Applying Ecosystem and Landscape Models in Natural Resource Management

Robert E. Keane

• Publisher: CRC Press
• Year: 2019
• Print ISBN: 9780367340001
• eBook ISBN: 9781000732832
• Edition: 1
• Topic: Biological Sciences
• Subtopic: Ecology
• Index: Biological Sciences

Managing today's lands is becoming an increasingly difficult task. Complex ecological interactions across multiple spatiotemporal scales create diverse landscape responses to management actions that are often novel, counter-intuitive and unexpected. To make matters worse, exotic invasions, human land use, and global climate change complicate this complexity and make past observational ecological studies limited in application to the future. Natural resource professionals can no longer rely on empirical data to analyze alternative actions in a world that is rapidly changing with few historical analogs. New tools are needed to synthesize the high complexity in ecosystem dynamics into useful applications for land management.

Some of the best new tools available for this task are ecological and landscape simulation models. However, many land management professionals and scientists have little expertise in simulation modeling, and the costs of training these people will probably be exorbitantly high because most ecosystem and landscape models are exceptionally complicated and difficult to understand and use for local applications.

This book was written to provide natural resource professionals with the rudimentary knowledge needed to properly use ecological models and then to interpret their results. It is based on the lessons learned from a career spent modeling ecological systems. It is intended as a reference for novice modelers to learn how to correctly employ ecosystem landscape models in natural resource management applications and to understand subsequent modeling results.

1

Introduction

Who Needs This Book?

“The computer is not, in our opinion, a good model of the mind, but it is as the trumpet is to the orchestra—you really need it. And so, we have massive simulations in computers because the problem is, of course, very complex.”

Gerald Edelman

An ecosystem model—an abstract, usually mathematical, representation of an ecological system (ranging in scale from an individual population, to an ecological community, or even an entire biome), which is studied to better understand the real system (Hall and Day 1977).

ABSTRACT

Predicting what will happen as a consequence of management actions, or lack of action, is a fundamental goal of landscape planning. Historically, land management professionals used results from empirical studies, coupled with their own expertise and wisdom, to make these predictions. But then climate change made most of the findings of field studies and the accrued wisdom of professionals over the last 100 years somewhat limited for the management of tomorrow’s landscapes. To fill this void will be ecosystem simulation models, which may become critically important tools for managing landscapes in the future because they synthesize highly complicated ecological interactions into complex computer programs that can be used to simulate alternative management actions. This book was written for those people who are unsure of how to properly develop and implement a modeling project to assess the impacts of management actions in local to regional applications. It is meant to be used as a generalized manual for employing ecological models to solve resource management problems. This chapter describes why this book is important and who needs to use this book.

Why Do We Need Modeling?

So much of natural resource management depends on scientifically credible projections of future conditions under both passive and active management. Predicting what will happen as a consequence of management actions, or lack of action, is a fundamental requirement of the planning process. Historically, land management professionals used results from empirical studies, coupled with their own expertise and wisdom, to make these predictions. But then something happened. The rapid influx of carbon dioxide and other greenhouse gases into the atmosphere from human activities over the last 100 years has changed everything (IPCC 2007). Changes in the earth’s climate systems because of greenhouse warming are now rapidly creating new climate futures that have no analog in the recent past (Flannigan et al. 2009, Fei et al. 2017), and as such, most of the findings of empirical studies and the accrued wisdom of professionals over the last 100 years may not be entirely valid for the management of tomorrow’s landscapes (Gustafson 2013). Yet, the valuable information gained from results of past studies should never be tossed in the trash bin or relegated to distant archives as these results still have great value to management. They do, however, need to be interpreted in a brand new context and used in other ways. No longer can we assume that statistical correlations and empirical analyses done in the past will hold in our uncertain climatic future to effectively manage tomorrow’s resources (Scheller 2018). We need to integrate past study findings into a tool that is based on physical, chemical, and biological foundations to predict the consequences of alternative land management actions under various climate futures. And that tool has been around for the last 50 years—ecosystem simulation modeling: the computation of ecological responses over time using a computer program.

Ecosystem simulation models will be a critically important tool for managing landscapes in the future for a number of reasons. First, models are needed because ecosystems are remarkably complicated. Myriad interacting ecological processes result in complex biotic and abiotic responses (Bachelet et al. 2000, Allen 2007, Bockino 2008). It is nearly impossible for one person to understand the complexity of all possible ecological interactions, nor is it possible to collect enough data on these interactions to completely understand their consequences. Models also provide a means to synthesize state-of-the-art knowledge, current research findings, and general information into a tool that explicitly recognizes interactions and predicts various ecosystem responses as a result of changing conditions. Next, models can be used to extrapolate spotty empirical data over larger areas and for longer time spans to provide greater spatiotemporal scope for management decisions. Models also identify those ecological processes that are poorly understood and need further research. Comparisons of alternative management actions are generally thought to be the greatest strength of ecological modeling in natural resource management, but models can also be used for many other phases of management, from planning to real-time decision making, risk and hazard analysis, prescription development, and treatment prioritization, and they can be used across the many scales of management actions (Figure 1.1). Most importantly, models can be used to evaluate ecosystem and landscape responses as novel climates and disturbance regimes change over time. Many are now finding that describing landscape and ecosystem response as climates change is best done with ecological simulations (Loehle and LeBlanc 1996, Bachelet et al. 2003). While models are far from perfect, they may be the best tools we have for future land management.

Figure 1.1. A summary of the ways ecological models can be used in natural resource decision-making across temporal and spatial scales. Management decisions are often made at three broad spatial scales—stand (< 100 ha), landscape (< 200,000 ha), and region (> 200,000 ha)—and at three temporal scales—years, decades, and centuries. There are specific uses of models for each combination of space and time scale. Adapted from Reinhardt et al. (2001). Citations for models: ABM-agent based models as reviewed by An (2012); Biome-BGC-(Running and Hunt Jr. 1993); Climate Models-operational climate forecast models; FireBGCv2-Keane et al. (2011); FVS-FFE-Beukema et al. (1997); IBM-Individual based models such as Butler (2003); JABOWA-Botkin and Schenk (1996); LANDIS-Mladenoff (2004); LPJ-Spitfire-Bachelet et al. (2003); TOPMODEL-Beven and Freer (2001).

Why is This Book Needed?

This book was written for those people who are unsure of how to properly develop and implement a modeling project to assess the impacts of management actions in local to regional applications. It is meant to be used as a generalized manual for employing ecological models to solve resource management problems. This book is not an introduction to modeling—the literature is teeming with books on that subject. It is also not a manual for a specific model nor does it describe the use of models for different ecosystems, and it is not a handbook for building models—that would be a difficult and complicated task and far beyond the scope of only one book. More importantly, this book is not a comprehensive literature review of existing models and it is not a comparison of ecosystem models, as both would require a specific context in which to base the evaluation or comparison. This book was written with the primary purpose of providing general guidance on executing a modeling project for common natural resource management problems. I’ve found that, in many of the modeling projects, most people are not completely confident in how they prepared the model’s input and how they interpreted the output. This book is for them. If you are someone who has been assigned to evaluate landscape responses to management activities and you have decided that modeling is the appropriate tool but you have limited expertise in modeling, then this book is for you. In summary, the goal of this book is to provide enough information on how to employ a model for a natural resource management project so that the user has high confidence in the modeling results and is able to interpret model results in the proper context.

What Does This Book Contain?

It is important that the reader know that this book only deals with ecological simulation models, and specifically emphasizes those models that are implemented for terrestrial ecosystems in a spatial domain. It does NOT cover statistical modeling using multivariate analysis, although many ecological models have statistical sub-models embedded within their structures. Examples of statistical models are timber growth and yield models for simulating stand timber volume changes, species distribution modeling to determine species current and future ranges, and phenomenological modeling relating empirical data to measured biophysical gradients often using multivariate modeling. Moreover, this book does not cover coarse scale ecosystem models, such as Dynamic Global Vegetation Models (DGVMs), as these models simulate dynamics over spatial scales that are rarely used by resource managers.

There are many computer models that simulate various environmental concerns for natural resource management. These include hydrological models, wildlife models, tree growth models, aquatic ecosystem models, and a host of other resource models. This book deals mostly with a special class of ecological models that are often applied when solving natural resource management issues—Landscape Ecosystem Simulation Models (LESMs). The word “landscape” in LESM is used to denote that these models simulate processes in a spatial domain, which may or may not include spatially explicit models that directly simulate spatial processes such as seed dispersal and fire spread. The word “ecosystem” is used to emphasize that these models simulate multiple interacting ecological processes, and the word “simulation” is used to signify that the ecological processes are modeled over time. And last, the word “model” is used to represent that these are computer programs written to synthesize complex ecological ecosystem behaviors into algorithms that are programmed in a specific computing language. While LESMs are the central theme of this book, nearly all of the material would also apply to other non-spatial ecosystem simulation models (e.g., stand models) and other landscape simulation models that ignore spatial processes.

In general, the models that this book covers simulate four basic factors for terrestrial systems—climate, vegetation, disturbance, and management actions—at various resolutions and detail. Climate is important because it is the fundamental top-down process to which all ecosystems respond. However, due to the sheer number and incredible diversity of climate models, this book will only cover the types of data used in ecological models to represent climate; this book does not cover how to use climate models or how to select climate data sets. Vegetation is the major biota that is primarily responding to top-down processes of climate and it is also providing the energy and habitat for all other biota. Disturbances are perturbations that depend on climate and vegetation, and in turn, modify the two. Last, management actions are the primary feedbacks between ecosystem dynamics and human land use, and they provide managers the ability to craft alternative scenarios in natural resource management—the primary context of this book.

As mentioned, this book is for the person or team of people that is at the beginning of implementing a modeling project for natural resource management. A generalized set of steps to complete any modeling project is found in Figure 1.2. The first step is to become familiar with modeling terminology and fundamentals to better understand the modeling process (Chapter 2). Then, the preliminary modeling project design steps, such as setting an objective, forming the sideboards, selecting a model, defining the landscape, and designing the project are needed to develop a blueprint for the project (Chapter 3). Once the modeling project has been designed, the six essential phases of modeling must be completed (Figure 1.2). Initialization concerns quantifying the values of the starting conditions in the selected model (Chapter 4) and parameterization is quantifying the values of all parameters in the model (Chapter 5). Once these are done, it is necessary to calibrate the model to produce realistic output by adjusting input parameter and initial conditions (Chapter 6). Next, all modeling projects need some sort of evaluation to determine the quality and accuracy of simulated results, so Chapter 7 details steps used in how to validate model behavior using extensive evaluations. Then, the model must be executed to implement the project’s design (Chapter 8). Once all simulations are completed, the output must be analyzed to answer the project’s objectives. And last, there are various issues that are often encountered during any modeling project and information on how to address these issues is presented in Chapter 10. Readers of this book should be able to implement a comprehensive and successful modeling project using material from this book.

Figure 1.2. A generalize flow chart of all the steps involved in the planning, design, and implementation of a ecological modeling project as covered by this book. At the center of this chart are six critical phases in a modeling project (light gray) that form the critical chapters in this book. The remaining steps (white) are detailed in the context of the six critical phases.

This book may appear to be a linear, step-by-step guide for implementing ecological models, but many phases presented in this book could overlap in some modeling projects. T...

Table of contents

• Cover
• Title Page
• Copyright Page
• Dedication
• Acknowledgments
• Preface
• Table of Contents
• 1. Introduction—Who Needs This Book?
• 2. Modeling Fundamentals—What You Need to Know to Use This Book
• 3. Project Design—How to Plan a Modeling Project
• 4. Initialization—How to Begin a Simulation
• 5. Parameterization—How to Tune the Model for Local Applications
• 6. Calibration—How to Tuning the Model for Realism
• 7. Validation—Determining Model Uncertainty
• 8. Execution—Implementing the Model Project
• 9. Analysis—Evaluating Model Results
• 10. Issues—Things to Think About When Using Models
• Index