Andrew Fall

A domain-specific language for models of landscape dynamics

Gaining insight into the dynamic nature of landscapes often involves the use of simulation models to explore potential changes over long time frames and extensive spatial areas. However, bridging the gap between conceptual models of landscape dynamics and their simulation on a computer can lead to many pitfalls. If implemented using a general-purpose programming language, the underlying model becomes hidden in the details of the computer code, making it difficult to compare the conceptual and implemented models, and to modify the model. Alternatively, previously built models may contain hidden assumptions and have limited adaptability. Domain-specific languages have been developed in a number of areas to facilitate the construction of models at a level closer to the conceptual model, thereby making model implementation more accessible to domain experts. Such tools to support modelling in the domain of landscape ecology can achieve a balance between the flexibility of programming and the structure and ease of using pre-built models. One of the goals of SELES (Spatially Explicit Landscape Event Simulator) has been to create a language for modelling landscape dynamics that provides ecologists and planners with an appropriate tool to address some of the problems that arise in model development. Our high-level, structured language separates the specification of model behaviour from the mechanics of its implementation, freeing landscape modellers from programming and allowing them to focus on the underlying model. This language is declarative and thus permits a clear representation of the conceptual model, which the SELES engine converts into a computer simulation of landscape change. Our structured framework guides the development of a broad class of spatio-temporal landscape models by aiding model prototyping, modification, verification, comparison, and re-use.

Spatial Graphs: Principles and Applications for Habitat Connectivity

A BSTRACTWell-founded methods to assess habitat connectivity are essential to inform land management decisions that include conservation and restoration goals. Indeed, to be able to develop a conservation plan that maintains animal movement through a fragmented landscape, spatial locations of habitat and paths among them need to be represented. Graph-based approaches have been proposed to determine paths among habitats at various scales and dispersal movement distances, and balance data requirements with information content. Conventional graphs, however, do not explicitly maintain geographic reference, reducing communication capacity and utility of other geo-spatial information. We present spatial graphs as a unifying theory for applying graph-based methods in a geographic context. Spatial graphs integrate a geometric reference system that ties patches and paths to specific spatial locations and spatial dimensions. Arguably, the complete graph, with paths between every pair of patches, may be one of the most relevant graphs from an ecosystem perspective, but it poses challenges to compute, process and visualize. We developed Minimum Planar Graphs as a spatial generalization of Delaunay triangulations to provide a reasonable approximation of complete graphs that facilitates visualization and comprehension of the network of connections across landscapes. If, as some authors have suggested, the minimum spanning tree identifies the connectivity “backbone” of a landscape, then the Minimum Planar Graph identifies the connectivity “network”. We applied spatial graphs, and in particular the Minimum Planar Graph, to analyze woodland caribou habitat in Manitoba, Canada to support the establishment of a national park.

The sensitivity of least-cost habitat graphs to relative cost surface values

Maintaining and restoring connectivity among high-quality habitat patches is recognized as an important goal for the conservation of animal populations. To provide an efficient measure of potential connectivity pathways in heterogeneous landscapes, least-cost route analysis has been combined with graph-theoretical techniques. In this study we use spatially explicit least-cost habitat graphs to examine how matrix quality and spatial configuration influence assessments of habitat connectivity. We generated artificial landscapes comprised of three landcover types ranked consistently from low to high quality: inhospitable matrix, hospitable matrix, and habitat. We controlled the area and degree of fragmentation of each landcover in a factorial experiment for a total of 20 combinations replicated 100 times. In each landscape we compared eight sets of relative landcover qualities (cost values of 1 for habitat, between 1.5 and 150 for hospitable matrix, and 3–10,000 for inhospitable matrix). We found that the spatial location of least-cost routes was sensitive to differences in relative cost values assigned to landcover types and that the degree of sensitivity depended on the spatial structure of the landscape. Highest sensitivity was found in landscapes with fragmented habitat and between 20 and 50% hospitable matrix; sensitivity decreased as habitat fragmentation decreased and the amount of hospitable matrix increased. As a means of coping with this sensitivity, we propose identifying multiple low-cost routes between pairs of habitat patches that collectively delineate probable movement zones. These probable movement zones account for uncertainty in least-cost routes and may be more robust to variation in landcover cost values.

How Fire History, Fire Suppression Practices and Climate Change Affect Wildfire Regimes in Mediterranean Landscapes

Available data show that future changes in global change drivers may lead to an increasing impact of fires on terrestrial ecosystems worldwide. Yet, fire regime changes in highly humanised fire-prone regions are difficult to predict because fire effects may be heavily mediated by human activities We investigated the role of fire suppression strategies in synergy with climate change on the resulting fire regimes in Catalonia (north-eastern Spain). We used a spatially-explicit fire-succession model at the landscape level to test whether the use of different firefighting opportunities related to observed reductions in fire spread rates and effective fire sizes, and hence changes in the fire regime. We calibrated this model with data from a period with weak firefighting and later assess the potential for suppression strategies to modify fire regimes expected under different levels of climate change. When comparing simulations with observed fire statistics from an eleven-year period with firefighting strategies in place, our results showed that, at least in two of the three sub-regions analysed, the observed fire regime could not be reproduced unless taking into account the effects of fire suppression. Fire regime descriptors were highly dependent on climate change scenarios, with a general trend, under baseline scenarios without fire suppression, to large-scale increases in area burnt. Fire suppression strategies had a strong capacity to compensate for climate change effects. However, strong active fire suppression was necessary to accomplish such compensation, while more opportunistic fire suppression strategies derived from recent fire history only had a variable, but generally weak, potential for compensation of enhanced fire impacts under climate change. The concept of fire regime in the Mediterranean is probably better interpreted as a highly dynamic process in which the main determinants of fire are rapidly modified by changes in landscape, climate and socioeconomic factors such as fire suppression strategies.

A Framework and Software Tool to Support Collaborative Landscape Analysis: Fitting Square Pegs into Square Holes

For landscape models to be applied successfully in management situations, models must address appropriate questions, include relevant processes and interactions, be perceived as credible and involve people affected by decisions. We propose a framework for collaborative model building that can address these issues, and has its roots in adaptive management, computer-supported collaborative work and landscape ecology. Models built through this framework integrate a variety of information sources, address relevant questions, and are customized for the particular landscape and policy environment under study. Participants are involved in the process from the start, and because their input is incorporated, they feel ownership of the resulting models, increasing the chance of model acceptance and application. There are two requirements for success: a tool that supports rapid model prototyping and modification, that makes a clear link between a conceptual and implemented model, and that has the ability to implement a wide range of model types; and a core team with skills in communication, research and analysis, and knowledge of ecology and forestry in addition to modelling. SELES (Spatially Explicit Landscape Event Simulator) is a tool for building and running models of landscape dynamics. It combines discrete event simulation with a spatial database and a relatively simple modelling language to allow rapid development of landscape simulations, and provides a high-level means of specifying complex model behaviours ranging from management actions to natural disturbance and succession. We have applied our framework in several forest modelling projects in British Columbia, Canada. We have found that this framework increases the interest by local experts and decision-makers to participate actively in the model building process. The workshop process and resulting models have efficiently provided insight into the dynamics of large landscapes over long time frames. The use of SELES has facilitated this process by providing a flexible, transparent environment in which models can be rapidly implemented and refined. As a result, model findings may be more readily incorporated into decision-support systems designed to assist resource managers in making informed decisions.

Concentrating anthropogenic disturbance to balance ecological and economic values: applications to forest management

To maintain healthy ecosystems, natural-disturbance-based management aims to minimize differences between unmanaged and managed landscapes. Two related approaches may help accomplish this goal, either applied together or in isolation: (1) concentrating anthropogenic disturbance through zoning (with protected areas and intensive management); and (2) emulating natural disturbances. The purpose of this paper is to examine the effects of these two approaches, applied both in isolation and in combination, on the structure of the forest landscape. To do so, we use a spatially explicit landscape simulation model on a large fire-dominated landscape in eastern Canada. Specifically, we examine the effects of (1) increasing the maximum size of logged stands (cutblocks) to better emulate the full range of fire sizes in a fire-dominated landscape, (2) increasing protected areas, and (3) adding aggregated or dispersed intensive wood production areas to the landscape in addition to protected areas (triad management). We focus on maximizing the amount and minimizing the fragmentation of old-growth forest and on reducing road construction. Increasing maximum cutblock size and adding protected areas led to reduced road construction, while the latter also resulted in less fragmentation and more old growth. Although protected areas led to reduced harvest volume, the addition of an intensive production zone (triad management) counterbalanced this loss and resulted in more old growth than equivalent scenarios with protected areas but no intensive production zone. However, we found no differences between aggregated and dispersed intensive wood production. Our results imply that differences between unmanaged and managed landscapes can be reduced by concentrating logging efforts through a combination of protected areas and intensive wood production, and by creating some larger cutblocks. We conclude that the forest industry and regulators should therefore seek to increase protected areas through triad management and consider increasing maximum cutblock size. These results add to a growing body of literature indicating that intensive management on a small part of the landscape may be better than less intensive management spread out over a much larger part of the landscape, whether this is in the context of forestry, agriculture, or urban development.

Backtrackable State with Linear Affine Implication and Assumption Grammars

A general framework of handling state information for logic programming languages on top of backtrackable assumptions (linear affine and intuitionistic implications ranging over the current continuation) is introduced. Assumption Grammars (AGs), a variant of Extended DCGs handling multiple streams without the need of a preprocessing technique, are specified within our framework. Equivalence with DCGs is shown through a translation from AGs to DCGs and through use of an implementation-independent meta-interpreter, customized for handling both DCGs and AGs.

The Foundations of Taxonomic Encoding

Taxonomies (partially ordered sets and lattices) are important in many areas of computing science, particularly object‐oriented languages, machine learning, and knowledge representation. Taxonomic encoding strives to enhance the efficiency of taxonomic representation and use, which becomes increasingly important as the size of taxonomies grows. In this paper, we describe a formal structure, called a spanning set, in which taxonomic encoding techniques can be characterized. Any taxonomic encoding scheme implements a mapping from the original ordered set into a structure, such as the lattice of bit‐vectors or logical terms, in which operations can be performed efficiently. We analyze the fundamental properties any such mapping must satisfy in order to preserve subsumption, joins, or meets. Spanning sets are an abstract framework within which we portray and compare existing encoding techniques, and provide a context in which new encoding problems can be analyzed, leading to existing, related algorithms or, using other results we develop in this paper, guiding the development of new algorithms. We also explore the limits of minimal‐sized encodings, proving a lower bound for simple forms of encoding and showing that, in general, finding minimal‐sized encodings is NP‐hard. This paper can thus be viewed as both a synthesis of current research in taxonomic encoding and a repository of new results and directions for encoding as viewed from the perspective of spanning sets.

Tardis: a visual exploration environment for landscape dynamics

This paper presents the creation of a visual environment for exploring landscape patterns and changes to such patterns over time. Dynamic landscape patterns can involve both spatial and temporal complexity. Exploration of spatio-temporal landscape patterns should provide the ability to view information at different scales to permit navigation of a vast amount of information in a manner that facilitates comprehension rather than confusion. One way of achieving this goal is to support selection, navigation and comparison of progressively refined segments of time and space. We have entitled this system Tardis after the time machine of Dr. Who, to emphasize the exploration of time dependent data and because our use of elastic presentation has the effect of providing more internal space than the external volume suggests. Of special concern in this research is the extent of the data and its inter- relationships that need to be understood over multiple scales, and the challenge inherent in implementing viewing methods to facilitate understanding.

Sparse Term Encoding for Dynamic Taxonomies

Many domains of knowledge are related by a partial order. In Conceptual Graphs, both the type lattice and the generalization hierarchy exhibit a partial order structure. In many practical systems, the vast amount of taxonomic knowledge necessitates encoding that knowledge in a form that facilitates fast answers to taxonomic operations, such as computing greatest lower bounds. A number of researchers have proposed encoding algorithms and implementations. These techniques each have advantages and disadvantages, in terms of the space and time efficiency of the resulting encoding, the efficiency of the encoding algorithm (which is particularly important for dynamic knowledge) and the operations supported. Our main goal in this paper is to propose sparse logical terms as a universal encoding implementation. We show how sparse terms generalize bit-vectors, logical terms, integer vectors and interval sets, all of which have been used for encoding. As such, the algorithms developed for these implementations are directly applicable to sparse terms. We also argue that simple encoding algorithms (e.g. transitive closure and compact) implemented with sparse terms provide the efficiency and flexibility required for dynamic orders. We justify this claim using encoding results for theoretical and empirical partial orders.