Disturbances can be defined as relatively discrete events that disrupt ecosystem, community or population structure, and change the availability of resources or the physical environment (Pickett and White 1985). Analysis of the disturbance regime of a forest can be of great value for understanding patterns of structure and composition, as well as being important for defining appropriate management interventions. Detailed treatments of forest disturbance are provided by Attiwill (1994), Barnes et al. (1998), Oliver and Larson (1996), Peterken (1996), and Pickett and White (1985).
Characterizing the disturbance regime typically involves assessing the severity, timing, and spatial distribution of the different types of disturbance affecting the forest. It is useful to note the difference between the intensity and severity of disturbance. Intensity refers to its physical force, or the amount of energy released, whereas severity refers to the amount of living biomass that is either killed or removed as a result of the disturbance (Spies and Turner 1999). Timingrefers to the seasonality (time of year), duration, frequency, and the return interval or rotation time of disturbance (Spies and Turner 1999).
Spatial distribution may include assessments of the extent and shape of disturbances, and can be characterized at different scales, for example at local, landscape, and regional scales (Barnes et al. 1998).
In order to characterize the disturbance regime of a particular forest, the following variables should therefore be measured (Gibson 2002):
● extentand spatial pattern of the disturbed area
● intensity, or the strength of the disturbance (for example, fire temperature or wind speed)
● severity, or the amount of damage that occurred to the forest (for example, number of individual trees killed or stems damaged)
● timing, including the frequency (the number of disturbances per unit time), the turnover rate or rotation period (the mean time taken for the entire forest area to be disturbed) and the turnover time or return interval (the mean time between disturbances)
● interactions between different types of disturbance (for example, drought increases fire intensity).
For each of these variables, the central tendency (mean, mode, or median) and variation (standard deviation or range, for example) should be calculated, and presented together with frequency distributions in order to fully characterize a disturbance regime (Gibson 2002).
The main types of natural disturbance affecting forests are fire, weather (including wind, ice, temperature, and precipitation), soil disturbance (erosion, deposition, flooding, and movement) and herbivory (Gibson 2002). Anthropogenic disturb- ances include logging, fuelwood cutting, road building, drainage, fire, forest clearing for agriculture and urban development, livestock husbandry, application
of chemical fertilizers and pesticides, aerial pollution, and many others.
Assessment of the impact of human activities on forests is of fundamental import- ance for conservation planning and management. Often, a key objective is to determine whether a given forest is able to withstand or tolerate a particular anthropogenic disturbance regime. Such analyses lie at the heart of defining approaches to sustainable forest management, and depend not only on character- izing the disturbance regime, but on understanding how the forest responds to different types and patterns of disturbance. An important principle is the idea of limiting human impacts to the frequency, size, and severity of disturbance to which species are adapted, leading to the development of forest management plans based on natural disturbance regimes (Spies and Turner 1999). However, this is often difficult because of a lack of understanding about past disturbance regimes, and the ecological impacts of current human activities.
The following sections describe methods for characterizing some of the most important types of forest disturbance. The methods described here are field techniques; remote sensing methods and GIS are also very widely used for assessing forest disturbance and are described in Chapter 2. Although the methods described here may be useful for characterizing the current disturbance regime of a particular forest, it can also be helpful to consider the likelihood(White 1979) or risk of disturbance occurring in the future. This is central to assessment of vulner- ability, an important consideration in conservation planning, which is described in section 8.4. Types of disturbance not considered here in detail include aerial pollution and insect attack, both of which are of major concern to foresters, who have consequently developed a range of techniques for their assessment (see Horn 1988, Innes 1993, Knight and Heikkenen 1980).
4.2.1 Wind
It is helpful to differentiate the typical (‘chronic’) wind disturbance of a site from the effects of relatively rare wind events, such as hurricanes or storms (Ennos 1997). Rare events can have a major impact on forest ecology by destroying trees.
The breakage or uprooting of trees by wind is referred to as windthrow or blowdown. This form of natural disturbance results from the interaction between climate, topography, stand structure, soil characteristics and the growth and form characteristics of individual trees. Damage can be in the form of stem failure, root failure or uprooting (Mergen 1954). Windthrow can be termed endemic or catastrophic. The latter results from winds with relatively long return periods and is influenced primarily by local wind speed and wind direction, whereas endemic windthrow results from peak winds with return intervals of less than 5 years and is influenced more strongly by site conditions (Lanquaye-Opoku and Mitchell 2005). Endemic windthrow is generally more predictable than catastrophic windthrow, and therefore most assessments of windthrow risk focus on the former rather than the latter (see section 8.4).
The impacts of wind disturbance on a forest can be assessed through a field survey, involving measures of the structure and species composition of the area Forest disturbance regimes | 149
affected (see Chapter 3) and the canopy gaps created (see section 4.4). Windthrow also results in production of woody debris (see section 7.2 for techniques for assess- ing deadwood volume). Treefalls not only create gaps in the canopy but also cause disturbance to the soil and influence the micro-environmental heterogeneity of the forest by creating pit-and-mound micro-relief with exposed root mats, bare mineral soil, and humus, as well as fallen logs. These features may provide suitable sites for seedling colonization in many forests, and are therefore often included in field surveys. For example, Ulanova (2000) mapped the distribution of pits and mounds caused by treefalls, as potential sites of seedling establishment. Some authors have used a decay scale for logs, with between five and nine divisions, to determine gap age or the time since a windthrow event (see, for example, Liu and Hytteborn 1991). However, the degree of decay of a log depends not only on the date of the windthrow but also on log size and species, its position (i.e. whether it rests directly on the ground or on broken branches), and its status at the time of the windthrow event (i.e. dead or alive) (Ulanova 2000).
Surveys may also be designed to assess tree condition, by recording different kinds of damage such as broken or bent stems, branch loss, snapped or uprooted trees, and root damage. The extent of canopy loss can be estimated visually by using methods for estimating canopy cover (see section 3.6.4) (Brommit et al.
2004). Wounding is often common in wind storms, caused by falling trees and branches, twisting, etc., and may also be recorded. Different tree species vary in the amount of crown damage caused by storms, reflecting differences in green wood strength and crown shape (Zimmermann et al. 1994). The capacity for recovery following wind damage, by resprouting of stumps or branches, also varies between species (Del Tredici 2001) and can usefully be recorded in a survey (Bellingham et al. 1994, Paciorek et al. 2000).
The amount of damage caused by wind is a function of wind speed, duration, and the direction from which the winds originate. Wind speed is usually measured by using a cup anemometer, consisting of three hemispherical or conical cups mounted on arms and attached to a spindle so that they can rotate in the wind (Coombset al. 1985). An alternative design, the vane anemometer, consists of a number of light vanes radially mounted on a horizontal spindle. The main sources of error for such mechanical anemometers is that they have a threshold below which the friction of the system prevents rotation, and their inertia causes an overestimation of wind speed when the latter suddenly drops (Coombs et al.
1985). Continuous measurements from anemometers can be recorded by using data loggers (see section 4.6). If such instruments are not available, records may be obtained from national weather centres or meteorological surveys. Further details of devices for wind measurement are provided by Grace (1977).
Tatter flagsoffer a relatively simple alternative method of measuring exposure to wind. Described by Lines and Howell (1963) and Rutter (1966, 1968a, b), the method employs flags made out of cotton mounted on poles or wires. It is import- ant that material is used that degrades with time—not the material usually used to make flags! The material should slowly degrade to lose 50–75% of its initial mass.
Tatter flags are also commercially available. The loss in mass or area of the flag measured over time can be attributed to attrition caused by wind. Studies of the tattering of flags in both controlled and field conditions have shown that the rate of tatter is closely correlated with wind exposure, but is also influenced by factors such as rainfall and atmospheric moisture (Jack and Saville 1973). Another possible source of error is freezing of flags in winter. As wind varies seasonally and from year to year, flags should be flown for 2–3 years and should be distributed throughout the area to be surveyed. Use of tatter flags in the UK has been described by Quine and White (1994).
Susceptibility of areas to wind damage can be assessed in several ways (Reed and Mroz 1997). It may be possible to identify soil types susceptible to wind damage because of shallow rooting depth. The incidence of wind damage in the past can potentially be ascertained from forest survey notes or weather records, and used to assess relative risks of different areas in the future. Climatic atlases often indicate severe storm frequencies. For example, Boose et al. (2001) describe methods for reconstructing hurricane disturbance in New England by using a combination of historical research and computer modelling. Wind disturbance history can also be inferred by means of dendrochronological techniques (see section 4.3). The pres- ence of physical evidence such as broken trees, fallen trees, and tip-up mounds can also be used to infer the previous occurrence of wind damage.
Examples of studies that have assessed the damage caused by windstorms on forests include Bellingham et al. (1994, 1995), Brommit et al. (2004), Burslem et al. (2000), Peterson (2000), Peterson and Rebertus (1997), and Zimmermann et al. (1994). Further information on the effects of wind on forests is provided by Coutts and Grace (1995), Ennos (1997), and a special issue of the journalForest Ecology and Management(2000), vol. 135. Approaches to modelling wind risk are briefly considered in section 8.4.
4.2.2 Fire
Fire temperatures can be measured by using thermocouples or temperature- sensitive paints. Thermocouples can be used to measure temperature by monitoring the voltage produced by the difference in temperature between two dissimilar metals, which are used in construction of the thermocouple. The temperature response of a thermocouple should be calibrated against a reference measurement.
Ideally a thermocouple with a small bead (or tip) and small wire diameter should be used; sheathed thermocouples are also preferred because they protect the thermocouple junction from soot, which can influence the measurements made (Saito 2001). However, useful information can also be gained from thicker wire thermocouples (Iverson et al. 2004). Thermocouples need to be connected to a data logger for measurements to be recorded (see section 4.6). Arrays of thermo- couples can be arranged in a grid and at different heights in a vegetation canopy, in order to characterize spatial variation in fire intensity (Jacoby et al. 1992).
Alternatively, pyrometerscan be constructed by painting spots of temperature- indicating paint (such as Temiplaq), which melt at different temperatures, on to Forest disturbance regimes | 151
the unglazed side of ceramic tiles or metal tags, which are placed in the forest before the fire then collected afterwards for inspection (Gibson 2002). The minimum temperature attained in the fire around each tile is determined as the highest- temperature paint spot that melted in the fire (Hobbs et al. 1984). Maximum fire temperatures can also be estimated through the use of Tempil tablets, which melt at different temperatures (Grace and Platt 1995). The tablets should be wrapped in aluminium foil before they are placed in the forest, to facilitate relocating them after the fire, and their melting points should be calibrated in the laboratory (Gibson 2002).
Iverson et al. (2004) compared thermocouples and temperature-sensitive paints to measure fire intensity (Figure 4.1) and found that maximum tempera- tures recorded by the two measuring systems were highly correlated. These authors recommended the use of temperature-sensitive paints if only maximum temperature is required, because of their substantially lower cost involved.
However, additional information can be collected by using thermocouples. For example, positioning of the thermocouples in a grid enables information on the rate of fire spread to be collected.
A further technique for measuring fire temperature involves use of an infrared camera and an image recording and analysis system, which is capable of obtaining a two-dimensional thermal image from a remote location (Saito 2001). This method has been used to produce temperature profiles in forest fires (Clark et al. 1999c).
Fire velocity can be measured by using a pitot tube, which can measure velocities in the range from a few metres per second to above 100 m s–1, covering most of the wind velocity range in forest fires (Saito 2001). The technique is relatively simple and can measure a one-dimensional velocity component, although three- dimensional measurements can be obtained by changing the direction of the pitot tube head when the flow is at a steady state (Saito 2001). The method is described by Sabersky et al. (1989).
Partially buried, capped 15 cm PVC canister containing
data logger
Aluminium tag with temperature-sensitive
paints
Temperature probe
25 cm
Buried cable
Approximately 2 m
Fig. 4.1 Arrangement of thermocouple temperature probes and temperature- sensitive paints, used in a comparative trial to measure fire intensity. (From Iverson et al. 2004. International Journal of Wildland Fire, CSIRO Publishing.
http://www.publish.csiro.au/nid/115/issue/871.htm.)
Trees may survive a fire but be damaged by it. Death of the cambium can lead to production of a fire scar, which is visible as a gap in the bark that is usually triangular in shape, becoming narrower with height, found on the leeward side of trees. Fire scars have been widely used to date fires, particularly in conifer forests, but also with some hardwood tree species. Such scars can be detected as blackened areas or damaged rings in increment cores (see following section), or in discs obtained from the trunks of felled trees (Schweingruber 1988). However, errors can be caused by missing, false, or indistinct rings, which should be corrected by cross-dating (Fritts 1976). This can be achieved by sampling trees that were undamaged by the fire, or sampling both sides of the trees affected. Further details of these methods are provided by Arno and Sneck (1977), Madany et al. (1982), and Schweingruber (1988).
Byram (1959) provided a simple index of fire intensity:
IHwV
where His the heat of combustion of fuel, wis the mass of fuel consumed per unit area, and Vis the heading rate of spread of the fire. This index is thought to correlate closely with tissue necrosis and possible tree mortality. Iverson et al.
(2004) indicate that measurements from thermocouples can be used to estimate I.
Fire intensity is usually measured as the rate of heat energy released per unit length of fire line per unit time (W m–1s–1) or sometimes per unit area and time (W m–2s–1).
An ability to predict the spread of fires is something of great value to forest managers, and consequently this has received much attention from researchers.
However, fire spread is influenced by a complex set of phenomena occurring over a range of scales, involving turbulent flow influenced by wind and topography, and the spatial distribution and amount of fuel available. A range of models have been developed in different parts of the world, which are reviewed by Weber (2001).
Further information about the ecology of forest fires is provided by Agee (1993), Johnson (1992), Johnson and Myanishi (2001), and Whelan (1995). Methods for evaluating fire risk are considered further in Chapter 8.4.
4.2.3 Herbivory
In order to clearly demonstrate the effects of herbivory on plant populations, prop- erly designed, replicated field experiments are required. Below-ground herbivory is generally assessed by applying some form of either physical or chemical exclusion as treatments in a field experiment. Chemical treatments range from the non- specific killing of most soil organisms through the use of chloroform fumigation (see, for example, Sarathchandra et al. 1995), to the use of biocides that are targeted to the control of specific organisms. Organophosphates such as isofenphos and ethoprop can be used to reduce populations of soil invertebrates, including coleopteran and lepidopteran larvae (Gibson 2002). Freezing of soil to -20 C can be used to reduce the activity of soil nematodes (Sarathchandra et al. 1995).
Experimental control of above-ground herbivory by invertebrates is often achieved Forest disturbance regimes | 153
through spraying insecticides such as carbaryl, chlorpyrifos, dimethoate, and malathion (Brown and Gange 1989, Gibson 2002). However, great care should be employed in the use of any chemical treatments; many can have unforeseen and potentially deleterious effects on components of the ecological communities other than those being investigated, and can have undesirable effects of the physical and chemical properties of soils. Some can alter plant growth, either positively or neg- atively, and this should be evaluated as part of the experiment (Gibson 2002).
Insecticides that might be of use in ecological research are listed by appropriate guidebooks such as Page and Thompson (1997), which should be consulted care- fully to ensure that the chemicals are selected, handled, and applied correctly (Gibson 2002) in a way that minimizes impact on the environment.
An important method of assessing herbivory impacts in forests is through the use of fenced exclosures, designed to exclude browsing mammals such as deer.
Comparison of areas within and outside such exclosures can provide insights into the impacts that the animals are having on variables such as the extent and composition of ground flora and the seedling establishment of trees. The effectiveness of the exclosure in excluding the animals should be evaluated; the presence of a fence is no guarantee that it will be effective. Some species of deer are capable of jumping fences of at least 2 m in height, and can prove remark- ably persistent in attempting to gain access to fenced areas. An appropriate design and size of fence should be used for the type of animal to be excluded, and animal populations should be monitored both within and outside the exclosures (see Sutherland (1996) for methods appropriate for censusing animal populations).
Fences should be regularly checked for damage and repaired, and fence supports need to be strong enough to withstand damage from any large animals that are present (Gibson 2002). For smaller mammals, an appropriate mesh size for the fencing material must be selected (i.e. 3–4 cm for rabbits, 0.5 cm for voles and mice; Gibson 2002). To exclude burrowing animals (such as rabbits), the fence must be partly buried in the soil. Care should also be taken to ensure that the fences do not have any negative effects on wildlife; for example, in Scotland, collisions with deer fences were found to be a significant cause of mortality in woodland grouse (Summers 1998).
The potential impact of herbivory on individual plants can be examined experi- mentally by introducing herbivores in appropriate cages, or by simulating the effects of herbivory by clipping leaves or other plant parts with scissors (Canham et al. 1994b). However, removal of plant material in this way may fail to mimic the effects of herbivory precisely (Gibson 2002). Generally, the effects of herbivory on individual plants are assessed by measuring damage to the plant parts consumed, although it should be noted that patterns of dry mass allocation within the plant may change in response to herbivory, and therefore this may need to be measured for a comprehensive assessment of impacts. Plant damage as a result of herbivory can be most simply assessed as the proportion of damaged leaves or shoots per plant (Gibson 2002). Alternatively, individual leaves can be assessed on a scoring system