Below is the uncorrected machine-read text of this chapter, intended to provide our own search engines and external engines with highly rich, chapter-representative searchable text of each book. Because it is UNCORRECTED material, please consider the following text as a useful but insufficient proxy for the authoritative book pages.
3 Forest Disturbance and Management Effects on Hydrology Forest hydrology is the study of water in forests: the movement, distribu- tion, and quality of water as regulated by forests. Forest hydrology addresses the hydrologic processes within forested areas and the output of water resources from forested areas. Disturbancesâboth planned and unplannedâand manage- ment activities in forests (see Chapter 2) can significantly alter hydrologic proc- esses. These changes can affect nearly all components of forest ecosystems, including surface erosion, slope stability, nutrient cycling, channel morphology and aquatic organisms, and the quantity and quality of water. This chapter defines forest hydrology, describes the factors that produce change in forests, and lists the general principles of forest hydrology (Figure 3- 1). The chapter then provides an overview of forest hydrology findings to date and evaluates how this science supports the management of forests for water. FOREST HYDROLOGY SCIENCE Forest hydrology draws on forestry, including silviculture and forest water- shed management, as well as civil, environmental, and hydraulic engineering; ecohydrology; geomorphology; soil science; and water resources engineering. It combines field measurements, experiments, and modeling to characterize and predict hydrologic processes and water resources. Principal instruments include precipitation and streamflow gages; devices for collecting and measuring water in tree canopies; thermocouples and other devices to measure sap flow in trees; wells, piezometers, lysimeters, and other devices to measure soil water tension, soil water content, and water table depth; and many types of devices to charac- terize chemical composition of water in trees, soils, and streams. The principal questions of forest hydrology are: â¢ What are the flowpaths and storage reservoirs of water in forests? â¢ How do modifications of the forestâincluding both trees and forest soilsâinfluence water flowpaths and storage? and â¢ How do changes in forests affect water resources from forests? Forest hydrology science relies on watershed studies, plot studies, process studies, and modeling. âPaired watershedâ studies are an important approach to forest hydrology. In a paired watershed study, stream gages are installed at the mouths of two or more watersheds, and the watersheds are manipulated to de- termine the effects of experimental forest treatments on streamflow or water quality. Paired watersheds are similar in size, land use or land cover, and other 45
46 HYDROLOGIC EFFECTS OF A CHANGING FOREST LANDSCAPE Modifiers of forest hydrology Forest disturbance: Forest management: Wildfire Forest harvest & silviculture Insects & disease Road networks Species changes Grazing Hydrologic response: General principles 1. Changes in 2. Changes in 3. Changes in forest flowpaths in water, soil structure soil and subsoil chemistry Specific hydrologic responses Hydrologic responses within forests: Interception & transpiration Infiltration & overland flow Water flowpaths in soil and subsoil Changes in watershed outputs: Water yield Floods Lowflows Sediment Chemistry Temperature Managing forests for water FIGURE 3-1 Forest hydrology examines the flowpaths and storage of water in forests, and how forest disturbance and management modify hydrologic responses. Hydrologic re- sponses to changes in forests fall into three categories of general principles, as well as specific hydrologic responses, discussed in the text. The final section of this chapter evalu- ates the state of knowledge of forest hydrology and its implications for managing forests for water, including feedbacks to processes that modify forests.
FOREST DISTURBANCE AND MANAGEMENT EFFECTS ON HYDROLOGY 47 attributes. After a reference period (usually five or more years), the âtreatedâ watershed is subjected to manipulations (i.e., forests cutting, road building, fire, herbicides, etc.), but the âcontrolâ or reference watershed is not. The change in the relationship of streamflow and water quality between the treated and control watersheds before and after treatment is defined as the effect of the forest treat- ment. Most paired watershed studies in forest hydrology were begun in the 1940s through the 1960s, and many of these studies had been abandoned by the 1980s. Field measurements from plots (segments of hillslopes), plot-scale experi- ments, process studies, and hydrologic modeling also are important components of forest hydrology. Field measurements from control and treated (e.g., burned, watered, cleared) forested plots have provided important insights into key hy- drologic processes such as interception, transpiration, infiltration, and overland flow. Hundreds of hydrologic models have been developed to represent and predict water quantity and quality, and many of these models have been applied to forested areas; models are especially important for areas lacking field meas- urements and experiments. In the twentieth century, forest hydrology science was conducted primarily in relatively small areas: segments of hillslopes (âplotsâ) or small watersheds that ranged in size from a few square meters to 1-2 km2. Time scales for plot- scale research commonly spanned a few growing seasons. Watershed studies have mostly been limited to one or two decades, but some, especially those in U.S. Forest Service (USFS) experimental forests and Long-term Ecological Re- search (LTER) sites sponsored by the National Science Foundation (NSF), have spanned periods up to 60 years. Most watershed studies occurred on publicly owned forestland, but some were conducted on privately owned land. MODIFIERS OF FOREST HYDROLOGY Forests are dynamic systems. Forests can be modified by: (1) fire; (2) spe- cies changes; (3) insects and disease; (4) forest harvest and silvicultural activi- ties; (5) roads and trails, including skid trails; and (6) grazing (see Figure 3-1). Hydrologic Effects of Fire Record-breaking fires in recent years and projected increases in high- severity wildfires in the western United States (Fried et al., 2004; Westerling et al., 2006) have contributed to increased interest in how fire in forested systems affects water (Neary et al., 2005a). The type and magnitude of post-fire effects on runoff and water quality vary greatly with fire severity (Box 3-1), vegetation type, topography, soil type, subsequent amount and type of precipitation, and other local conditions. Despite these variances, fires can increase runoff and erosion rates by one or more orders of magnitude (Tiedemann et al., 1979; Helvey, 1980).
48 HYDROLOGIC EFFECTS OF A CHANGING FOREST LANDSCAPE BOX 3-1 Fire Severity and Hydrologic Effects Most researchers and resource managers use fire severity as the primary means to characterize and predict the hydrologic effects of fires. Fire severity differs from fire inten- sity. Forest fire severity refers to the amount of tree mortality, which is related to fire effects on the ground surface. Fire intensity is the heat released per unit time per unit length of flame front. Fire severity is usually divided into three classesâlow, moderate, and high. In low- severity fires the surface of the litter layer is blackened or partially consumed and some understory vegetation may be scorched or burned, but some charred or unburned litter is still present on the soil surface (Wells et al., 1979). In moderate-severity fires the litter layer is completely consumed, but the underlying mineral soil is not physically or chemically al- tered. High-severity fires not only burn all of the surface litter, but the greater soil heating also consumes some of the surface organic matter. This results in a disaggregation of particles in the uppermost layer of mineral soil and, depending on the soil type and amount of heating, a discoloration of the soil surface. In high-severity fires, more than 75 percent of the forest canopy is killed, and the forest floor may contain cavities where stumps and roots have been completely consumed as well as lines of white ash where coarse woody debris once covered the surface. High-severity fires typically have a much greater hydrologic effect than low- or moder- ate-severity fires. The effect of wildfires on life, property, and aquatic resources means that wildfire risk is a major factor driving forest management, particularly on public lands. Ap- proximately 45 percent of the U.S. Department of Agriculture Forest Service budget is now being devoted to fire suppression and fuels management, which is nearly double the pro- portion allocated in the 1980s and 1990s. Hydrologic Effects of Changing Species Composition in Forests The composition of forest species changes as a result of natural disturbance and forest management. During forest succession after disturbance, early suc- cessional species are replaced by late successional species, leading to changes in water yield (Swank and Crossley, 1988; Jones and Post, 2004). Forest manage- ment may modify forest species, such as replacing deciduous forest with conifer species; these changes modify water yield and timing (Swank and Crossley, 1988). In addition, wildlife management, particularly the reduction of predators, may increase native ungulates and indirectly modify forest species composition by altering the intensity of browsing (Gill and Bearsall, 2001). The hydrologic responses to these indirect effects of wildlife management are not known. Hydrologic Effects of Insects and Diseases A wide array of insects and diseases can be found in forest ecosystems. In- sect and disease dynamics are closely coupled to climate and forest disturbances such as wind, ice, and fire. The two groups of insects of primary concern to forest managers are bark beetles and defoliators (Schmid and Mata, 1996). Bark beetles bore into the living tissue and weaken or kill trees by introducing fungi
FOREST DISTURBANCE AND MANAGEMENT EFFECTS ON HYDROLOGY 49 and disrupting the transport of water and nutrients between the tree crown and the roots. Recent bark beetle infestations have killed extensive areas of mature trees in short periods of time. Defoliators generally lay their eggs in the buds of trees and feed on the emerging new leaves or needles. In contrast to bark bee- tles, defoliators take several years of repeated, heavy attacks to kill a tree. Large, relatively infrequent outbreaks of insects or disease in forests are of- ten a matter of public concern; they are also ubiquitous disturbance mechanisms in forest ecosystems. Low levels are termed endemic, whereas a rapid increase in insect populations or the incidence of disease is termed an outbreak. Histori- cal photographs and tree ring records indicate that outbreaks of native species of insects and native diseases are intrinsic processes that are a natural part of forest ecosystems. However, introduced insects and pathogens are primary drivers of forest change and have transformed forests in the United States. Large outbreaks of native insects and dramatic forest decline due to the in- troduction of exotic pests and diseases fueled early interest in the effects of in- sects and pathogens on forest hydrology (Bue et al., 1955; Bethlahmy, 1974). Recent increases in forest area affected by insect outbreaks and possible links to fire suppression (Fleming et al., 2002; Bebi et al., 2003) have reignited scientific interest in the effects of insect and pest outbreaks on water quantity and quality. However, very few studies have been conducted on the hydrologic effects of insects and disease. The hydrologic effects of insects and disease can be ex- trapolated from general principles derived from studies of timber harvest and fire (MacDonald and Stednick, 2003; Uunila et al., 2006), but much remains to be understood about hydrologic effects of insects and disease. Hydrologic Effects of Timber Management and Silviculture Since the 1950s and 1960s, when timber harvesting expanded on federal forest lands and industrial forestry developed on private lands, an extensive lit- erature has examined the effects of forest management. Forest hydrology stud- ies have addressed the effects of silvicultural practices (such as site preparation, herbicide treatment, and thinning); forest protection (such as post-harvest slash burning); and timber harvest (especially removal of trees, and construction of roads and trails) on water quantity and quality. Along with fire, insects, and disease, these are the primary processes that modify forests. Most of these stud- ies have occurred in small plots or in small, experimental watersheds. Forest management effects on water quantity, quality, and timing vary with the area treated, the type of treatment, forest type, soils, climate, and time after treatment (Hibbert, 1967; Anderson et al., 1976; Bosch and Hewlett, 1982; Swank and Crossley, 1988; Hornbeck et al., 1993; MacDonald and Stednick, 2003; Jones and Post, 2004; Brown et al., 2005; Moore and Wondzell, 2005).
50 HYDROLOGIC EFFECTS OF A CHANGING FOREST LANDSCAPE Hydrologic Effects of Roads and Trails Forest management for timber and firefighting in the latter half of the twen- tieth century relied heavily on trucks and other heavy machinery for skidding logs to landings, constructing fire breaks, and hauling logs to mills. Starting in the 1940s, extensive road networks were constructed on public and private for- estlands in the western United States, and heavy machinery was used on forest soils throughout the country. Obvious soil disturbances associated with mechanized harvesting equip- ment and conspicuous landsliding associated with forest roads led to early inter- est in the effects of roads and trails on forest hydrology (Megahan, 1972; Ander- son, 1974; Harr et al., 1975; Swanson and Dyrness, 1975; Ziemer, 1981). Con- tinued hydrologic and sedimentation effects of lengthy road networks combined with efforts to decommission roads have fueled continued study of the effects of roads and trails on water quantity, timing, and quality (Reid and Dunne, 1984; King and Tennyson, 1984; Wemple et al., 1996; Bowling and Lettenmaier, 2001; Lamarche and Lettenmaier, 2001; Wemple et al., 2001; Wemple and Jones, 2003; Coe, 2006). Roads affect water timing and water quality, but the magnitude of the effect depends on road design, slope steepness, soils, and the configuration of the road system relative to the stream and river drainage net- works. Hydrologic Effects of Grazing in Forests In 1970, about half (85 million acres) of the western forests and about four- tenths (161 million acres) of the eastern forests were grazed, and about one-half of the areas grazed in eastern forests was âexploitatitve,â or beyond acceptable management (Anderson et al., 1976). Forest Service researchers (Platts, 1981) estimated that more than 800 million acres in the United States were grazed by livestock in 1970, furnishing 213 million animal unit months of forage. Over- grazing (animal densities in excess of the carrying capacity of the range) was common on forestlands and became a major research and management concern in the 1920s (Platts, 1981). Overgrazing in forests was associated with de- creased infiltration capacity, increased overland flow and surface erosion, in- creased peak flows, and increased sedimentation in streams (Johnson, 1952; Dissmeyer, 1976; Anderson et al., 1976). In larger watersheds, overgrazing by domestic livestock is associated with ecological damage to thousands of linear miles of riparian forest cover and associated ecosystems, spurring policy state- ments by the American Fisheries Society (Armour et al., 1994). HYDROLOGIC RESPONSES: GENERAL PRINCIPLES Twelve general principles of forest hydrology (Table 3-1) describe the
FOREST DISTURBANCE AND MANAGEMENT EFFECTS ON HYDROLOGY 51 TABLE 3-1 General Principles of Forest Hydrology Describing the Direct Effects on Hydro- logic Processes of Changes in Forest Structure, Changes in Water Flowpaths, and Applica- tion of Chemicals Principles of Hydrologic Response to Changes in Forest Structure 1 Partial or complete removal of the forest canopy decreases interception and in- creases net precipitation arriving at the soil surface 2 Partial or complete removal of the forest canopy reduces transpiration 3 Reductions in interception and transpiration increase soil moisture, water avail- ability to plants, and water yield 4 Increased soil moisture and loss of root strength reduce slope stability 5 Increases in water yield after forest harvesting are transitory and decrease over time as forests regrow 6 When young forests with higher annual transpiration losses replace older forests with lower transpiration losses, this change results in reduced water yield as the new forest grows to maturity Changes in Water Flowpaths in Soils and Subsoils 7 Impervious surfaces (roads and trails) and altered hillslope contours (cutslopes and fillslopes) modify water flowpaths, increase overland flow, and deliver over- land flow directly to stream channels 8 Impervious surfaces increase surface erosion. 9 Altered hillslope contours and modified water flowpaths along roads increase mass wasting Hydrologic Response to Application of Chemicals 10 Forest chemicals can adversely affect aquatic ecosystems especially if they are applied directly to water bodies or wet soils 11 Forest chemicals (fertilizers, herbicides, insecticides, fire retardants) affect water quality based on the type of chemical, its toxicity, rates of movement, and persis- tence in soil and water 12 Chronic applications of chemicals through atmospheric deposition of nitrogen and sulfur acidify forest soils, deplete soil nutrients, adversely affect forest health, and degrade water quality, with potentially toxic effects on aquatic organisms NOTE: These general principles are not predictions, so qualifying adjectives such as âmay,â âusually,â etc., are omitted. direct effects or first-order responses to changes in forest structure, changes in water flowpaths in soil and subsoil, and application of chemicals. These princi- ples tie together the storage and movement of water in forests, how disturbance and management modify water storage and movement within forests, and how these internal changes are translated into changes in watershed outputs (Figure 3-1). These principles embody the state of knowledge of forest hydrology based on process, plot, and watershed studies conducted mostly in the second half of the twentieth century. HYDROLOGIC RESPONSES WITHIN FORESTS Forest disturbances and management affect the pathways of water within the forest system. Interception, evapotranspiration, infiltration, and overland (or surface) flow respond to forest disturbance and management (Figure 3-1). In turn, these changes affect watershed outputs.
52 HYDROLOGIC EFFECTS OF A CHANGING FOREST LANDSCAPE Interception and Evapotranspiration Interception is the net loss of precipitation, by evaporation, between the top of the forest canopy and the forest floor; this water is returned to the atmosphere and does not enter the soil. When forest canopies temporarily capture raindrops or suspend ice and snow, they slow the rate at which precipitation arrives at the forest floor. If this captured moisture evaporates, it effectively decreases the amount of precipitation available for soil moisture storage, transpiration, or run- off. In dispersing raindrops or suspending ice or snow, interception slows the rate at which precipitation hits the forest floor and, in doing so, effectively de- creases the net effect of precipitation. Removal of trees reduces leaf area and hence, interception. Reductions in leaf areaâfrom fire, harvest, insects, or dis- easeâand differences in leaf area among different forest types and ages all af- fect hydrology in the same way (Verry, 1976; Schmid et al., 1991): a reduction in interception increases the amount of water that reaches the mineral soil. If infiltration rates are not changed, an increase in net precipitation increases soil moisture, water availability to plants, and the proportion of precipitation that is available for streamflow (Helvey and Patrick, 1965; Helvey, 1971). Reduced leaf area decreases interception rates in both rain- and snow-dominated sys- tems; in snow-dominated systems an increase in net precipitation increases wa- ter stored in the snowpack (Neary and Ffolliott, 2005; Woods et al., 2006). Where forest canopies capture additional moisture from clouds, a reduction in leaf area can decrease net precipitation (Harr, 1982; Hutley et al., 1997; Reid and Lewis, 2007). A reduction in leaf area also increases the amount of light reaching the for- est floor, increasing energy exchange between soil or snow and the atmosphere and altering the energy budget (Figure 1-3). Increased exposure of the snow- pack to solar radiation and to turbulent heat transfer by wind increases snowmelt rates relative to undisturbed forest canopies. In snow-dominated forest systems a reduction in leaf area can lead to increased snow accumulation as well as an earlier onset of snowmelt and faster melt rates (Helvey, 1980; Megahan, 1983; Hornbeck et al., 1997; Jones and Post, 2004). The process of transferring moisture from the earth to the atmosphere by evaporation of water and transpiration from plants is called evapotranspiration. In North American forests, evapotranspiration accounts for 40 to more than 85 percent of gross precipitation. A reduction in leaf area from forest harvest, fire, or insect or disease outbreaks reduces evapotranspiration and increases water available for runoff. The magnitude and persistence of the reduction in transpira- tion depends on the amount and type of the vegetative canopy removed and the rate at which the vegetative cover is reestablished. However, it has only re- cently become possible to accurately measure transpiration in trees, and few studies have quantified transpiration rates for forest stands (but see Ryan et al., 2000; Moore et al., 2004).
FOREST DISTURBANCE AND MANAGEMENT EFFECTS ON HYDROLOGY 53 Infiltration and Overland Flow Most forests have an organic surface layer that protects the soil surface and facilitates infiltration. In most cases this water moves by subsurface pathways to the stream. Because forest soils have high infiltration rates, water rarely flows over the ground surface as infiltration excess (also called Horton overland flow). In flatter, low-lying or convergent zones, the saturated zone may rise to the sur- face and produce saturated overland flow. Forest management activities and forest disturbances may remove or alter the surface layers of forest soils, and thereby reduce infiltration and increase Horton overland flow (Figure 1-1). Forest management activities and distur- bances also create impervious surfaces (e.g., as roads) and modify hillslopes in ways that alter water flowpaths in soils and subsoils, shift subsurface flow to surface flow, and increase runoff and erosion rates. When organic surface layers are removed or burned, underlying mineral soil is exposed to raindrop splash and fine soil particles can accumulate on the surface, reducing infiltration and increasing overland flow. If soils are compacted to the extent that infiltration rates are lower than rainfall or snowmelt rates, the resulting overland flow can greatly increase runoff rates and surface erosion. CHANGES IN WATERSHED OUTPUTS Forest hydrology science describes direct changes in watershed outputs re- sulting from fire, timber harvest, and roads and trails (Table 3-2). These find- ings are summarized below. Fire Fire, Infiltration, and Overland Flow Burning can greatly reduce infiltration rates and thereby increase surface runoff and erosion rates through several mechanisms: development of a water- repellent (âhydrophobicâ) layer at or near the soil surface; exposure of the soil surface to raindrop impact and soil sealing; increased soil erodibility; and de- creased surface roughness (Box 3-2). Large post-fire increases in runoff and erosion are often attributed to an increase in soil water repellency after burning (Box 3-2), but soil sealing may play an equal or even larger role in increased runoff and erosion in some areas.
54 HYDROLOGIC EFFECTS OF A CHANGING FOREST LANDSCAPE TABLE 3-2 Magnitude and Duration of Direct Effects on Watershed Outputs of Three Sets of Processes That Modify Hydrology in Forests: Fire, Forest Harvest and Silviculture, and Roads and Trails Processes That Modify Hydrology in Forests Forest Harvest and Watershed Output Fire Silviculture Roads and Trails Water yield High-severity fire Increase water yield; Little or no effect increases annual magnitude and dura- water yields; little tion of response effect of low-severity varies (see text) fire Peak flows High-severity fire Increase peak flows; Increase peak increases peak magnitude and dura- flows; effects may flows; effect is short- tion of response be long-lived and lived varies (see text) affect extreme events Low flows High-severity fire Increase low flows in Little or no effect increases low flows; short term; deficits little effect of low- may develop as severity fire forests regrow Erosion, land- High-severity fire Increase surface Increase surface slides, sedimenta- increases erosion erosion, landslides, erosion (road sur- tion and sedimentation in and sedimentation; faces and gullies streams; less effect effects may be long- below culverts) and from low-severity lived landslides; increase and prescribed fire sedimentation in streams Water temperature Increases water Increase water tem- Deliver road and chemistry temperature due to perature due to ripar- chemicals (e.g. riparian forest re- ian forest removal; salt, oil) to streams moval; fire retar- effects of fertilizer dants and ash affect mostly small and chemistry; effects short-lived; short- are short-lived lived post-harvest increases in nitrate Research gaps Uncertainty about Uncertainty about Uncertainty about effects beyond a few effects beyond one road effects on years; magnitude or two decades; extreme floods and and persistence of magnitude and per- in watersheds >1 downstream effects; sistence of down- km2 effects of salvage stream effects; ef- logging fects on habitat and aquatic ecosystems NOTE: These are general effects, not predictions, so qualifying adjectives such as âmay,â âusually,â etc., are omitted. See text for factors that influence when, where, and to what extent these effects apply.
FOREST DISTURBANCE AND MANAGEMENT EFFECTS ON HYDROLOGY 55 BOX 3-2 Physical and Chemical Causes of Water Repellency in Soils Many soils are water repellent without being burned, particularly in coniferous forests and xeric shrublands. Waxy and other aromatic compounds in the foliage of these vegeta- tion types leach out and accumulate on the soil surface. Fungal hyphae also can generate very strong, localized soil water repellency near soil surfaces. In the absence of burning, these compounds are rarely sufficient to reduce infiltration rates at the hillslope scale. However, when burned at roughly 175-200ÂºC, these compounds vaporize and are driven by steep heat gradients down into the soil, where they condense on cooler underlying soil particles. Thus, burning can create a semicontinuous or continuous water repellent layer at or beneath the soil surface, whose depth and thickness depend on the duration and magni- tude of soil heating (DeBano, 2000; Letey, 2001). Temperatures above 280-400ÂºC con- sume most waxy and aromatic compounds, so very hot fires produce a nonrepellent, dis- aggregated soil layer above a water-repellent layer (DeBano, 2000; Doerr et al., 2006). Coarse-textured soils are more susceptible to the formation of a water-repellent layer than fine-textured soils because of their lower surface area and greater air permeability (Huff- man et al., 2001; DeBano et al., 2005). Fire-induced soil water repellency has been well documented for certain vegetation types, particularly coniferous forests and chaparral-type ecosystems (e.g., DeBano, 2000). Fire-induced soil water repellency is spatially heteroge- neous (Woods et al., 2007) and can persist for a few weeks or several years (Shakesby and Doerr, 2006). Snowmelt or prolonged rainfall may overcome water repellency until soils dry out (Doerr and Thomas, 2000; MacDonald and Huffman, 2004). Thus, burning may have less effect on infiltration and runoff during winter wet seasons or in snowmelt-dominated areas than in drier areas subjected to summer thunderstorms. Fire-induced soil water repellency breaks down by a combination of physical, chemical, and biological processes over time as plant regrowth provides a protective cover of vegetation and litter (e.g., Robichaud and Brown, 1999; Benavides-Solorio and MacDonald, 2005). There- fore, runoff and erosion rates usually return to reference or pre-fire levels within one to four years (Shakesby and Doerr, 2006). Soil sealing refers to a reduction in infiltration as a result of breakdown of soil aggregates and rearrangement of soil particles at the surface. After moder- ate- and high-severity fires, rain splash can detach soil particles and reduce infil- tration rates by a sealing effect at the soil surface. Combustion and the loss of organic matter also lead to a loss of soil cohesion. These effects contribute to greater overland flow (Neary et al., 1999; DeBano et al., 2005; Moody et al., 2005) and soil erodibility (Cerda, 1998; Larsen and MacDonald, 2007; Woods and Balfour, 2007). Severe fire leads to greater reductions in infiltration and greater increases in overland flow than moderate- or low-severity fire (Shakesby and Doerr, 2006). In the Colorado Front Range, for example, summer thunderstorms with 60 mm of rainfall per hour often produce no surface runoff or erosion, but after a high- severity fire, surface rainfall intensities of only 10 mm per hour can generate overland flow (Moody and Martin, 2001; Kunze and Stednick, 2006; Wagen-
56 HYDROLOGIC EFFECTS OF A CHANGING FOREST LANDSCAPE brenner et al., 2006). Compared to high-severity fires, low-severity wildfires and most prescribed fires result in little or no exposure of the mineral soil sur- face, smaller changes in soil water repellency (Robichaud, 2000), and little ef- fect on overland flow (Van Lear and Danielovich, 1988; Robichaud, 2000; Benavides-Solorio and MacDonald, 2005). Salvage logging is often conducted post-fire, and although its effects on forest ecosystems are being debated (Donato et al., 2006), few studies have ex- amined the hydrologic effects. Ground-based salvage logging generally results in more ground disturbance and less ground cover (Klock, 1975; McIver and McNeil, 2006), and extensive ground disturbance can increase soil erodibility and erosion rates. Non-ground based activities, such as the use of helicopters in logging, result in relatively little ground disturbance and may have minimal ef- fect on post-fire runoff and erosion rates. Standard salvage logging practices are unlikely to significantly reduce or break up the water-repellent layer from a high-severity wildfire. Fire and Water Yield High-severity fires occur in unpredictable locations, at unpredictable times (Carpenter, 1998), which makes their study difficult and has resulted in few studies that confirm relationships between fire and water yield. By analogy to clear-cutting, high-severity fires are expected to increase water yield. Changes in interception, transpiration, and runoff processes usually lead to higher annual water yields after a high-severity fire, although the increases are highly variable between and within ecoregions (Berndt, 1971; Campbell et al., 1977; Helvey, 1980; Neary and Ffolliott, 2005; Neary et al., 2005b). Low-severity fires gener- ally do not consume or kill enough vegetation to alter water yields significantly (e.g., Douglass and Van Lear, 1983; Gottfried and DeBano, 1990). Fire and Peak Flows High-severity fires can increase peak flows by one or two orders of magni- tude (Scott, 1993; Moody and Martin, 2001; Neary et al., 2005b). The greatest increases in peak flows occur in areas with summer thunderstorms or fall rains, where burning has altered infiltration and overland flow processes. Reported changes in peak flows after wildfires include (1) minimal change following a high-severity burn in a snow-dominated Wyoming fir forest (Troendle and Bevenger, 1996): increase of 1.4 times in a Douglas fir forest in Oregon (Ander- son, 1974); (2) increase of 6.5 to 870 times in California chaparral (Hoyt and Troxell, 1934; Sinclair and Hamilton, 1955; Krammes and Rice, 1963); and (3) increase of 20 to more than 2,000 times in ponderosa pine forests in Arizona and New Mexico (Campbell et al., 1977; Bolin and Ward, 1987; Ffolliott and Neary, 2003).
FOREST DISTURBANCE AND MANAGEMENT EFFECTS ON HYDROLOGY 57 Fire and Erosion The most important effects of fire are increases in overland flow and ero- sion and resulting effects on flooding and water quality. Fire can enormously increase surface erosion. Fire exposes the mineral soil and increases surface ero- sion; it also may increase soil moisture and landslides (Wells et al., 1979; Moody et al., 2005; Neary et al., 2005; Shakesby and Doerr, 2006). Fire-induced higher rates of erosion increase sediment delivery to streams (Helvey, 1980; Ewing, 1996; Moody and Martin, 2001; Ffolliott and Neary, 2003; Wondzell and King, 2003; Libohova, 2004; Kunze and Stednick, 2006). Low-severity and prescribed fires produce smaller effects on erosion and sedimentation in streams (Douglas and Van Lear, 1983; Van Lear et al., 1985; Van Lear and Danielovitch, 1988; Gottfried and DeBano, 1990; Wright et al., 1982). Post-fire salvage logging can further increase erosion and sediment delivery to streams (McIver and McNeil, 2006). Studies of erosion and sedimentation after high- severity fires have become more frequent in the past few decades as wildfire activity has increased on forestland in the western United States (Westerling et al., 2006). Fire Effects on Water Temperature and Chemistry Fire can affect a series of water quality parameters (see recent summaries by Landsberg and Tiedeman, 2000; Ranalli, 2004; Neary et al., 2005b). The effects of fire depend in large part on the pre-fire composition of organic matter and the fire intensity. The chemistry of unburned organic matter varies with plant spe- cies, underlying geology, time elapsed since the last disturbance, and atmos- pheric deposition of elements such as mercury and lead. Fires usually affect water quality by the indirect pathway of increasing stream water temperature and two direct pathways, atmospheric deposition and surface runoff. Extensive burning of the riparian forest canopy removes shade, increases the amount of solar radiation, and raises stream water temperatures. Increased organic carbon and temperature in streams can reduce concentrations of dissolved oxygen (Neary et al., 2005c). During a fire, gases and particulate matter are carried aloft and transported for varying distances before being deposited on water surfaces. In the Yellow- stone fires of 1988, for example, increases in nitrogen in lakes and rivers were attributed to the diffusion of smoke into the water bodies under active fire condi- tions (Spencer et al., 2003). Ash deposition can increase the pH of surface water and soil (Neary et al., 2005b). Post-fire pH values in stream water rarely exceed U.S. Environmental Protection Agency (EPA) standards (Landsberg and Tiedemann, 2000), but tran- sient pH values of 9.5 were measured in streams after a fire in eastern Washing- ton (Tiedemann, 1973; Tiedemann et al., 1979). Fire can cause a short-term in- crease in stream nitrate concentrations, and the delivery of ash and fine sediment
58 HYDROLOGIC EFFECTS OF A CHANGING FOREST LANDSCAPE can increase phosphorus concentrations in streams. In most cases these in- creases do not exceed standards for drinking water (Neary et al., 2005c). During forest fires, chemical fire retardants are applied aerially to forests and inadvertently (perhaps unavoidably) to streams and rivers. The effects of these chemicals on water quality may be important, especially since recent stud- ies have shown that they persist for years after application (Morgenstern, 2006). Fire retardants can contain nitrate and possibly sulfate, phosphate, and some trace elements (Landsberg and Tiedemann, 2000), which can contribute to eu- trophication, especially when applied directly to streams. When these materials enter rivers, streams, and lakes, they react with sunlight to form compounds that are toxic to aquatic organisms (e.g., Buhl and Hamilton, 1998, 2000). Increased concentrations of other chemicals, such as manganese, sulfate, and mercury, also have been documented after forest fires. Elevated concentrations of both lead and mercury were detected in the post-fire runoff from the Bobcat fire outside Fort Collins, Colorado. Elevated post-fire concentrations of manganese and other constituents forced the Denver Water Board to initiate additional special- ized treatments to maintain drinking water quality. In most cases the adverse effects of forest fires on chemical water quality persist for no more than two or three years. Forest Harvest Timber Management, Silviculture, and Water Yield Dozens of paired watershed forest harvest experiments have demonstrated that forest removal increases water yield (Bosch and Hewlett, 1982; Hornbeck et al., 1993; Ice and Stednick, 2004; Jones and Post, 2004; Brown et al., 2005). The magnitude of water yield increases can be expressed as an absolute increase (e.g., millimeters of water, Figure 3-2) or as a percentage. Water yield increases are highest after 100 percent forest harvest, and are almost always highest in the first year after forest harvest, or the wettest year in the early post-harvest period, when changes in interception and transpiration have the greatest effect on the water balance (Bosch and Hewlett, 1982; Sahin and Hall, 1996). Water yield increases after forest harvest vary according to several factors: â Climate. The largest absolute water yield increases have occurred after cutting of forests in climates with relatively abundant precipitation (1,500 to 2,500 mm per year) and relatively low evapotranspiration. High water yield increases (300 to 500 mm per year) have been measured in the Pacific North- west (H.J. Andrews Experimental Forest, site 1), the Northeast (Hubbard Brook Experimental Forest, site 14), and the Southeast (Coweeta Experimental Forest, site 18) (Figure 3-2). Much smaller water yield increases have been measured in regions where mean annual precipitation is low (<500 mm per year) and
FOREST DISTURBANCE AND MANAGEMENT EFFECTS ON HYDROLOGY 59 (a)
60 HYDROLOGIC EFFECTS OF A CHANGING FOREST LANDSCAPE (b) (c) FIGURE 3-2 (a) Potential annual water yield (precipitation minus potential evapotranspira- tion, P â PET) in millimeters for the United States, and mean monthly precipitation, tem- perature, and discharge at four small watershed sites in the Northwest, Northeast, Rocky Mountains, and Southeast. Mean elevation (meters), mean annual precipitation (millime- ters), and mean annual temperature (degrees celsius) are shown above the graph for each of these four sites. (b) First year water yield increase (millimeters) versus percent basal area removal for 21 paired watershed forestry experiments in the conterminous United States. The trend line, shown for comparison, was taken from Figure 1 of by region were developed by Stednick (1996), which represented 95 catchments spread across similar geographic regions as shown here. (c) Locations of paired watershed sites show in (b). Black lines on map are USFS ecoregions.
FOREST DISTURBANCE AND MANAGEMENT EFFECTS ON HYDROLOGY 61 potential evapotranspiration is high. In these areas, such as the interior West and Southwest, removal of forest cover is largely offset by increased soil evapo- ration and evapotranspiration by any remaining vegetation, and water yield in- creases are often <150 mm per year (Bosch and Hewlett, 1982) (Figure 3-2). Small water yield responses in the Southwest illustrate the limited potential for increasing water yields in dry forest types. â Seasonal timing of precipitation. In regions where precipitation is evenly distributed throughout the year (Northeast, Southeast; Figure 3-2), water yield increases typically occur during the growing season (Martin et al., 2000; Jones and Post, 2004). In regions with dry summers and wet winters (western forests; Figure 3-2), the largest water yield increases occur in the late fall and early winter due to a reduction in transpiration and resultant increase in soil moisture carryover (Jones and Post, 2004). In snowmelt-dominated regions, most of the water yield increase occurs in spring because larger snowpacks ac- cumulate in cutover areas (Harr et al., 1979; Troendle and King, 1985; Troendle and Reuss, 1997; Jones and Post, 2004). Thus, in both eastern and western for- ests, water yield increases after forest harvest often occur during seasons when water is abundant, not scarce (Harr, 1983; Troendle et al., 2001). â Amount of forest removed. Forest harvest experimental treatments have ranged from 100 percent clear-cutting to partial cuts, overstory thinning, or selective harvest of a fraction of watershed area (Figure 3-2). In areas with more than 500 mm of mean annual precipitation (the Pacific Northwest, North- east, and Southeast), water yield increases are roughly proportional to the amount of forest area cut (Hibbert, 1967; Bosch and Hewlett, 1982) (Figure 3- 2). Water yield increases are difficult to detect when less than 20 percent of the basin area has been harvested (Stednick, 1996) (Figure 3-2). The spatial ar- rangement of cutting within a watershed also affects whether a water yield in- crease is detected; controlling for the amount of forest cut, there is less detect- able water yield increase for thinning or selective harvests than for patch cuts (Site 4, Figure 3-2) (Perry, 2007). â Harvest treatments, such as burning, herbicides, or buffer strips. For- est harvest has different effects on water yield increase depending on whether the area is burned, herbicide is applied, or the treatment is conducted in stages (compare vertical scatter of points from site 14 in Figure 3-2). â Storage of water in soil and snow. Year-to-year storage of water in deep soils or poorly drained areas may offset water yield increases in areas with deep compared to shallow soils (compare site 20 versus 18, and 16 versus 14 in Figure 3-2). Post-harvest changes in snow accumulation and melt rates also can affect water yield increases after harvest (Verry et al., 1983; Troendle and Reuss, 1997; Jones and Post, 2004) (compare vertical scatter of points from site 1 or site 14 versus 13 in Figure 3-2), â Type and age of forest removed. Post-harvest water yields are higher when old-growth forests with high leaf area are harvested, compared to when younger forests with low leaf area are cut (Swank and Crossley, 1988; Jones and Post, 2004). When forests of low interception (or lower annual transpiration
62 HYDROLOGIC EFFECTS OF A CHANGING FOREST LANDSCAPE losses), are replaced by forests with higher interception (higher transpiration losses) the net water yield can be reduced as the new forest grows to maturity (e.g., Swank and Crossley, 1988; Jones and Post, 2004). â Time since harvest, or the amount of forest regrowth. As forests re- generate after harvest, water yield increases disappear. Water yield increases have persisted for as little as a decade in some areas, but for multiple decades in other areas, depending on the type and history of the forest, soils, climate, refor- estation methods, and harvest treatments (Bosch and Hewlett, 1982; Troendle and King, 1985; Swank and Crossley, 1988; Hornbeck et al., 1993; Hornbeck et al., 1997; Troendle et al., 2001; Jones and Post, 2004; Brown et al., 2005). In some cases, water yields drop below pre-harvest levels after a couple of decades of forest regrowth (Hornbeck et al., 1993; Swank and Crossley, 1988; Jones and Post, 2004; Brown et al., 2005). Many paired watershed experiments established to test forest management effects on water yield were terminated after the first 5-10 years of post-treatment, so only a few paired watersheds are still providing information about the long-term consequences of past forest management for water yield. In summary, water yield increases from forest harvest are highly variable. The highest increases in water yield occur from 100 percent forest harvest in the first years after harvest and in areas where water is relatively abundant. Because of the inherent variability in water yield responses, the amount of forest harvest necessary to produce a water yield increase varies according to regional and site- specific characteristics (e.g., amount and type of precipitation, slope, soil thick- ness, silvicultural methods, harvest treatments). There is little evidence that timber harvest can produce sustained increases in water yield over large areas. Because of high evapotranspiration relative to precipitation, and dry summers, the potential for augmenting water yield on a sustainable basis in western forests and rangelands is very low (Harr, 1983; Hibbert, 1983; Troendle et al., 2001). Water yield increases from the harvest of western forests occur in winter when water is relatively abundant, and these in- creases would have to be stored for up to six months to effectively augment wa- ter supplies when water is scarce in late summer (Harr, 1983). Maintaining wa- ter yield increases requires continued forest harvest or conversion of forests to other land uses such as pastures, annual crops, and urban areas. Although the potential for augmenting water yield is higher in eastern than western forests, achieving this potential would require major changes in forest management ob- jectives and land use (Douglass, 1983). Timber Management, Silviculture, and Low Flows Relative to peak flows or annual water yields, few studies have examined the effects of forest harvest on low flows. Most studies show an initial increase in low flows immediately after forest harvest (Harr et al., 1979, 1982; Keppeler
FOREST DISTURBANCE AND MANAGEMENT EFFECTS ON HYDROLOGY 63 and Ziemer, 1990; Hicks et al., 1991; Hornbeck et al., 1997; Johnson, 1998; Swank et al., 2001; Jones and Post, 2004). Observed water increases in low flows after harvest change are often short-lived, usually persisting for less than 10 years due to the relatively rapid recovery of leaf area, interception capacity, and transpiration rates. These short-term surpluses during the low-flow period change to deficits as forests regrow. As in the case of annual water yields, the increase in low flows often is followed by a decrease in low flows to below pre-harvest levels (Hicks et al., 1991; Hornbeck et al., 1997; Swank et al., 2001). These decreases occur when a forest with relatively high transpiration and/or interception replaces a forest with relatively low transpiration or interception, such as during (1) species conversion (e.g., deciduous to evergreen) (Swank et al.,1988); (2) regeneration of a young stand with higher water use than the mature stand it replaces (Hicks et al., 1991; Perry, 2007); or (3) establishment of different riparian vegetation with greatest water demands (Moore et al., 2004; Ice and Stednick, 2004). Be- cause relatively few studies have examined long-term trends in low flows, there is much uncertainty about this subject. Timber Management, Silviculture, and Peak Flows Decreases in transpiration and interception after forest harvest increase soil moisture, and higher initial soil moisture at the beginning of a storm increases storm runoff (peak flow) (Box 3-3). Recent compilations of studies examining forest management effects of peak flows show wide variability in the magnitude of peak flow response to forest harvest (Austin, 1999; Moore and Wondzell, 2005; Grant et al., 2008). Much of this variation is explained by the differences in how peak flows were defined and analyzed, dominant hydrologic regimes (e.g., rain or snow), differences in forest management, and other differences in site conditions (Austin, 1999). Peak flow responses to forest harvest vary according to several factors: â Event size. Often, the percentage increase in peak flows after forest harvest decreases as the magnitude of the peak flow increases (Harr, 1976; Beschta et al., 2000; Grant et al., 2008). However, in many cases the absolute increase in peak flows is larger with larger storms (Box 3-4; Verry, 1986; Jones and Grant, 1996; Jones, 2000; Moore and Wondzell, 2005). As storm magni- tude (the total amount of rainfall or snowmelt) increases, the proportion of pre- cipitation that can be stored by vegetation decreases. Therefore, large peak flows often experience smaller relative increases than small peak flows. Nevertheless, small percentage increases in very large floods as a result of forest harvest (Fig- ure 3-3) may be quite large in absolute terms; a 10 percent increase in a typical 50-year flood is the same amount of water as a 50 percent increase in a 1-yr flood. Small increases in extreme floods affect more people and may be of greater concern for managers than increases in small floods.
64 HYDROLOGIC EFFECTS OF A CHANGING FOREST LANDSCAPE BOX 3-3 Does Timber Harvesting Cause Floods? The effect of forest management on flooding has been a recurrent scientific, social, and political theme (Eisenbies et al., 2007). The notion that deforestation leads to wide- spread land degradation and exacerbates the risk of flooding dates to antiquity (Hillel, 1994), and the role of forest management on extreme floods is an important concern for policy makers and the public (Figure 3-3). There is little doubt that forests influence the storage and movement of water, particularly at annual and seasonal time scales. Under- standing the role of forest management in moderate and large floods requires a clear defi- nition of terms and careful consideration of the various processes by which forest man- agement can affect the size of peak flows. Floods are variously defined by scientists and affected populations. Floods are com- monly described as flows that exceed channel capacity and result in overbank inundation (Brooks et al., 2003), which can occur as frequently as every one to two years, since chan- nels tend to adjust their shape to accommodate more frequently occurring events (Leopold et al., 1995). Hydrologists typically define floods according to their probabilities of recur- rence or return period (e.g., the 5-year flood, the 100-year flood). Public concerns about floods are commonly limited to the more extreme events that result in a loss of life or prop- erty. The largest floods are associated with extreme storm events, such as tropical cy- clones. Some recent assessments have attempted to link the growth in flood damage in recent decades to development in flood-prone lands and to discount the role of deforesta- tion on large-scale extreme floods (FAO, CIFOR, 2005). Nevertheless, considerable pub- lic and political pressure tends to follow extreme floods, and forest protection is often an important element of debates and policy formulation (Eisenbies, 2007). FIGURE 3-3 Public perception of flooding is often linked to land use activities. SOURCE: Reprinted, with permission, from Associated Press (2002). Copyright 2002 by Associated Press.
FOREST DISTURBANCE AND MANAGEMENT EFFECTS ON HYDROLOGY 65 BOX 3-4 Regional Variability in Peak Flow Response: Northern Lake States The effect of timber harvesting (or forest clearing for other uses) on the timing of peak flows varies regionally, as illustrated in the northern Lake States, particularly where snow- melt dominates the annual hydrograph. In mountainous areas, snowmelt is naturally de- synchronized by the heterogeneity of terrain features such as slope, aspect, and elevation. The snowpack generally melts first at low elevations and last at high elevations, more quickly on south slopes than on east and west slopes, and most slowly on north slopes. Differences in forest canopy type and density (e.g., dense, even-aged conifers versus sparse, deciduous stands) also change the timing and rate of snowmelt. In the relatively flat terrain of the northern Lake States the effect of topography on energy exchange is minor in comparison to the influence of forest cover. Therefore, when timber harvesting (e.g., patch or strip cuts) changes the energy balance and microclimate of a watershed, it tends to de- synchronize snowmelt (Figure 3-4). In effect, forest harvest changes the snowmelt hydro- graph from one peak to two, with the first peak coming from the harvested openings and the second peak from the remaining mature forest. As the proportion of the harvested area within a watershed increases (e.g., greater than ~70 percent open), snowmelt occurs ear- lier and at a more rapid rate. FIGURE 3-4 First-year increase in peak discharge versus percent forest clearing for wa- tersheds in the northern Lake States. SOURCE: Reprinted, with permission, from Ice and Stednick (2004). Copyright 2004 by the Society of American Foresters.
66 HYDROLOGIC EFFECTS OF A CHANGING FOREST LANDSCAPE â Rain versus snow. Storm events involving rain respond differently to forest harvest than those involving snow. In rain events, forest harvest affects peak flows directly through changes in soil water. In events involving snow, the effect of forest harvest on peak flows depends on how forest harvest changes snowpack size and snowmelt, as well as soil moisture (Verry et al., 1983; Tro- endle and King, 1985; Jones, 2000; MacDonald et al., 2003; MacDonald and Stednick, 2003; Schnorbus and Alila, 2004). â Season. Peak flow increases after forest harvest are proportionately larger in spring, summer, and fall compared to winter, because soil moisture levels are sensitive to transpiration by forest vegetation in spring, summer, and fall. However, peak flow increases (in millimeters) after forest harvest are abso- lutely (see Figure 3-2) larger in winter than in spring, summer, or fall, because peak flows are higher in winter, and reductions in transpiration in previous sea- sons carry over into winter (Jones, 2000; MacDonald and Stednick, 2003). â Proportion of area harvested. The larger the proportion of area har- vested, the greater is the increase in peak flows (Jones, 2000; Moore and Wondzell, 2005). Peak flow increases have been detected after only 25 percent harvest of a small watershed (Harr et al., 1979, 1988; Jones and Grant, 1996; Caissie et al., 2002). â Topographic relief and elevation. The effect of forest harvest on the energy balance, and resulting changes in snow accumulation and melt, vary with elevation and aspect. Forest harvest may increase peak flows during rain-on- snow events in the Pacific Northwest (Harr, 1981; Harr, 1986). In the flatter topography of the northern Lake States, harvesting 20 â 50 percent of the water- shed desynchronizes snowmelt and reduces annual snowmelt peak flows by as much as 40 percent, while harvesting over 60 percent of a basin can increase the size of snowmelt peak flows by more than 140 percent (Box 3-4 and Figure 3-4) (Verry, 1986). â Time since harvest. As forests regenerate, peak flows return to pre- harvest levels (Troendle and King, 1985; Jones, 2000). â Roads and skid trails. Many studies of forest harvest effects on peak flows include some roads and skid trails, which can accentuate the effect of har- vest on peak flows (Jones and Grant, 1996). Because roads and trails influence different components of the water balance, they are discussed separately below. Timber Management, Silviculture, and Erosion, Mass Movement, and Sedimentation Many studies have shown that timber harvest practices greatly increase sur- face erosion (summarized in Dunne and Leopold, 1978; Brooks et al., 2003). Overland flow and surface erosion are rare in undisturbed forests, but logging operations expose surface soils and lead to surface erosion. Multiple studies have shown that surface erosion is most significant in areas of soil disturbed by cable yarding and skidding of cut logs to landings (e.g., Johnson and Beschta,
FOREST DISTURBANCE AND MANAGEMENT EFFECTS ON HYDROLOGY 67 1980). Many forestry regulations govern surface erosion and sediment produc- tion. The effects of skid trails and unpaved roads on surface erosion are de- scribed below in the discussion of roads and trails. In steep landscapes, extreme storms trigger landslide events and the associ- ated input and transport of bedload and woody debris (MacDonald and Coe, 2007). Portions of the Pacific Northwest, northern Rocky Mountains, and central Appalachians are especially prone to shallow landslides. After forest harvest on steep slopes, decreasing root strength and increased soil moisture and pore water pressures contribute to decreased slope stability and can increase the likelihood of shallow landslides (debris avalanches) during precipitation events. Higher soil moisture (from reduced interception and transpiration) increases the forces generating slope movement. Higher soil moisture also increases pore pressures and reduces soil cohesion; combined with loss of root strength after harvesting, these factors reduce the forces resisting slope movement (Swanson and Dyrness, 1975; Sidle et al., 1985; Montgomery et al., 2000; Miller and Burnett, 2007). A number of studies have documented increased landslides from forest harvest relative to undisturbed forested areas (Swanson and Dyrness, 1975; Sidle and Ochiai, 2006). Forest clearcutting may increase the landslide erosion rate by two to nine times relative to undisturbed areas (Montgomery et al., 2000; Sidle and Ochiai, 2006; Miller and Burnett, 2007). Sediment and woody debris deliv- ered to stream channels by landslides can be transported downstream by debris flows, which may create debris dams and exacerbate flooding in lowland set- tings, causing extensive property damage. Steep forestlands also are prone to deep slow-moving earthflows. During large storms, these earthflows may con- tribute material to streams, including fine clays, which may create persistent turbidity in downstream reservoirs and water supply systems (Bates et al., 1998). Many studies have shown that fine sediment contributed to streams by sur- face erosion from exposed soils or landslides can greatly increase suspended sediment in streams, adversely affecting aquatic habitat, especially in steep, coarse-bedded streams (e.g., Campbell and Doeg, 1989). Suspended sediment levels may remain elevated for many years or decades after timber harvest (Grant and Wolff, 1991). Timber Management, Silviculture, Water Temperature, and Chemistry Removal of riparian vegetation along streams causes peak water tempera- tures to increase as a result of increased solar radiation. The largest stream tem- perature increases occur in the summer (Levno and Rothacher 1967, 1969; Beschta and Taylor, 1988; Binkley and Brown, 1993a; Johnson and Jones, 2000). As streamside forests regenerate and provide shade, temperatures usually return to pretreatment levels (Johnson and Jones, 2000; Ice et al., 2004). Main- taining streamside forests of sufficient width to shade streams helps mitigate temperature increases from forest harvest (Stednick, 2000; MacDonald and Coe, 2007), but groundwater contributions and hyporheic flow also mitigate stream
68 HYDROLOGIC EFFECTS OF A CHANGING FOREST LANDSCAPE temperature increases (Story et al., 2003; Johnson, 2004). In silviculture, fertilizers, herbicides, fungicides, insecticides, and fire retar- dants are used to protect and enhance tree growth. Plant available nitrogen and phosphorus are occasionally added to forests as fertilizer if one or both of these nutrients are limiting growth. Generally, increases in stream concentrations after fertilization are minimal because forest soils efficiently retain nutrients, and because fertilizer that is not absorbed is often volatilized (Binkley et al., 1999; NCASI, 1999; Stednick, 2000). If fertilization increases stream nitrogen concentrations, the maximum value is usually reached within two to four days, and concentrations decrease rapidly thereafter (although a return to pre- treatment conditions could take six to eight weeks). Greater effects on stream nutrient concentrations occur when fertilizer is applied directly to the stream, applied during rainy periods (Stednick, 2000), or applied to sites already af- fected by nitrogen pollution from non-fertilizer sources (Fenn et al., 1998). Relatively few studies have addressed the effects of fertilizer at large watershed scales or over the long term (Anderson 2002; McBroom et al., 2008). When pesticides are applied according forestry regulations, they are unlikely to impair water quality (Michael, 2000; Dent and Robben, 2000 [from Ice et al., 2004]; Tatum, 2003, 2004). If pesticides enter streams, they are usu- ally at low concentrations and only remain for a short period of time (Ice et al., 2004). Nevertheless, members of the public, nongovernmental organizations (NGOs), and the scientific community remain concerned that risk assessments are incomplete and that manufacturers funded most studies. Concerns focus on (1) chemicals that have yet to be investigated; (2) synergistic, cumulative effects of mixtures in the environment; (3) effects of degradation products in the long term through transformation and transport in groundwater (Michael, 2000); and (4) inadequate tests of how stream ecosystems might react because native am- phibians may be more sensitive than laboratory animals (Tatum, 2003). Most forests are naturally nitrogen-limited, and stream concentrations of ni- trogen are lowest in young forests and increase as forests mature (Edwards and Helvey, 1991; Swank and Vose, 1997; Vitousek, 1997). Decades of elevated atmospheric deposition of nitrogen (and sulfur, in the eastern United States) have greatly altered the nitrogen dynamics of eastern forests (Likens et al., 1977, 1996; Lovett and Kinsman, 1990; Likens and Bormann, 1995; Fenn et al., 1998) and increasingly, western forests (Riggan et al., 1985; Binkley, 2001). Resulting nitrogen saturation of soils and streams (Aber, 1992), as well as soil acidifica- tion and loss of basic cations, has affected forest growth and tolerance to cold stress, elevated stream nitrogen concentrations, and acidified streams, especially in the eastern United States (Adams et al., 1993; Lovett and Lindberg, 1993; Flum and Nodvin, 1995; Rustad et al., 1996; Lawrence et al., 1997; De Hayes et al., 1999; Lawrence and Huntington, 1999; Lovett et al., 2000; Baldigo and Lawrence, 2001). Riparian forest buffers have been adopted to mitigate these effects (Lowrance et al., 1997). After forest harvest, concentrations of nitrate-nitrogen (nitrate-N), one of the most mobile nutrients in disturbed forests, typically increase in streams as
FOREST DISTURBANCE AND MANAGEMENT EFFECTS ON HYDROLOGY 69 soil moisture content and subsurface flow rates increase (Likens et al., 1970; Miller and Newton, 1983; Martin and Harr, 1989; Binkley and Brown, 1993; Martin et al., 2000). Elevated stream nitrogen concentrations return to pre- harvest levels within a few growing seasons or less as young forests grow on a harvested site (Likens et al., 1970, 1977; Martin and Harr, 1989; Likens and Bormann, 1995; Martin et al., 2000; Swank et al., 2001). The duration of ele- vated post-harvest nitrogen in streams varies among sites according to soil ni- trogen levels and microbial activity, atmospheric deposition, and forest re- growth (Swank, 2000). A review of more than 40 studies found that nitrate-N on averaged doubled in concentration to 0.44 mg/L for one to five years after har- vest on 75 percent of the locations, but the remaining studies could not detect increases and a few (5) even had 24-95 percent declines (Binkley et al., 2004). Elevated peak flows and surface erosion after forest harvest may increase phosphorus delivery to streams. Phosphorus export occurs most during storm- flow events that mobilize the fine particulate matter to which phosphorus is sorbed (Hobbie and Likens, 1973; Meyer and Likens, 1979). Riparian forest buffers have been successful in reducing particulate phosphorus (phosphorus attached to sediment) by minimizing overland flow but less effective in remov- ing dissolved phosphorus (de la CrÃ©taz and Barten, 2007; Stednick, 2000). Roads Roads, Trails, Infiltration, and Overland Flow Roads and skid trails modify surface and subsurface flowpaths of water (Figure 3-5). Forest roads alter runoff processes in two ways (Figure 3-5) (Megahan, 1972; Wemple and Jones 2003). First, roads and skid trails (com- pacted soil surfaces) generate overland flow because they have very low infiltra- tion rates (Johnson and Beschta, 1980; Luce and Cundy, 1994; Ziegler and Giambelluca, 1997). Cutslopes above roads and hillslopes below roads also often have lower infiltration rates than forest soils, and they also may generate overland flow. Roads constructed on steep slopes also can intercept water flow- ing in the subsurface, further increasing overland flow (Megahan, 1972; Wem- ple and Jones, 2003). During precipitation or snowmelt events, water flows on road surfaces or in ditches that are connected to streams, so the road network delivers this water directly to the stream network. Roads also increase overland flow by intercepting water flowing in subsur- face flowpaths at cutbanks on steep hillslopes (Figure 3-5) and converting this water to surface flow, which the road network then delivers to streams (Wemple
70 HYDROLOGIC EFFECTS OF A CHANGING FOREST LANDSCAPE 1 2 3 4 FIGURE 3-5 Schematic diagram of how roads modify water flowpaths in soils and sub- soils. Roads generate Horton overland flow and intercept subsurface flow: 1 = shallow subsurface flow in the hillslope; 2 = subsurface flow intercepted by the roadcut; 3 = water flowing in roadside ditch; 4 = Horton overland flow on impervious road surface. Adapted from Wemple et al. (1996). and Jones, 2003). The greater the depth of the soil profile exposed by the cut- bank, the greater is the potential for subsurface flow interception (see photo be- low). With these two effects, roads convert relatively slow moving subsurface flow to overland flow, which has higher flow velocities. Roads, Trails, and Water Yield Roads have their greatest impact on hydrologic effects on infiltration and overland flow (see preceding section). Neither in empirical studies nor in mod- eling studies have roads been identified as having significant effects on annual water yields in forested systems.
FOREST DISTURBANCE AND MANAGEMENT EFFECTS ON HYDROLOGY 71 An unpaved forest road in steep terrain with a large cutslope. Photo courtesy of B. Wemple. Roads, Trails, and Peak Flows Roads redistribute water locally and alter flow routing. Roads can contrib- ute to an increase in the size of peak flows by increasing the amount of surface runoff from impervious surfaces, intercepting subsurface stormflow, and speed- ing the delivery of this runoff to the stream network through ditches or gullies (Megahan, 1972; Wemple et al., 1996; Wemple and Jones, 2003). The percent- age of unpaved roads that are connected to the stream network is directly pro- portional to mean annual precipitation and decreases with the presence of engi- neered road drainage structures such as waterbars, rolling dips, and relief cul- verts (Coe, 2006). Through the combination of increasing the amount of surface runoff and delivering this runoff more rapidly to the stream channel, roads can produce detectable changes in peak flows at the small watershed scale (Harr et al., 1975; Ziemer, 1981; King and Tennyson, 1984; Wright et al., 1990, La- Marche and Lettenmeier, 2001). Modeling studies have replicated measured road effects on peak flows (Bowling and Lettenmeier, 2001; LaMarche and Let- tenmeier, 2001; Tague and Band, 2001; Cuo et al., 2006).
72 HYDROLOGIC EFFECTS OF A CHANGING FOREST LANDSCAPE It is debated how much roads affect very large peak flows in small or large watersheds (Jones and Grant, 1996; Megahan and Thomas, 1998; Beschta et al., 2000; Jones, 2000; Jones and Grant, 2001a, 2001b). A small number of studies designed to test road effects and a lack of long-term records that capture extreme floods contribute to uncertainty about the magnitude of road effects on very large floods or in large watersheds. Very few experimental studies have been conducted with road-only treatments, but many paired watershed forest harvest experiments include roads. Very few large watersheds lack roads, and fewer of these have streamflow records. Hence peak flow responses to forest harvest often include the effects of forest removal, the effects of roads, and the interac- tion between them (Jones and Grant, 1996). Roads, Trails, Erosion, Mass Movements, and Sedimentation High rates of overland flow along unpaved road surfaces entrain sediment, erode road surfaces, and contribute fine sediment to forest streams (Reid and Dunne, 1984). Overland flow on road surfaces and in roadside ditches and cul- verts concentrates soil moisture on oversteepened fillslopes below roads and cutslopes above roads, increasing susceptibility to landslides (Swanson and Dyrness, 1975; Larsen and Parks, 1998; Wemple et al., 2001; May, 2002). For- est roads can increase the landslide erosion rate by 30-300 times relative to un- disturbed areas (Sidle and Ochiai, 2006), much more than the effects of forest harvest. Road fills and cutslope areas are subject to landslides during storm events, and these can contribute large volumes of sediment to downstream re- ceiving waters (Swanson and Dyrness, 1975; Wemple et al., 2001). Landslides contribute fine sediment to streams, which can be detrimental to water quality and aquatic habitat, as well as coarse sediment and large pieces of wood, which are important structural elements for stream ecosystems but also can exacerbate downstream flooding and serve as tools for damaging roads and bridges. In addition to their effects on landslides, compacted road surfaces, cut- slopes, fillslopes, ditches, and areas below culverts are exposed to chronic sur- face erosion as a result of the generation and concentration of overland flow; roads deliver much of this fine sediment directly to streams, where it may be- come suspended sediment (Reid and Dunne, 1984; Bilby 1985). Suspended sediment is the most ubiquitous nonpoint pollution source from forests (Lands- berg and Tiedemann, 2000) and can degrade aquatic ecosystems by reducing water clarity, reducing interstitial flow and dissolved oxygen levels, and altering stream channel morphology (Waters, 1995; Stednick, 2000; Swanson et al., 2000). Sediment-laden water increases water treatment costs, reduces water storage facility storage volume and life span, and interferes with disinfection processes (NRC, 2000; Scatena, 2000; Stednick, 2000). Sediment particles also can bind with and become a transportation vehicle for contaminants such as nu- trients, metals, organic compounds, and pesticides.
FOREST DISTURBANCE AND MANAGEMENT EFFECTS ON HYDROLOGY 73 MANAGING FORESTS FOR WATER Forest hydrology principles elucidate direct effects of forest management and disturbance on hydrologic processes. Direct effects of forest management and disturbance on hydrology include increased net precipitation, temporary increases in water yield, and increased suspended sediment concentrations. Gen- eral principles of forest hydrology science indicate that increased water yield is one of the direct effects of forest harvest (Tables 3-1 and 3-2). Increases in water yield occur locally and may last up to a few decades after forest harvest. Although in principle forest harvest can increase water yield, in practice a number of factors make it impractical to manage forests for increased water. Water yield increases from vegetation removal are often small and unsustain- able, and timber harvest to augment water yield may diminish water quality. Increases in water yield tend to occur at wet, not dry, times of the year, and tend to be much smaller in relatively dry years. In addition, harvesting enough area to achieve a sustainable increase in water yield will have potential effects on wild- life fisheries and aquatic ecosystems Forest hydrology principles also describe indirect and interacting effects of forest management on hydrologic processes. Indirect effects are responses to forest management that are displaced in time or space, such as fire suppression leading to insect outbreaks that affect forest hydrology. Interacting effects occur when two or more management practices coincide, such as when post-salvaging logging and road building have a different collective effect on forest hydrology than their individual effects. The state of knowledge of forest hydrology pro- vides a strong foundation for knowledge of the direct effects of forest manage- ment and disturbance on hydrology. Contemporary forest management and dis- turbance processes raise issues that require extending the science from these principles to prediction, including indirect and interacting effects of changes in forest landscapes. These research needs are discussed in Chapter 4.