The Physical Template

Transcription

1 The Physical Template Instructor: K. McGarigal Assigned Reading: Turner et al (Chapter 4); Swanson et al. (1988) Objective: Provide an overview of the physical template created by the abiotic environment as the foundation for landscape pattern. Highlight importance of the abiotic environment in creating a template for pattern formation via interaction with biotic processes and disturbances. Topics covered: 1. Conceptual framework agents of pattern formation. 2. Abiotic factors geology, geomorphology, hydrology, soils. 3. Climatic factors temperature, precipitation, solar radiation, meteorology. 4. Climate-landform effects on vegetation pattern, ecological flows, disturbance regimes and geomorphological processes. 5. Resource management implications climate change scenarios, conservation planning. Comments: Material taken from Turner et al. (2001) and Urban s Landscape Ecol course notes. 4.1

2 1. Conceptual Framework Given the myriad ways to define landscapes, it should not be too surprising that there are numerous agents operating at a wide variety of scales in space and time to create the heterogeneous patterns we observe and seek to understand. For heuristic purposes, it is useful to organize these pattern-forming agents into a conceptual framework to facilitate discussion, even though we acknowledge that this framework is an artificial construct that vastly oversimplifies how patterns are formed. Nevertheless, we recognize three primary agents of pattern formation in landscapes. The abiotic environment presents a physical template, the arena in which biotic processes and disturbance regimes interact to generate pattern. These agents will occupy us for the next three lectures. We begin by considering the physical template of landscapes, especially in terms of the manner in which terrain influences microclimate and edaphic factors. These features of the physical template have implications for studies that attempt to explain vegetation or landscape pattern, especially if those studies involve assessments of potential future scenarios of land use management or climatic change. Next we will consider the physical template of abiotic environmental heterogeneity as the arena in which a variety of biotic processes interplay to contribute to landscape pattern. Here we will consider just some of these biotic processes. For 4.2

3 vegetation pattern, these processes include the demographics of establishment, growth, and mortality as well as dispersal; competition may also play an important role. We will also be concerned with the interplay of these demographic processes and the way in which they interact with the physical template. In the third section, we will layer disturbance regimes onto the pattern generated by abiotic constraints and biotic processes. In this section of the course, we will focus in two ways: 1. First, we will focus on the role of these factors in pattern formation; the ecological consequences of those patterns will be the subject of subsequent sections. 2. Second, much of what we humans observe as landscape patterns is actually the spatial distribution of dominant vegetation types; e.g., forest versus grasslands versus desert. The dominant vegetation establishes the resource base for the rest of the ecosystem. The pattern in the dominant vegetation, therefore, affects the spatial patterning of all components of the system. Therefore, it behooves us to focus on the major factors affecting vegetation patterns. 4.3

4 2. Climate and Landform Constraints on Vegetation Pattern The abiotic environment presents a physical template, the arena in which biotic processes and disturbance regimes interact to generate pattern. We begin by considering the physical template of landscapes, especially in terms of the manner in which terrain influences microclimate & edaphic factors. These features of the physical template have implications for studies that attempt to explain vegetation or landscape pattern, especially if those studies involve assessments of potential future scenarios of land use management or climatic change. Vegetation patterns result, in part, from variability in climate and landform. Climate refers to the composite, long-term, or generally prevailing weather of a region, and acts as strong control on biogeographic patterns through the distribution of temperature, energy and moisture. Climate effects are modified by landform, the characteristic geomorphic features of the landscape, which result from geologic processes producing patterns of physical relief and soil development. Together, climate and landform establish the physical template on which the soils and biota of a region develop. A long tradition in gradient analysis by vegetation scientists has pointed to climate, specifically temperature (proxied as elevation), energy (proxied as various exposure indices) and moisture (proxied as various landform-based indices), as primary constraints on vegetation pattern. 4.4

5 2.1. Life zones The broad scale climatic control on vegetation is one of the foundational concepts in ecology. Each forest type and each species is distributed within a characteristic range of conditions and resources. The main driving conditions for vegetation are water, temperature and energy. These vary broadly at regional and continental scales, producing coarse patterns of biomes latitudinally and life zones altitudinally. Latitudinally, climate dictates biome zones. There are analogous changes in climate with altitude, producing life zones that are analogous to latitudinal biomes. As a general rule of thumb, 1 mile (5,280 ft or 1,609 m) change in elevation is the equivalent of approximately 800 miles (1,287 km) change in latitude. However, as discussed later, there are often inconsistencies between predictions of regional scale modeling based on climate and conditions at landscape and local scales, because local microclimate and landform have large influences on biophysical conditions and resulting vegetation. 4.5

6 2.2. Ecological Classification That broad-scale vegetation patterns along latitudinal and longitudinal gradients are strongly controlled by climate and landform is also the foundation for many ecological classification schemes, such as the National Hierarchy of Ecological Units (table below). In this classification, ecoregion boundaries are largely defined by major transitions in climate and landform. 4.6

7 2.3. Empirical Examples A number of recent research efforts have focused on predicting current and future vegetation conditions as functions of climatic and biophysical drivers. One recent study is McKenzie et al (2003) Climatic and biophysical controls on conifer species distributions in mountain forests of Washington State, USA. This study took existing vegetation survey data for several thousand vegetation plots on National Forest lands in Washington State and used climatic, physical, hydrological and soils variables to predict the distribution of conifer trees. 4.7

8 The main conclusions of their study include: Most species were relatively well predicted by moisture and temperature gradients, and most of these species showed a unimodal response of species occurrence to these gradients, consistent with Gaussian niche theory and indicating that the sampling regime was sufficient to capture the full gradient of these limiting factors. This kind of biophysical modeling of species niches is valuable to forecast redistribution on the landscape in response to climatic change, evaluate the predictions of simulation models, and alert managers to sensitive ecosystems and landscapes. 4.8

9 Rehfeldt et al (2006) Empirical analysis of plant-climate relationships for the western United States addresses the same question over a larger geographic area using related methods. This study also used existing vegetation plot data and climatic variables to predict species distributions. The authors predicted the climatic niche associated with the occurrence of trees and projected expected changes in species distributions due to climate change. They predicted that climate change will result in fairly dramatic changes in climatic regimes in much of the west over the next 90 years. 4.9

10 Their analysis projected regional climate changes from Global Circulation Models for 100 years into the future. They found that at 2060 much of the Great Plains and desert southwest was expected to have a climate regime that is extramural to existing climate (dark gray areas in the figure), which would mean it would support different plant species than are presently occurring. By 2090 they predicted that much of the intermountain west and Pacific Northwest would also have extramural climate regimes. This projected shift in climate would certainly have major impacts on species distributions, forest ecosystem structure and disturbance regimes. 4.10

11 They then projected changes in species distributions as a result of changing climate. For example, western Larch (Larix occidentalis) is projected to be largely eliminated from its current range in the United States by Similarly, changes in regional climate were projected to greatly reduce the distribution of Engelmann spruce (Picea engelmannii) in the lower 48 states. 4.11

12 The main conclusions of their study include: Mapped climate profiles of the species were in solid agreement with range maps. Climate variables of most importance for segregating the communities were those that generally differentiate maritime, continental, and monsoonal climates, while those of importance for predicting the occurrence of species varied among species but consistently implicated the periodicity of precipitation and temperature-precipitation interactions. Thus, they demonstrated that climate was an effective predictor of at least broad scale distribution of biotic communities and individual plant species. 4.12

13 Moreover, under projected future climate change scenarios, they concluded that: Projections showed that unmitigated global warming should increase the abundance primarily of the montane forest and grassland community profiles at the expense largely of the subalpine, alpine, and tundra communities and arid woodlands. However, the climate of 47% of the future landscape may be extramural to contemporary community profiles. Effects projected on the spatial distribution of species- specific profiles were varied, but shifts in space and altitude would be extensive. Species-specific projections were not necessarily consistent with those of their communities. 4.13

14 A number of studies have demonstrated the complex interaction between climate and landform, which can make predictions of vegetation changes under future climate scenarios challenging. One such study addressing relationships between regional scale climate and regional scale vegetation is McKenzie et al (2001) Recent growth of conifer species of western North American: assessing spatial patterns of radial growth trends. This study sought to separate trends from quasi-periodic changes due to drought cycles to identify the signature of climate change effects on tree growth. The main conclusions of their study include: In the southwest, many sites showed quasi-periodic patterns of drought, but few showed strong directional changes in growth rates of 185 of the sites showed significant increases in growth between 1850 and Most of the sites with large increases in growth were associated with high-elevation and high latitude sites in maritime climates. These are sites in which vegetation typically is temperature/energy limited, rather than water limited. Ecosystems that are water or nutrient limited would not be expected to necessarily exhibit strong increases in growth with increasing temperatures. 4.14

15 Another recent study is Goetz et al Satellite-observed photosynthetic trends across boreal North America associated with climate and fire disturbance. This paper measured recent changes in the productivity and growth of vegetation across the American arctic and boreal biomes. The authors charted changes in seasonal productivity over the 22 year period

16 The main conclusions of their study include: Growth in high latitude vegetation is expected with rising CO2 and temperature. Tundra vegetation has conformed to this expectation, with increased photosynthetic activity. In contrast, the response of interior forest areas to temperature change have been inconsistent with expectations of direct positive relationships between temperature and plant growth. This is because the growth of trees is dependent on a complex interaction between landscape position and moisture availability. In particular, drought stress, nutrient limitation, insect and disease damage change nonlinearly with temperature and CO

17 While climate clearly has been shown to play a dominant overarching role in driving the distribution of vegetation communities and individual plan species, landform also plays an important role and interacts with climate to exert even greater control on the distribution of vegetation. Swanson et al. (1988), in a seminal paper on the role of landform, described four general effects of landform on ecosystem patterns and processes: (1) Landform affects on local climate Elevation, aspect, parent materials, and slope of landforms affect air and ground temperatures and quantities of moisture, nutrients, and other materials available at sites within a landscape. For example, southwest facing slopes receive more solar radiation than north facing slopes, resulting in warmer, drier conditions. These topographic patterns are strongly related to the distribution of vegetation across a landscape. 4.17

18 (2) Landform affects on ecological flows Landforms affect the flow of energy, material, and organisms across the landscape. The funneling of winds by landforms, e.g., may influence dispersal pathways for wind-blown seeds. The position of lakes relative to groundwater flow pathways may strongly influence the biological and chemical characteristics of these lakes. Landforms may influence or constrain the movement of animals which can ultimately alter biological interactions and disturbance regimes. 4.18

19 (3) Landform affects on disturbance regimes Landforms affect the frequency and spatial characteristics of disturbances such as fire, wind, or grazing. In fire-dominated landscapes, e.g., topographic position can influence the likelihood of lightning-caused ignitions and topography can markedly influence the spread of fire once started. Similarly, in New England, topographic position can greatly influence susceptibility to hurricane damage. Grazing by domestic livestock is often concentrated on gentle slopes and in riparian zones along valley bottoms. These disturbance processes have a marked influence on vegetation patterns. 4.19

20 (4) Landform affects on geomorphological processes Landforms constrain the spatial pattern and rate or frequency of geomorphic processes, the mechanical transport of organic and inorganic material, that alter biotic characteristics and processes. Mass soil movements, e.g., such as landslides, are strongly influenced by landforms. Stream channel morphology, e.g., sinuosity, is strongly influenced by valley floor morphology. 4.20

21 One example of recent research showing how important microtopographical variability is in influencing vegetation is Bunn et al (2005) Topographical mediation of growth in high elevation foxtail pine forests in the Sierra Nevada, USA. 4.21

22 The main conclusions of their study include: They found a strong association between biophysical setting and age structure, and with ring-width patterns in foxtail pine. The strong relationship between biophysical setting and growth implies that forest vegetation composition and performance must be evaluated at multiple scales, with particular attention to the scale at which individual trees directly interact with the environment. Broad scale regional climate modeling is insufficient to predict forest conditions and responses at local and landscape scales. 4.22

23 3. Decomposing the Soil Water Balance Beyond being a primary constraint of vegetation pattern, the soil water balance provides an instructive illustration of the way in which physical factors, in particular, climate and landform, interact to generate complex environmental gradients in montane landscapes. The soil water balance (Stephenson 1990) is the balance between water supply and demand. Here we consider the proximate variables affecting each of these terms, with special attention to how one might estimate these at the landscape scale. 4.23

24 There are a number of well-established procedures for predicting some of these terms from physical terrain data and limited meteorological data, a process of statistical extrapolation and/or interpolation. For example, Running et al. (1987) illustrate methods for interpolating microclimate in mountainous terrain. Nikolov and Zeller (1992) developed a general model to predict solar radiation from temperature, precipitation, terrain, and latitude. Daly et al. (1994) have used lapse rates constrained by local physiography to interpolate precipitation data at landscape to continental scales. Other physical variables prove less tractable: soil is heterogeneous on all spatial scales and difficult to predict or interpolate. This is a problem for empirical as well as modeling studies. 4.24

25 4. Confounding Issues Inferences about effects of the physical template on vegetation are confounded by several issues: (1) In modeling vegetation responses to the physical environment researchers have often used topographical surrogates instead of direct measurements of the key drivers: temperature, water and energy. For example, elevation is often used as a predictor, implicitly as a proxy for temperature and water effects. However, vegetation does not respond to elevation per se, and elevation is an imperfect proxy for temperature and water. For example, growth may have a strong unimodal relationship with temperature, but there will be substantial variability in the relationship due to the variability in other important factors such as water, energy, soils. Elevation may have a moderately strong relationship with temperature, but there will be even more variability in the relationship due to orographic and topographical characteristics, such as aspect and slope. The relationship between growth and elevation then will be much weaker than either the growth-temperature or the elevation-temperature relationship because the error in these two combines multiplicatively in formulating the growth-elevation relationship. The main point here is two-fold: 1) the key resources that drive organism responses are physical characteristics, such as temperature, water and energy, and 2) these drivers vary in complex ways with topography and orography. This implies that attempts to use coarse topographical variables alone as proxies for species distribution and performance may not be very successful. 4.25

26 (2) The physical components of the gradients are intercorrelated and confounded. For example, in the mountains, temperature decreases while precipitation increases with increasing elevation. Consequently, it is exceedingly difficult to separate the effects of temperature and moisture on plant distribution in montane landscapes. 4.26

27 (3) Biotic responses to these physical factors may be through differential establishment, growth, or mortality; these may be further confounded by competitive interactions; and dispersal may either limit or amplify these patterns. Further, disturbance regimes also interact with the physical template. Thus, all three primary agents of pattern formation are interwoven in landscape pattern. 4.27

28 (4) Each physical factor has its own characteristic spatial scale, and so behaves as a "variable" at its own scale. For example: Soil depth and texture: centimeters (and every other scale!) Topographic convergence: 10's to 100's of meters (similar to the illustrations of the topographic moisture index presented with scaling techniques). Temperature: 100's of meters (elevation) or 100's kilometers (latitude). For example, oc temperature decreases roughly 4-6 for a 1,000 m increase in elevation (or 1,000 km latitude), depending on moisture content in the air. Precipitation: 100's of meters elevation in mountains. Precipitation also varies substantially within regions due to local physiography (e.g., via rain shadow effects). In the U.S., these gradients run east-west due to the effects of our mountain ranges (which run north-south) and the prevailing westerly winds. 4.28

29 Example: Scaling physical variables in Sequoia National Park: The characteristic scaling of the physical environment can be examined quantitatively using semivariance analysis. Urban et al. (2001) illustrated the characteristic scaling of four physical variables in Sequoia National Park. They used elevation as a proxy for temperature and precipitation as governed by lapse rates. They measured topographic effects on drainage as a topographic convergence index (TCI) computed as ln[a/tan(beta)], where A is upslope contributing area and beta is local slope angle (Moore et al. 1990). They used slope aspect, transformed as A'=cos(45-Asp)+1, where Asp is aspect in degrees (0=360=N). As transformed, this index aligns on a SW-NE axis and can be used as a proxy for radiation load. They also measured soil depth at some scales (but not all, due to logistical constraints). They computed variograms at three spatial scales: (a) a mixed-conifer reference stand (2.5 ha in extent, 50-cm horizontal resolution); (b) Log Creek Watershed (50 ha in extent, 5-m horizontal resolution); and the Kaweah Basin (~90,000 ha in extent, 30-m resolution). The variograms at the three scales suggest the scaling of the physical template. Scaling at the level of a forest stand (2.5 ha): At the scale of a single stand, elevation shows the signature variogram of a gradient, with no range obvious in the variogram. Aspect, TCI, and soil depth all show large nugget 4.29

30 variances, suggesting considerable variability at finer scales than the minimum resolution of the data. Scaling at the level of a small watershed (50 ha): At the scale of a small watershed, nugget variances are much reduced but elevation still shows the variogram of a gradient. Soil depth and aspect show ranges on the order of ~100 m, with TCI approaching its sill at slightly shorter distances. Scaling at the level of a large basin (90,000 ha): At the scale of the large basin, elevation still shows no characteristic grain, while TCI and aspect show ranges at distances of ~200 and ~500+ m, respectively. No soils data are available at this scale. Urban et al. (2001) also noted that the range of variability measured in any of these variables might also be expected to vary with scale. This was certainly the case with elevation: the standard deviation of measured values decreased by two orders of magnitude with a change in spatial extent from the large basin to a small forest stand. Thus, elevation was an important variable as measured over the basin, but it was essentially a constant at the stand level. By contrast, the other variables (TCI, aspect, and soil depth) showed similar ranges of variation (as standard deviations) at each of the three scales, reflecting the finer-grained measurements possible over the smaller extents. That is, changing scale (grain and extent) for these variables merely discovered finer-resolution expressions of the same variables. 4.30

31 Implications of Scaling Considerations: The important implication of this scaling lies in our ability to use these variables to explain or predict vegetation pattern: Explanatory variables vary with scale from microscale to regions. For example, in the Sierra Nevada, elevation gradients in temperature and precipitation predict the location of the mixed conifer zone as being a narrow range between "too dry" at low elevations and "too cold" at high elevations. Thus, either temperature or precipitation (or more proximately, elevation) could be used to predict the location of various community types (the white fir type, the Ponderosa pine type, and so on). But within each zone, neither temperature nor precipitation is a useful predictor of individual species distributions because all sites witness similar temperatures and precipitation. Rather, individual species distributions are predicted by finer-scale variables such as topographic position and soil depth. 4.31

32 Predictive models based on field data must vary locally as well as regionally; and a model derived for one site might not work elsewhere. In short, empirical models of the physical template are scale- and site-specific. 4.32

33 5. Implications of the Physical Template under Climate Change Climate change scenarios typically imply increases in temperature (T), and either increases or decreases in precipitation (P). In considering potential vegetation response, it is important to recognize that T and P affect different terms of the water balance (T affects water demand; P affects water supply); moreover, these are only two of several effective variables in the water balance. Work by Stephenson (1990) suggests that decomposing the water balance into its effective terms can have major implications under climate change scenarios; e.g., a site might be dry because of increased evaporative demand (increased T) or because of decreased water supply (decreased P), but these "different kinds of dry" might elicit different responses in vegetation. In Sierra Nevada, mid-elevation sites that are dry due to high evaporative demand tend to support Ponderosa pine, while sites that are dry due to low water supply tend to support Jeffrey pine. Sometimes a "space for time substitution" is invoked to make inferences about how systems might respond to climatic change. In this, the future condition of a site under climatic change would be predicted to be similar to present-day sites subject to similar climates. This approach is compromised by 2 considerations: Many of the effective variables in the water balance will not change under greenhouse scenarios (e.g., soils, topography). Predictions based on regional-scale variables (temperature, precipitation) cannot embrace the factors that drive the water balance--hence explain species distributions--within landscapes. 4.33

34 6. Applications in Conservation Planning Increasingly, conservation planning involves the preselection of potential sites for nature reserves, based on "proxies" or variables that are readily available and can be used in the absence of actual species distributional data. Given the strong role of the physical template in governing species distributions, terrain analysis offers itself as a compelling starting point for these approaches. For example, terrain analysis might help locate topographically distinctive sites that support unusual communities (e.g., cove forest, rock outcrops). Similar analyses might help delineate regions of especially high local heterogeneity with respect to topography, microclimate, soils, and so on. These sites would be good candidate sites for high beta-diversity communities. In either case, the proxy analysis would serve to identify potential sites, which would then be selected through further field assessments. Given the sometimes extreme logistics of field surveys in conservation planning or natural resource management, invoking the physical template in preliminary screening of potential sites could be a powerful research/ planning aid. The Nature Conservancy, for example, has been using this approach to build a portfolio of conservation reserves for each ecoregion. 4.34

35 More recently, TNC has proposed as an overarching biodiversity conservation strategy a conserving the stage approach in which the focus is on conserving the geophysical stage upon which species and communities develop and ecosystem functions occur. The basic approach involves classifying the landscape into a set of geophysical settings and within each setting prioritize locations based on the connectedness of a diversity of landforms and elevation (and wetland density in low-lying areas); the idea being that areas that are highly connected in terms of natural cover and that provide access to a diversity of abiotic environments will allow species and communities to adapt to climate and other environmental changes over time. In other words, the players can shift over time so long as the stage is there. 4.35