-Stikine case study to provide information for a succession model. We tested soil moisture, wetland and alluvial fan models in the Lakes, Kispiox and North Coast Forest Districts as a basis for determining Grizzly Bear Habitat Units (Eadie 2001, MacLennan and Senichenko 2001 and MacLennan et al. 2001).

Disturbance events, often fire, but also wind throw, landslide, and insect outbreaks, shape the pattern of seral stages in the landscape. Disturbed areas may revert to early seral stages, and recover forest structure gradually, over several hundred years. With ongoing occurrence of disturbance events, the landscape becomes a mosaic of seral stages with different forest structures. For an excellent, in-depth discussion of landscape pattern refer to Eng, 1998.

Characterising the natural disturbance regime in an area is particularly important to provide a basis for ecosystem management. Problematically, historic age class distributions cannot be recreated by simply making all younger forest old, because disturbance hits younger stands as well as old stands (i.e., some young stand were young stands historically). Using historical inventory data we reconstructed a snapshot of the Invermere landscape as it would appear if there had not been any logging, producing an “adjusted age class map”. The age of salvaged sites was set based on the date of the wild fire and cover attributes were taken from the largest burned neighbour. Similarly, stands that had been logged within the mature forest matrix where assigned the age class and cover of their largest mature neighbour.

This adjusted age class map was used to evaluate landscape pattern and composition. We did a number of comparisons of the age class map derived from the current inventory with the “adjusted age class” map to get a better understanding of the differences in the natural landscape patterns and the ones resettling from harvesting activities. This information was used to help parameterise the landscape event simulator (SELES),used for temporal landscape modelling. We also compared adjusted maps to projected maps to determine whether our present management regimes approximate natural patterns..

We used the “Roll-Back” technique as presented by DeLong and Tanner (1996) to compare the adjusted and un-adjusted age class distributions in the Invermere case study. Briefly, if we assume that fire burns the landscape indiscriminately, then stands of all ages are then equally susceptible to fire. Therefore to determine historical disturbance size we can allocate younger forests to older forests in proportion to their current percent occurrence in the landscape.

We implemented two measures of connectivity for use in landscape analysis. Such measures employ inter-patch distances, sometimes in conjunction with patch sizes. One example, the “Minimum Spanning Tree” (MST) (A. Fall, pers. com.), is the sum of the shortest edge-edge inter-patch distances necessary to connect all patches. In a sense, MST is the minimum distance you would have to travel to visit all patches in the landscape. Another index is the “Centroid Connectivity” (CCE) index described by He et al (2000). This index calculates the average of all the patch-patch interactions by habitat class in the landscape. The strength of an interaction between patches is the product of the mass (area) of the 2 patches, divided by the square of the distance between patch centroids (mean x and y co-ordinates of the patch). These are attractive indices as there is evidence that below threshold habitat abundance levels, the spacing and size of suitable patches potentially affects plant and animal movements or distribution (Keitt et al. 1997; Collingham and Huntley 2000). We tested this work in the Morice Forest District

Results

In this section we describe the reports, software products and database scripts generated by this project. We do not list all the workshops held. At one level, workshops are part of the research process, however, they also serve to pass methodology to an interested audience.

Reports

Morgan, D. 1998. Invermere Enhanced Forest Management Pilot Project: Natural Disturbance Analysis. Unpublished Report. Research Branch. BC Ministry of Forests, Smithers, BC /invpage/index.html – start at the map

Olivotto G. 1998. Landscape Lego Software Design – Phase 1. Unpublished Report. Research Branch. BC Ministry of Forests, Smithers, BC –

Morgan D. 1999. Invermere Enhanced Forest Management Pilot Project: "Standing up the Trees" Historical Inventory Reconstruction. Unpublished Report. Research Branch. BC Ministry of Forests, Smithers, BC – (html) (pdf)

M. Jennings and D.G. Morgan. 2000. Landscape Lego Documentation. Unpublished Report. Research Branch. BC Ministry of Forests, Smithers, BC – (html) (pdf)

Fall, A. and D.G. Morgan. 2000. Ecological Soil Moisture Prediction Models. Unpublished Report. Research Branch. BC Ministry of Forests, Smithers, BC - (html) (pdf)

Steventon, D. 2001. Effects of Landscape Pattern on Species Conservation: Working Paper. – in review – (html) (pdf)

Morgan, D.G., Olivotto, G. and D. Daust. 2000. Landscape Lego - playing with landscape pattern. Extension Note – in review. – (html) (pdf)

Software products

Fall, A. and D.G. Morgan. 2000. Ecological Soil Moisture Prediction Model. Research Branch. BC Ministry of Forests, Smithers, BC (Download)

Fall, A. and D.G. Morgan. 2001. Floodplain and Fan Terrain Models. Research Branch. BC Ministry of Forests, Smithers, BC - (Download)

Jennings, M. and D.G. Morgan. 2001. Landscape Lego. Research Branch. BC Ministry of Forests, Smithers, BC – (Download)

Discussion

In the past, modelling has failed to reach its potential as a decision-support tool in the forest management arena. Since the emergence of computer science, simulation modelling has defined a solid niche in many scientific fields, judging by the numerous published models. The application of models in forest management seems less successful—the calculation of timber-supply being the only consistent application of modelling. Yet the complexity and large spatial and temporal scales that characterise forest management problems suggest modelling should help. Land-use decisions can profoundly impact our economy and our ecology, with consequences that may reach forward several generations. Wise decisions seem essential. While the analysis methodologies and software discussed in this report aim to improve the quality of decisions, passable software and modelling expertise have existed for many years. We believe decision-support projects have failed primarily for two reasons. First, modellers and decision-makers do not understand each other’s objectives and limitations. Decision-makers must often decide quickly, using available information and subjective judgement, and they may have unreasonable expectations about what can be modelled. Modellers, on the other hand, often aim for efficient elegant models, backed by thorough research, and are not unduly constrained by time. Second, influential stakeholders have not been adequately included in modelling projects (McLain and Lee 1996, Selin et al. 1997). Thus, to support our analytical methods developed, we have developed a framework for involving stakeholders—the framework simply aims to align the objectives of the modellers and the decision-makers (including stakeholders). If this main aim is remembered, then we believe modelling can inform forestry decisions.

Our original objective in characterising natural landscape patterns involved the application of automated mapping tools to identify features such as forest age class distribution, ecosystem pattern, riparian habitat and forest matrix and gaps. The case studies in the Lakes, Invermere, Morice, North Coast and Kispiox went well beyond this objective. We developed landscape reconstruction techniques in Invermere to compare historic forest management practises with a pristine landscape, which continues to be used for comparison with different operational strategies. In the Morice we have analysed the historic range of natural disturbance across the entire TSA and evaluated each landscape unit and how its industrial forest development fits in to the historical range. The research in the Morice is now being included in the Lakes/Morice IFPA’s biodiversity scenario project. We have made major contributions to ecosystem mapping with our terrain modelling techniques. These technique are being applied in a number of new projects, such as Lakes PEM and Morice Forest District Alpine Ecosystem Mapping. We are validating our automated ecosystem mapping with terrain ecosystem maps, aerial photos and ground sampling.

Acknowledgements

DULP is a BC Ministry of Forest Research Branch project with funding assistance from Forest Renewal BC.