Agricultural change

The pressures on the environment resulting from agricultural changes have received much attention. Major changes in agricultural activity since the 1940s in the UK include:

1. 1 Increasing levels of farm mechanisation, energy inputs and decline in the labour force

2. This long-term change accelerated after the Second World War.

3. 2 Development of highly productive strains of crops and livestock

4. For example, as a result of genetic improvements, spring barley yields have increased by an average of 0.84% per year in the last 30 years.

5. 3 Increasing use of fertilisers and pesticides

6. Fertiliser use has increased about 8 times in the last 50 years, with nitrate application showing the greatest increase at about 16 times. Coupled with the development of highly productive strains of crops, there has been a marked rise in crop productivity. Agricultural food production in the UK has increased by a factor of about 100% since 1955, for example yields per hectare of wheat and barley have nearly doubled, root crop production has increased by 50–75% and milk volumes per cow have increased by about 50%.

7. 4 Changes in farm size

8. Between 1949 and 1979 the number of farms which were larger than 122 ha in the UK increased from 12 317 to 16 765. Accompanying this has been an increase in field sizes to accommodate larger machinery and reduce the proportion of non-productive headlands used for turning. The removal of hedgerows, as a consequence, has been substantial (see below).

9. […]

Introduction

Investigators worldwide are fundamentally concerned about vegetation changes within forest ecosystems across spatial and temporal scales. The nature and extent of the impacts of these changes, as well as the feedbacks to global climate, are being addressed through modeling the interactions of the vegetation, soil, and energy components of the forest ecosystem. Mathematical models of the dynamics of forest succession and soil processes, combined with observations of forest ecosystems, can provide much of the insight required to comprehend these processes and construct initial cause-and-effect relations. In practice, model testing is complicated by the difficulties of making necessary observations at the required scales. Observations of ecological factors, such as the successional stage of landscape units, at regional scales and over periods of decades or longer are particularly difficult. To effectively utilize the modeling strategies described elsewhere in this book, large amounts of information are required concerning the scope and state of boreal forest ecosystems. Remote sensing technology provides the only feasible means of acquiring this information, repeatedly, with a synoptic view of the landscape. In addition, the incorporation of remotely sensed data into models of forest ecosystem dynamics can be used to characterize northern–boreal forest ecosystems, especially with regard to the interpretation of landscape patterns and processes at local and regional scales.

The launch, by the National Aeronautics and Space Administration (NASA), of the Landsat multi-spectral scanner (MSS) in 1972 ushered in a new era for terrestrial ecologists by providing a synoptic view of the global landscape over an extended period of time.

Introduction

The reproductive process plays an important role in determining species composition and distribution. At the landscape level, the variation in density and species composition of forests growing on similar, but geographically separated, sites is related to the quantity and type of reproductive material, the nature and severity of disturbance, and the growth requirements of the species comprising the forest. Within a given site, the spatial and age distribution of trees and associated plant species is, in part, a consequence of the nature of the substrate following disturbance and the ability of residual or new reproductive material to survive and develop under the prevailing biotic and abiotic conditions.

Regeneration of boreal forests varies in time and space. Temporal variation takes into account such variables as the periodic nature of seed production and changing seedbed conditions. Spatial variation includes factors such as seed dispersal distance and pattern of surface conditions as related to dynamics of sexual and asexual reproduction.

The state and dynamics of regeneration vary with the condition of the forest ecosystem. Natural disturbances, for example fire and major wind damage, remove or destroy variable amounts of the above-ground system, including the forest floor, and below-ground reproductive material, creating a continuum of conditions that determine the relative importance of sexual reproduction and vegetative regeneration to secondary succession. Colonization and forest development on newly formed or exposed sites (e.g. primary succession along rivers or following glacial retreat) is usually dominated by sexual reproduction, owing to the absence of pre-existing vegetation.

Introduction

The objective of this chapter is to define the properties of a unified boreal forest model. Models are constructed for specific purposes, which themselves guide the selection of, for example, which factors to include, and the level of abstraction and realism. In the case of the unified boreal forest model, certain tasks are obvious and have strongly influenced model construction. Fundamentally we require a model that will permit us to examine long-term forest ecological behavior under a wide range of stable and chronically or catastrophically shifting conditions. However, we can describe more specific tasks for which a boreal simulator is needed.

Most importantly, we lack a vehicle for learning how realistically boreal forests can be defined as abstract ecosystems. Thus, we may wish to develop a mathematical model that can predict forest behavior over both time and space, based on quantitative relations between growth processes and forces intrinsic and extrinsic to boreal forest stands. Construction of the model would simply reflect in part the desire to formulate in testable ways our hypotheses on the importance of various processes that affect the long-term behavior of forest stands. Iterative model construction, comparison of its predictions with field data, further model modification, and so on, allows us to understand what the real forests ‘know’ concerning the processes of interest that the model does not, and to correct the model accordingly. The model may also serve for exploring long-term ramifications of current temporal or spatial vegetation patterns, which are not otherwise intuitively obvious.

Introduction

The perception that there is a relationship between the patterns observed on a landscape and the set of physical and biological processes that generate those patterns is central to modern ecology. The concept was perhaps best elucidated in the classic 1947 presentation of A. S. Watt but was also an important construct in earlier papers by Watt (e.g. 1925) and others. The basic premise is to view an ecosystem as a working mechanism (Tansley 1935; Watt 1947). Such a mechanism, as a consequence of its internal interactions and interactions with the environment, produces the patterns that we see in nature. When one inspects Tansley's (1935) original definition of the ecosystem, one finds the same concepts that one sees in hierarchy theory today (Allen & Starr 1982; Allen & Hoekstra 1984; O'Neill et al. 1986; Urban, O'Neill & Shugart 1987).

Of course, the Watt–Tansley ecosystem paradigm has been reintroduced as a major ecosystem construct in ecological studies. One conspicuous reintroduction of these concepts was Whittaker's (1953) review, which used the Watt pattern-and-process paradigm to redefine the ‘climax concept’. These same ideas are also found in ecosystem concepts developed by Bormann & Likens (1979a,b) in their ‘shifting-mosaic steady-state concept of the ecosystem’, as well as in what Shugart (1984) called a ‘quasi-equilibrium landscape’. Given the richness of concepts developed by ecologists in the first half of this century, it is foolish to propose that any idea is new, but we are now in a position to extend the pattern-and-process paradigm in what may be fundamentally important ways.

Introduction

The extent of the circumpolar boreal forest strongly corresponds to macroclimate. Within its climatic limits, the system functions as a complex interrelation between solar radiation, soil moisture, the forest floor organic layer, nutrient availability, forest fires, insect outbreaks and vegetation patterns (Bonan & Shugart 1989). Bonan (1989a and Chapter 15 of this volume) has specified a model for the environmental regimes, which sets the limits driving boreal forest dynamics. He has linked this model with a forest succession model, a gap model, which simulates the demographic processes of tree populations through time within the environmental constraints, and a model of moss dynamics. The combined model mimics the large-scale dynamics of boreal forest (Bonan 1989a; Bonan & Korzukhin 1989).

Bonan's model simulated different stands well with respect to species composition, biomass and density. The ability to simulate such trends in quantitative characteristics of tree species is a robust feature of gap models in general (Shugart 1984; Leemans & Prentice 1987). This robustness is largely a result of the coupling of growth responses of individual trees to environmental factors. If the annual growth of a tree declines as a result of environmental conditions, its chances of dying increase. Thus, the individual tree is removed from the plot, leaving room for better-adapted individuals. This aspect of the traditional gap models appears to provide a robust ability to reproduce composition, biomass and density. If more precisely defined forest structures are used to test such models, the models are often less successful.

Introduction

Major patterns of plant communities and species distribution are induced by various disturbance regimes operating at different spatial and temporal scales (Loucks 1970; White 1979; Bormann & Likens 1979b; Delcourt, Delcourt & Webb 1983). The development of temperate forests is controlled by canopy disturbance associated with single and multiple treefalls creating small gaps (Henry & Swan 1974; Oliver & Stephens 1977) and, occasionally, larger gaps along windstorm tracks (Canham & Loucks 1984). In contrast, fire disturbance in the boreal forest creates all sizes of canopy gap, and constitutes one of the most important community processes in vegetation development along complex environmental gradients (Heinselman 1981b). At the landscape level, ecological disturbances result in the spatiotemporal development of the vegetation mosaic composed of an assemblage of stand patches of different age, areal extent, and floristic composition (Pickett & White 1985).

The boreal forest is one of the world's two largest forest belts; it covers more than 126 km2 (Baumgartner 1979). The North American segment of the circumboreal forest, which constitutes a well-delineated biome both geographically and ecologically (Fig. 5.1), will be analyzed here with respect to fire disturbance. The boreal forest (excluding the Cordilleran region) spans more than 10° of latitude in eastern and western Canada; it is somewhat contracted south of Hudson Bay and James Bay, and in Alaska. At the continent scale, the boreal forest is a floristically poor biome (Takhtajan 1986) with only nine tree species dominating regionally or throughout the range, in coexistence with a subdued under-canopy flora in dense stands and an ubiquitous cryptogamic flora in open stands.

Introduction

Strong advances in understanding forest ecology and forest-ecosystem responses to disturbance and environmental change have come through the application of systems analysis and simulation in studies of small forested ecosystems. For example, the whole series of gap models of forest succession (e.g. JABOWA (Botkin, Janak & Wallis 1972), FORET (Shugart & West 1977), FORTNITE (Aber & Melillo 1982), LINKAGES (Pastor & Post 1985), FORENA (Solomon 1986) and FORSKA (Leemans 1989); see Shugart (1984) for an overview of several of these models) examines forest ecosystems at the spatial scale of one or several large trees, i.e. about 0.1 ha. These models represent areas of sufficient spatial extent to represent adequately important processes such as inter- and intra-specific competition, and soil–vegetation–atmosphere interactions. Results from such models might reasonably be scaled up spatially to the level of the stand, where stand is defined as an ecosystem with a relatively homogenous community of trees and relatively homogenous site conditions compared with neighboring ecosystems. However, generalization of results from stand-level models (as the gap and other microcosm models will now be called) to forest ecosystems containing many stands, without the use of specially formulated forest simulation models, can be dangerous, if not absolutely misleading.

Ecologists generally recognize that ecosystem boundaries are more or less arbitrary. Some such boundaries are easy to assign and are ecologically very meaningful, such as the perimeter of an island in a lake, or the perimeter of a farm woodlot surrounded by field crops.

Introduction

Models for simulating different aspects of vegetation dynamics have become increasingly popular during recent decades. Initially, mathematical modeling was only accessible to well-trained biomathematicians, but with the increasing availability of small and faster computers and with the development of modern software, it has been applied by more traditionally trained ecologists and foresters. Recently, many papers that present different models and applications within ecology have been published (e.g. Emanuel et al. 1984; van Tongeren & Prentice 1986; Running & Coughlan 1988; Tilman 1988; Costanza, Sklar & White 1990; Keane, Arno & Brown 1990).

Simulation models can help in the understanding and management of ecosystems. Such models are usually the only tool available for translating a collection of hypotheses for ecological processes into a testable representation of how the whole ecosystem functions. Simulation models can be used not only to evaluate hypotheses generated by field studies and ecological experiments, but also for situations where the more traditional ecological approach is less applicable, for example for studies that span several research generations, such as the study of processes involved in forest succession and gap-phase replacement of individual trees within a stand (Watt 1947). Ecological hypothesistesting by experiment and field studies for such long-term and largescale processes is almost inevitably incomplete and must be supplemented by simulation experiments.

Simulation models consist of a collection of hypotheses, most often in equation form. These hypotheses define how the major parts of the model change over time (Swartzman & Kaluzny 1987).

In the boreal zone, precipitation exceeds evaporation and the forests are inclined to be paludal. There are two contradictory points of view regarding the nature of the transitions between boreal forest and bog. The first such view conceives of irreversible paludification of the forests due to the advancement of bog and self-paludification of the forests. The second surmises a dynamic equilibrium between forest and bog shown initially by the general lessening of the paludification process in recent times and secondly by the periodic afforestations of bogs and the depaludification of paludified forests.

The transitions between boreal forest and bog may be represented as a phytocenotic, continual series of ecosystems. Transitions among these ecosystems are reversible in time and space. A paludification series consists of automorphous forest → paludal forest → bogged forest → treed bog → open bog → regressive lake-bog complex; a series of depaludification is regressive lake-bog complex → secondary open bog or secondary treed bog → secondary bogged forest → automorphous forest. This work covers the latitudinal belt of taiga of the West Siberian Plain (Fig. 9.1) that consists of northern-, central-, southern- and subtaiga subzones and is limited by permafrost in the north and soil salinity in the south.

Major works on the transitions between forest and bog

According to Sukachev (1914a), there are two opposite pathways of mire evolution. The first pathway involves intensification of moistness with peat accretion and an eventual transition to ombrogenic supply (Abolin 1914).

Introduction

It has been a long-standing tradition in mathematical ecology to use dynamical equations in the modeling of population dynamics. These equations are intended to describe the ecological mechanisms ruling the observed dynamics of a system, and radically excel formalized descriptions such as regression formulae, which feature widely in the modeling literature, including that for forest modeling.

The various forms of forest (or stand) dynamics models can be subdivided, for our purposes, into three types (for a full review and details see Shugart (1984)):

1. (a) Models with zero-level averaging over a tree population. They are based on the dynamics of individual trees with individual coordinates: spatial models in the exact sense of the word (Newnham & Smith 1964; Ek & Monserud 1974; Mitchell 1969; West 1987; Gurtzev & Korzukhin 1988).

2. (b) Models with an intermediate degree of averaging. They operate by considering individual trees but without their spatial location (Botkin, Janak & Wallis 1972; Shugart & West 1977; Shugart 1984; Huston & Smith 1987; Leemans & Prentice 1989), and form the family of so called ‘gap models’ (see Shugart & Prentice, this volume, Chapter 12).

3. (c) Models with the maximum degree of averaging. They operate by considering groups of trees – parts of a whole population – chosen depending on the particular task (usual subdivisions are by size and/or by age) (Ek & Monserud 1979; Kapur 1982; Korzukhin, Sedych & Ter-Mikaelian 1987, 1988; Tait 1988; Korzukhin, Antonovski & Matskiavichus 1989; Prentice et al. 1989; Kohyama 1989).

Introduction

The development of mechanistic approaches in the study of forest ecosystem dynamics over the past 16 years has increasingly drawn scientific attention to autecological characteristics of forest tree species. Model exercises have shown that emergent properties of forest ecosystems are predictable from interactions of species-specific life history attributes (e.g. Shugart 1984). This chapter presents silvical data for the dominant boreal tree species in the northern hemisphere to be used with ecosystem analysis and modeling studies of the circumpolar boreal forests. The selection of the species is based on distribution maps provided by Sokolov, Svyaseva & Kubly (1977) for the Eurasian species and by Fowells (1965) for the North American ones, as well as on boreal zone maps by Hämet-Ahti (1981). Fourteen tree species have been found to dominate the boreal zone in Fennoscandia and the USSR (Fig. 2.1); fifteen dominate North America. Section 2 of this chapter presents autecological reviews for the boreal tree species in Eurasia. Detailed reviews for the North American boreal tree species have been omitted since they are already available (e.g. Fowells 1965; Harlow, Harrar & White 1979) and because of publication limitations. However, North American boreal tree species have been included in Section 3 of this chapter, which provides species life-history data for parametrization of boreal forest simulation models. Data for silvics of the Eurasian boreal tree species have been extracted from both published and unpublished literature sources in six languages (Bulgarian, English, Finnish, German, Russian and Swedish) and have been collected according to the following scheme.

Introduction

Human activities are increasingly affecting the relation between the biota and the physical environment. That has long been true of resource developments that have transformed vegetation on a regional scale. Now, however, the scale of human influence has increased to a planetary one because of the modification of the atmosphere by the accumulation of greenhouse gases and industrial pollutants (Clark & Munn 1986). The result could well be a significant increase in global temperature, as most general circulation models predict, exaggerated in the northern regions now occupied by the boreal forest. But our state of knowledge of such global-scale processes is sufficiently incomplete that the magnitude and location of those changes are highly uncertain. How, then, can we assess the impacts on vegetation, when we are so uncertain of the changes that might occur to the physical environment of northern regions? One way is to turn to the past to gain insight.

Certainly geophysical processes have led to planetary changes in the past that were extreme enough to trigger profound shifts in climate and in vegetation. When those produced pronounced shifts between glacial and interglacial conditions, the vegetation was transformed and individual species interactions became uncoupled to form a variety of transient assemblages very different from either those that preceded the shift (Wright 1987) or those that now characterize major biomes (Davis 1981).

Even modest changes in climate can be amplified if the frequency and extent of disturbance of vegetation is changed.

This is an era of increased interest in the function and interaction of the major geophysical, geochemical and ecological systems of the earth. The interest in these large spatial scale studies has had diverse origins: the success of the ‘International Geophysical Year’ of global observations (1957–8) and a shared comprehension of just how much time has passed since this effort; the characterization of the surface of the earth from an ever increasing availability of images from space; the realization that humans are altering the composition of the atmosphere; a relative warming in international political tensions and the increased likelihood of sustained international scientific exchanges; an improved understanding of the past dynamics of the earth's surface resulting from radioisotope dating and analysis of paleoecological data; and the ramifications of computers with the power to solve complex equations of the fluid motion of the atmosphere and oceans. The conjunction of these and many other developments have turned the interests of many scientists in different disciplines to the issue of increasing the level of understanding of the earth as an interacting, dynamical system.

However, for all of these exciting developments, scientists in one of the important disciplines contributing to the study of the working of the planet, ecological sciences, are focused largely on the understanding of biota – environment interactions at very short time and space scales. Kareiva & Andersen (1988) found that about 80% of the studies in a sample of the journal Ecology were developed on areas less than 100 m2 Weatherhead (1986) sampled studies in the three areas of ecology, evolutionary biology and behavior and found the average duration of study to equal 2.5 years.

Introduction

The boreal forest environment is characterized by a wide range of site conditions. Climatic conditions range from extremely cold, dry continental regimes in interior Alaska and Siberia to warmer, moist, oceanic regimes in eastern Canada and Fennoscandinavia (Hare & Hay 1974). Extreme temperatures as low as −70°C are not uncommon in interior Alaska and Siberia (Rumney 1968). In more moderate regions such as Saint John, Canada, extreme lows average −23°C (Hare & Hay 1974). Annual precipitation can be as little as 10–20 cm in the dry continental regions of interior Alaska and Siberia, but as much as 50–90 cm in eastern Canada (Rumney 1968). Annual global solar radiation varies from less than 3352 MJ m−2 at the tree-line in western Canada to over 5028 MJ m−2 in the south (Hare & Hay 1974). Maximum day length varies from 16 hours at the southern edge of the boreal forest to 24 hours at the northern tree-line (Hare & Hay 1974).

In North America, soil moisture ranges from xeric jack pine (Pinus banksiana Lamb.) and black spruce (Picea mariana (Mill.) B.S.P.) forests to lowland black spruce and tamarack (Larix laricina (Du Roi) K. Koch) bogs (Larsen 1980). Soil temperature ranges from widespread warm, permafrost-free soils to scattered permafrost soils in interior Alaska and western Canada to extensive cold, permafrost soils in central and eastern Siberia (Larsen 1980). Local soil temperature gradients can be large. In the discontinuous permafrost zone of interior Alaska, growing season soil temperature sums above 0°C range from as low as 483 in wet, permafrost soils to as high as 2217 in dry, permafrost-free soils (Viereck et al. 1983).