Fire Effects on Water Quality: a Synthesis of Response Regulating Factors among Contrasting Ecosystems

The key components of watershed processes are inputs in precipitation, interactions of vegetation, soil and water including evapotranspiration (water yield), overland flow (erosion), and storage and filtering (nutrients), and outputs in streamflow. Fire effects occur at the vegetation-soil interface and can result in altering overland flow and infiltration rate of water. Fire can affect infiltration rates by collapsing soil structure and reducing soil porosity, contributing ash and charcoal residues which can clog soil pores, and raindrop splash can compact soil and firther contribute to loss of soil porosity. An extreme example is the development of hydrophobic soils as observed in the western U.S. following severe wildfire. Watershed responses to fire depend on intensity and severity. Many factors influence fire severity including the quality and quantity of fuels, soil properties, topography, climate, and weather. The most important factors influencing the response to fire are vegetation mortality and the loss of the forest floor which are directly proportional to fire severity. Vegetation mortality reduces nutrient and water uptake, soil stability with root death, and the litter source for forest floor replenishment. The forest floor litter and humus (duQ layers provide soil cover, act as a sponge, and enhance infiltration. Large storm events immediately after a fie can accelerate surface runoff and compact soil. INTRODUCTION Wildland fire has the potential to significantly impact hydrologic processes such as surface nmoff, sediment yield, and sediment and nutrient transport to streams. The magnitude and duration of watershed responses to fue depends on the interactions among burn severity, post-fire precipitation regime, topography, soil characteristics, and vegetative recovery rate. The typical impact of fire is an immediate change in vegetative cover, forest floor surfhce, physical properties of the soil, followed by midand long-term changes in biological pools and nutrient cycling processes. Vegetation and litter protect the soil against the forces of erosion by maintaining high infiltration rates and low levels of overland flow (Covert and others 2005). Vegetative cover and forest floor are the primary drivers of sediment responses to fire. Large reductions of vegetative cover, particularly the ground vegetation and forest floor, leave the soil prone to raindrop impact and reduce rainfall infiltration and storage so that erosive overland flow tends to occur more readily (Shakesby and Doerr 2005). Nutrient responses are also impacted by changes in vegetation and forest floor, as well as changes in biological processes that regulate cycling processes. The magnitude and duration of hydrologic and water quality responses vary greatly across ecosystem types in the continental U.S. As such, it has been difficult to generalize response or apply knowledge derived fiom one region of the U.S. to any other. Over the past several years, a growing body of research has provided hydrologyfwater quality response data across a range of ecosystems, fire types, soils, and climate regimes. In this paper, we synthesize current knowledge on fixtors regulating water quality in contrasting ecosystems. Defining Fire Severity Fire severity depends on the interaction between fire intensity (rate at which thermal energy is produced) and duration (length of time burning occurs at a particular point) and describes the magnitude of the disturbance and reflects the degree of change in ecosystem processes (Neary and others 2005). Fire severity is a qualitative measure of the effects of fire on site and soil resources; it can occur along a spectrum fiom high to low or can be described as a patchwork, mosaic, matrix or mixed-severity event. Debano and others (1998) describe a light severity bum as one that bums only surface fuels, leaves the soil covered with partially charred organic material, and little to no duff consumption (fermentation (Oe) + humus (Oa) layers). A moderate-severity burn results fiom a large proportion of the organic material burned away fiom the surface of the soil and the remaining fuel is deeply charred. A highseverity burn results fiom all of the organic matexial burned away from the soil surfiice, organic material below the surface is consumed or charred. Fire severity has been assessed by numerous methods such as degree of destruction of aboveground live and dead biomass (Neary and others ZOOS), amount of forest floor consumed, particularly the duff layer, or heat penetration into the mineral soil (Swift and others 1993). Precipitation Regime After fire, rainfall intensity and duration can influence the amount of sediment delivered to a stream channel. The detachment of soil particles by rainsplash or overland flow and their transfer downslope are sensitive to modifications in land surface properties caused by fire (Johansen and others 2001, Sakesby and Doerr 2006). In low rainfall ecosystems, surface runoff and erosion may not be observed if there is a long period of post-fire recovery before the first rainfall event. Even in ecosystems with low mean annual rainfall, a high-intensity rainstorm immediately after wildand fire can create runoff that alters the topography of the hillslope, which subsequently impacts stream channels. Rainstorm events need to have enough energy to transport sediment. Swift and others (1993) determined that rainfall events of >50 rnrn hr-' were required to transport material after a fell-and-burn prescribed fire in the southern Appalachians. Sediment yields are typically higher in the first year after burning, especially when the burned watershed has been exposed to high-intensity rainfall events immediately after the fire has exposed the soil swfhce. Some of the largest increases in surface runoff have been observed where shortduration, high intensity convective rainstorms occur. For example, after the 1996 Buffalo Creek Fire in Colorado, two short-duration, high-intensity rainstorms (-90 mm hr-I) removed ash from the hillslopes, rilled the hillslope surfiices, channelized subtle drainages, which led to a headward extension of the channel network, and deposited sediment in stream channels (Moody and Kinner 2006). Kunze and Stednick (2006) found that rainfall intensity explained more than 80% of the variability in sediment yields. After the 2000 Bobcat Fire in Colorado, a single storm with 30 min rainfall intensity of 42 mm hr" resulted in 370 kg ha-' and 950 kg ha'' sediment yields, on treated (erosion-control with contour log felling, grass seeding, and mulching) and untreated watersheds, respectively (Kunze and Stednick 2006). Vegetation Recovery Post-fire soil erosion amounts vary not only with rainfall but also with bum severity, topography, soil characteristics and amount of vegetative recovery. Under moderate to severe fire severity that removes vegetation and forest floor cover, transpiration, interception and surfice storage capacity for rain are temporarily reduced. Conversely, any fireinduced alterations to storage capacity and water repellency will decline as vegetation and ground litter recover. Ground cover protects the soil from raindrop impact and offers resistance to overland flow. Vegetation recovery rates are strongly affected by fire size and severity, post-fire erosion events and vary by climate and geographic area. Rapid vegetation establishment has been regarded as the most cost-effective method to promote water infiltration and reduce hillslope erosion (Robichaud 2005). In the western U.S., land management agencies have spent tens of millions of dollars on post-fire emergency watershed stabilization measures to minimize flood runoff, onsite erosion, offsite sedimentation, mud and debris flows, and other hydrologic damage to natural habitats (Robichaud 2005). Post-fire hillslope rehabilitation treatments include seeding for vegetative re-growth, ground covers or mulches, and barrier and trenches that physically hold runoff and sediment In the eastem U.S., such costly and dramatic post-fire rehabilitation efforts are typically not required. Even after severe fire, recovery rates of southern Appalachian watersheds are much faster than western forests due to rapid vegetative re-growth (Clinton and Vose 2000, Elliott and others 1999). Low severity burning, such as prescribed fires, can promote a herbaceous flora (Elliott and others 1999, Gilliarn 1988, Hutchinson and others 2005) increase plant available nutrients (Elliott and others 2004), and thin-fiom-below over-crowded forests. While large, severe fires can cause changes in successional rates, alter species composition, generate volatilization of nutrients and ash entrainment in smoke columns, produce rapid or decreased soil mineralization rates, and result in subsequent nutrient losses through accelerated erosion (Neary and others 1999). Surfirce Runoff and Erosion Wildland fires are often landscape-scale disturbances that can alter the hydrologic and erosion responses of catchments. Erosion can occur when ground cover is reduced or consumed, and subsequently infiltration rates are reduced, i.e., water repellency is high. Fire-induced or enhanced soil water repellency (hydrophobicity) is commonly viewed as a key contributor to the substantial increases in hillslope runoff and erosion observed following severe wildfire (Debano and others 2005, Doerr and others 2006, Hufhan and others 2001,), particularly in the western U.S. Soils do not all exhibit the same degree of water repellency; a water-repellent soil is classified as one on which a drop of water will not spontaneously penetrate. Water drop penetration time (WDPT) has been used extensively to characterize soil water repellency (Letey 200 1). Several factors associated with fire, such as removal of surface litter and higher raindrop impact, would produce higher runoff and erosion fiom burned compared with unburned catchments, independent of water repellency. High runoff and erosion occurs from the combined effects of canopy destruction and water repellency induced by fire (Letey 2001), typically higher water repellency results fiom high severity fires (Lewis and others 2006). Sediment yields, in the first year after fire, range fiom very low in flat terrain without major rainfall events, to extreme in steep terrain affected by high-intensity thunderstorms. In the first post-fire year, sediment yield can vary fiom 0.01 to more than 110 Mg ha" (Robichaud and others 2000). High-intensity rainstorms after wildfire can create runoff that alters the topography of the hillslope, which subsequently impacts stream channels. In the coastal plain region of the Southeastern U.S., surface runoff and erosion from forested land would be minimal because the terrain is flat. On steep mountain slopes, Hendricks and Johnson (1944) found that sediment yield ranged fiom 71 Mg ha" yeail on 43% slopes, 202 Mg ha-' on 66% slopes, and 370 Mg K' yeail on 78% slopes after a wildfire in mixed conifer forests of Arizona. After the 1998 North 25 Fire in north-central Washington, Robichaud and others (2006) reported a first year mean erosion rate of 16 mg ha" yil (Table l), and this decreased significantly in the second year to 0.66 Mg h i ' y-'. Mean canopy cover (percent cover provided by live plants) was 18% the first year and 53% the second year after the wildfire. Total precipitation was below average during the fouryear period of their study (Robichaud and others 2006), and most erosion occurred during short duration, moderate intensity summer rainfall events. In the southern Appalachian mountain region, terrain is steep and rainstorms events with enough energy to transport sediment (250 mm br") have been recorded (Swift and others 1993), but vegetative recovery is rapid minimizing hillslope erosion. Table 1. Sediment losses the f i t year after prescribed (Rx) fire and wildfires. Location Community Severitylactivity 1" year sediment Reference loss (Mg ha") North Carolina, Pine/hardwoods Low severity, fell-and-bum 0.087 Switt and others 1993 Mountains South Carolina, Pine/hardwoods Low severity, Rx site 0.137 Robichaud and Waldrop Piedmont preparation burn 1994 High severity, Rx site 5.748 preparation burn South Carolina, Loblolly pine Control 0.027 Van Lear and others 1985 Piedmont Low severity. Rx understory 0.042 burn Moderate severity, 0.151 Cut + Rx bum Arkansas, Foothills Shortleaf pine Control 0.036 Miller and others 1988 High severity, slash cut + 0.237 Rx site preparation bum East Texas, Foothills Loblolly pine Clearmt + Herbicide + Rx 0.885 Field and others 2005 site preparation bum Clearcut + Mechanical 1.273 tillage + Rx site preparation bum Colorado Front Range Mixed conifer Low to moderate 0.20 to 0.05 Benavides-Solorio and severity, Rx iire MacDonald 2005 High severity, wildfire 2.0 to 10.0 Colorado Front Range Mixed conifer Unburned hillslopes 0.30 Moody and Martin 2001 High severity, wildfire 6.20 Wagenbrenner and others 2006 North-central Subalpine fir High severity, wildfire 16.0 Robichaud and others Washington 2006 In the western U.S., erosion rates increase by several orders of magnitude fiom areas burned at high severity because of the loss of protective ground cover and increase in surface runoff (Benavides-Solorio and MacDonald 2001, 2002, Robichaud and others 2000). In the Colorado Front Range, highest mean sediment rates were 80 100 Mg hafiom plots burned at high severity in recent wildfires (Benavides-Solorio and MacDonald 2005). The percentage of bare soil explained most of the variability in sediment yields (Benavides-Solorio and MacDonald 200 1). Johansen and others (2001) found that post-fire sediment yield increased non-linearly as percent bare soil increased. Specifically, sediment yields increased little when percent bare soil varied fiom 0 up to 60%, then yield increased exponentially above 60% bare soil. Ground cover effects appeared to be more important in explaining hydrologic response than either surface roughness or slope (Johansen and others 2001). Hence, maintaining vegetation cover or a cover of forest floor organic layers on the soil surface is the best means of preventing excessive soil erosion rates (Debano and others 1998, Neary and Ffolliott 2005, Wagenbrenner and others 2006). In the southeastern U.S., several authors have reported f ittle to no soil erosion after lightto moderate-intensity fires (Neary and Currier 1982, Shahlee and others 1991, Van Lear and Danielovich 1988, Van Lear and Waldrop 1986, Swift and others 1993). For example, Douglas and Van Lear (1983) found no significant differences in runoff or soil export between burned and unburned watersheds in the Piedmont of South Carolina Swift and others (1993) reported only minor and localized movements of burned plant hgrnents and soil after a fell-and-burn treatment in xeric pine-hardwood stands in the southern Appalachian Mountains of North Carolina In their study, soil erosion was minimal primarily because the forest floor remained largely intact; i.e., duff consumption ranged fiom 30 to 67 percent (Swifi and others 1993). Overall, these fires were classified as high intensity and low to moderate severity. Severity was moderate on portions of the burn where topography increased the frre intensity, causing greater proportions of forest floor consumption in small patches (Swift and others 1993). Effects were severe in a few spots where ribbons of soil were exposed after partially decomposed logs in contact with forest floor ignited and smoldered until consumed. After the burns, the bare soil exposure ranged fiom 7 to 14%. Where soil was exposed, the material was trapped within a short distance by residual forest floor and wood debris; thus, only two of eight sediment traps collected transported material resulting in < 0.10 Mg ha-' sediment lost (assuming a bulk density of 1.2 Mg ha-' and 40% charcoal by volume) the first year after the fires (Table 1). Sediment deposited at the lower margins of the study areas was transported by only three rainfall events that had enough force (> 50 mm hr-') to move sediment. Thereafter, no fiuther sediment was lost because subsequent rainhll events were not of sufficient magnitude to transport material. In their study, the residual forest floor was resistant to erosion over the range of burn intensities and sediment was prevented from leaving the site by unburned brush and undisturbed forest floor at the lower margins of the burned areas (Swift and others 1993). In the Piedmont region of South Carolina, Robichaud and Waldrop (1994) calculated sediment yields for lowand high severity site preparation burns in pine/hardwoods. For low severity fire (7% bare soil), sediments yields were 13.6 kg ha-' mm-' during simulated intense, rainfall (100 mm hr-' rain event lasting 30 min) with a total annual sediment loss of 0.137 Mg ha-' under natural rainfall events; and for high severity fire (63% bare soil), sediment yields were up to 27.7 kg h i ' mm" during simulated intense, rainfall with a total loss of 5.75 Mg ha" yearunder natural rainfall events (Table 1). In loblolly pine forests in South Carolina, Van Lear and others (1985) reported 0.042 Mg ha-' yr" and 0.15 1 Mg hae' 1-1 sediment loss fiom understory burn and burn + cut sites, respectively (Table I). Field and others (2005) estimated annual soil losses of 1.273 and 0.885 Mg ha-' from mechanical tillage and prescribed fire, respectively. Stream Suspended Sediment Severe wildfires can cause damage to plant cover and, thus, increase streamflow velocity, sediment delivery to streams, and stream water temperatures, as contrasted to low severity, cool-burning prescribed fires, which have less severe consequences (Reardon and others 2005). If surface erosion via overland flow reaches stream channels, then stream sediment concentrations increase proportional to the sediment delivered. Excess sediment is the principal pollutant of stream water associated with forest management (Phillips and others 2000) and is considered the primary threat to the integrity of aquatic resources (Henley and others 2000). After fire, excess sediment delivery to streams typically occurs after a measurable storm event. Watersheds severely denuded by fire are vulnerable to accelerated rates of soil erosion. While many fires increase sediment transport, wildfire often produces more sediment than prescribed fire (Debano and others 1998). Generally, prescribed fires, by their design, are not intended to consume extensive layers of forest floor litter. Without sediment transport via overland flow or surface runoff, input of sediment to streams would be minimal following prescribed fire or wildfire. If the forest floor remains intact and little to no bare soil is exposed, there is no mechanism for long-distance transport of sediment to streams (Vose and others 19991, regardless of rainfall event. In the western U.S., suspended sediment concentrations in streamflow can increase to very high levels following severe fire. For example, Hauer and Spencer (1998) found that stream sediment concentrations increased fiom 3.0 mg L" before fire to 32.0 mg L" after and Fredriksen (1971) recorded an increase fiom 2.0 mg L-' before disturbance to 150 mg L" following clearcut + slash burn. In contrast, fire in the southeast and southern Appalachians typically does not create conditions that result in sediment delivery to streams (Table 2). Forested streams in the southern Appalachians with high TSS during storm events are usually influenced by roads or land-use conversion (Table 2). Table 2. Total suspended solid (TSS) concentration in headwater streams with varying disturbance types and severity. Location Community Severity/activity TSS (mg L-') References North Carolina, Mesic hardwoods Low severity, prescribed fm 1-11 Vose, unpublished Mount a h East Tennessee and Pine/hardwoods Low severity, prescribed fire North Georgia, Mountains 1 -6 Elliott and Vose 2005 South Georgia, Mixed oak-pine Military training using traded 4 (baseflow) Houser and others Coastal Plain vehicles, ~ 7 % catchment area 57-300 (stormflow) 2006 disturbed (low severity) >7% catchment area disturbed 10 (baseflow) (hi# severity) 847-1 88 1 (stormflow) North Georgia, Mixed hardwoods Roads, land-use conversion 1-10 (baseflow) Riedel and others Mountains >lo0 (stormflow) 2003 W. Oregon Douglas-fir Clearcut, slash burn Montana Mixed conifer Wildfire 150 Fredriien 1971 32 Hauer and Spencer For prescribed fires in southern Appalachian pine-hardwoods, Elliott and Vose (2005) found no significant differences in total suspended solids (TSS) concentrations between burn and control streams over a 10-month postburn sampling period (Table 2). Several factors were attributed to the explanation of their results; a small rain event did occur the first day after the bum treatments, but this event brought less than 15 mrn of rainfhll, the low intensitylow severity, prescribed fire consumed less than 20% of the forest floor mass, and the burns were in the spring when vegetative re-growth occurs. With other disturbance types and intensities, other researchers have found more distinct and larger increases in sediment concentrations in highly disturbed streams (Houser and others 2006, Webster and others. 1990) than undisturbed streams during storm events (Table 2). For example, Houser and others (2006) investigated a range of disturbance intensities for typical low-gradient, Southeastern Coastal Plain streams to illustrate the impact of upland soil and vegetation disturbance on stream sediments. In catchments with a disturbance intensity of c7%, the mean maximum change in TSS ranged fiom 57 to 300 mg L" during storm events. In catchments with a disturbance intensity of >7%, mean maximum change in TSS ranged fiom 847 to 1881 mg L-' during storm events (Table 2). Stream Nitrogen The potential for increased N0i-N in streamflow after burning is attributed mainly to accelerated mineralization and nitrification (DeBano and others 1998, Knoepp and Swank 1993, Vitousek and Melillo 1979) and reduced plant uptake (Vitousek and Melillo 1979). Several studies on effects of prescribed fire on streamwater quality (Beche and others 2005, Clinton and others 2003, Douglas and Van Lear 1983, Elliott and Vose 2005, Field and others 2005, Richter and others 1982, Vose and others 1999), have found little to no detectable changes in streamwater chemistry after burning. For the few cases where a measurable increase in NO<-N was detected, timing of wildland fire influenced N0Y-N delivery to streams. In the spring, less NO,'-N will be transported to streams when vegetation uptake and microbial immobilization are typically high, compared to burns in the kll when vegetation is dormant. For example, Clinton and others (2003) compared stream NO3--N responses fiom watersheds burned in the fall and those burned in the spring. The two sites that showed a stream NO3--N response were burned in the fall, whereas the sites that were burned in the spring showed no response (Table 3). Table 3. Stream nitrate-nitrogen (NOi-N) responses following prescribed fire (Rx)and wildfire in the southeastern U.S. NOS'-NSite locationTreatment Community Fire severity Season response Duration References(mg rl) Jacobs Branch, NC Fell and Midelevation; High intensity, Fall 0,065 30 weeks Knoepp &burn, Rx Pine/hardwood moderate severitySwank 1993Wine Spring, NC Restoration, High elevation; ModerateSpring 0 None Vose andRxPine/hardwood intensity, lowothers 1999severityJoyce Kilmer, NC Wildfire High elevation; Low intensity, Fall 0.100 6 weeks Clinton andold-growthlow severityothers 2003hardwoodsHickory Branch, NC Restoration, Mid elevation; ModerateSpring 0.004 2 weeks Clinton andRxPine/hardwood intensity, lowothers 2003severity Conasauga,Understory, Low elevation;TN&GA;RxPinehardwoods Robin Branch, NC Understory, High elevation;Roach Mill, GA RxMesic, mixed oakUwarrie, NCUnderstory, Piedmont;RxpineJhardwoods Croatan, NCUnderstory, Coastal Plain;Rxlongleaf pineLow-to-moderate Springintensity, lowseverityLow intensity, Springlow severityModeratespringintensity,moderateintensityLow to moderate Winterintensity, low0 None Elliott &Vose 2005 0 None Vose andothers 20050 None Vose andothers 2005 0 None Vose andothers 2005 Vose and others (2005)compared the effects of low severity prescribed fire in Piedmont and southern Appalachianmountain streams (Table 3). In streamwater, measured N W N was extremely law (CO.1 mg NO;-N L-')before andafter burning. Both sites were burned in early spring and fires were confined to the understory and forest floor.There was generally no overstory mortality to prevent the rapid vegetation N uptake and immobilization of soilnutrientstypical of the spring growth flush. Fires were of low enough intensity to prevent significant overland flowand movement of nutrients off-site via physical changes in hydrologic processes. Vose and others (2005)also used anutrient cycling model to simulated stream NO<-N response under three frre scenarios: moderateseverity prescribedfire, high-severity prescribed fire, and highseverity wildfire. Only under the wildfire scenario was thereasignificant increase in stream NO3'-N concentrations. Vose and others (2005)attributed this simulated increase toreduced nitrogen uptake since the wildfire simulation included 100% overstory mortality. Under their wildfirescenario, streamwater NQ--N concentrations only reached 0.20 mg L" even with these extreme fire effects. Unlikelow to moderateseverity prescribed fires, largesevere wildfires often result in dramatic increases in stream solutes,which may last for years after the fire (Earl and Blinn 2003, Minshall and others 200 1, Spencer and others 2003).For example, Hauer and Spencer (1998) observed stream NO<-N concentrations from 0.12 to 0.30 mg L-' inimpacted streams after a wildfire in the Rocky Mountains, which were concentrations >5 fold over those observed incontrol streams. However, not all prescribed fires are low severity burns. Prescribed fire in the Tharp's Creek 16-hacatchment, Sierra Nevada of California killed most of the younger trees and understory vegetation, and the largertrees were scared, but leftalive. Most forest litter was cornbusted in the fire leaving an ash Iayer throughout thecatchment (Williams and Melack 1997). This prescribed burn in the Sierra Nevada of California resulted in thestream NO3-3-Nconcentration briefly exceeding 0.84 mg E' the first month of streamwater runoff after the fire, thenexceeding 1.96 mg L" three months after the fire. The following spring NO;-N concentrations increased above1.68 mg L", persisted above 0.84 mg L-'for several weeks, then returned to pre-fire conditions for the remainingyears after the fire (Williams and Melack 1997). Whereas, prefire stream NO;-N concentrations seldom exceeded0.01 mg L-'. In a recent national evaluation of forested streams, NCASI (2001) found that NO3--N concentrations for smallforested watersheds averaged 0.3 1 mg K' (median 0.15 mg L-I), and some streams averaged 10 times that level. Instreams draining both mountain and Piedmont regions of the southeast, fiom a range of fire intensities (fiom low tohigh; prescribed fire and wildfire), impacts on inorganic stream nitrogen levels are much lower (Clinton and others2003, Eliiott and Vose 2005, Vose and others 2005) than the average reported fiom NCASI (2001). CONCLUSIONSHydrologic and water quality responses to fire in the continental U.S. vary considerably. Whena wildland fire occurs, the principal concerns for changes in water quality are delivery of sediment and nutrients, particularlynitrate, into the stream channet. Fire managers can influence the effects of prescribed fire on water quality bylimiting fire severity, limiting fire size, and avoiding burning on steep slopes. Wildfires are typically larger and more severe consuming more fuel and releasing more nutrients than prescribed fire, which increases susceptibility toerosion of soil and nutrients into the stream. Our synthesis of a wide array of studies fiom across the U.S. supportthe following conclusions: 1. Maintaining an intact forest floor and promoting rapid vegetation recovery is critical to minimizing themagnitude and duration of sediment transport (surface erosion), sediment delivery (suspended solids) andsubsequent water quality responses,2. Burned areas are most vulnerable to surface erosion immediately post-fire and during extreme rainfallevents,3. Generally, water quality responses are much lower in the eastern U.S. than the western U.S. due to moremoderate topography, lower fire severity, and rapid vegetation recovery These regional differences emphasize the need for localized assessment and analyses of fire prescriptions, post-wildfire rehabilitation, and associated monitoring efforts. LITERATURE CITEDBCche, LA., S.L. Stephens, and V.H.Resh. 2005. Effects of prescribed fire on a Sierra Nevada (California, USA)stream and its riparian zone. For. Ecol. Manage. 21 8: 37-59. Benavides-Solorio, J.D. and L.H. MacDonald. 2005. Measurement and prediction of post-fire erosion at thehillslope scale, Colorado Front Range. International Journal of Wildland Fire 14: 457-474. Benavides-Solorio, J.D. andL.H.McDonald. 2001. Post-fire runoff and erosion &om; simulated rainfall on smallplots, Colorado Front Range. Hydrological Processes 15: 293 1-2952. Benavides-Solorio, J.D.and L.H. MacDonald. 2002. Errata for: Post-fire runoff and erosion from simulated rainfidlon small plots, Colorado Front Range. Hydrological Processes 16: 1131-1133. Clinton, B.D., and J.M. Vose. 2000. Plant succession and community restoration following fklling and burning inthe Southern Appalachian Mountains. Pp. 22-29 in W.K.Moser, and C.F. Moser (eds.), Fire and forest ecology:innovative silviculture and vegetation management; Tall Timbers fire ecology confkrence proceedings, no. 21; 1998April 1416; Tallahassee, FL: Tall Timbers Research Station. Clinton, B.D., J.M. Vose, J.D.Knoepp, and K.J. Elliott. 2003. Stream nitrate response to different burningtreatments in southern Appalachian forests. Pp. 174-181 in K.E.M. Galley, RC. Klinger, and N.G. Sugihara (eds.),Proceedings of the fire conference 2000: The first national congress on fire ecology, prevention, and managementproceedings. Miscellaneous Publication No. 13, Tall Timbers Research Station, Tallahassee, FL. Covert, S.A., P.RRobichaud, W.J.Elliot, and T.E. Link. 2005. Evaluation of runoff prediction fiom WEPP-Basederosion models for harvested and burned forest watersheds. Transactions of the ASAE 48: 109 1-1100. DeBano, L.F., D.G. Neary, and P.F. Ffolliott. 1998. Fire's effects on ecosystems. New York: John Wiley & Sons,Inc. 333 p. Debano, L.F. 2000. The role of fire and soil heating on water repellency in wildland environments: a review.Journal of Hydrology 23 1-232: 195-206. Debano, L.F., D.G. Neary, and P.F. Ffolliott. 2005. Soil Physical properties. Pp. 29-51 in D.G.Neary, K.C. Ryan,and L.F. Debano (eds.), Wildland Fire in Ecosystems: Effects of fire on soil and Water. Gen. Tech. Rep. RMRS-GTR-42-vol. 4, Ogden, UT,USDA Forest Service, Rocky Mountain Research Station. Doerr, S.H., R.A. Shakesby, W.H.Blake, C.J. Chafer, G.S. Humphreys, and P.J.Wallbrink. 2006. Effects ofdiffering wildfire severities on soil wettability and implications for hydrological processes. Journal of Hydrology319: 295-311. Douglas, J.E., and D.H. Van Lear. 1983. Prescribed burning and water quality of ephemeral streams in thePiedmont of South Carolina. For. Sci. 29: 181-189. Earl, S.R., and D.W. Blinn. 2003. Effects of wildfire ash on water chemistry and biota in South-Western U.S.A.streams. Freshwater Biology 48: 1015-1030. Elliott, K.J.,RL. Hendrick, A.E. Major, J.M. Vose, and W.T. Swank. 1999. Vegetation dynamics after aprescribed fire inthe southern Appalachians. For. Ecol. Manage. 114: 199-2 13. Elliott, K.J., J.M. Vose, B.D. Clinton, and J.D. Knoepp. 2004. E&cts; of understory burning in a mesic, mixed-oakforest in the southern Appalachians, Pp. 272-283in T.T. Engstrom, K.E.M.Galley, and W.J. de Groot (eds.),Proceedings of the 22"d Tall Timbers Fire Ecology Conference: Fire in temperate, boreal, and montane ecosystems,Tall Timbers Research Station, Tallahassee, FL. Elliott,K.J., and J.M. Vose. 2005. Initial effects of prescribed fire on quality of soil solution and streamwater in theSouthern Appalachian Mountains. Southern Journal of Applied Forestry 29: 515. Field, J.P.,K.W.Farrish, B.P. Oswald, M.T. Romig, and E.A. Carter. 2005. Forest site preparation effects on soiland nutrient losses in east Texas. Transactions of the ASAE 48: 861-869. Fredriksen, R.L. 1971. Comparative chemical water quality -natural and disturbed streams following logging andslash burning.Pp. 125-137 in Proceedings of a symposium on forest land uses and stream environment, 1970October 19-21, Corvalis, OR, Oregon State University, Continuing Educations Publication. Gilliam, F.S. 1988. Interactions offire with nutrients in the herbaceous layer of a nutrient-poor Coastal Plain forest.Bull. Torrey Bot. Club 115: 265-271. Hauer, F.R, and C.N.Spencer 1998. Phosphorus and nitrogen dynamics in streams associated with wildfire: a studyof immediate and longterm effects. International Journal of Wildland Fire 8: 183-198. Hadricks, B.A., and J.M. Johnson. 1944. Effects of fire on steep mountain slopes in central Arizona Journal ofForestry 42: 568-571. Henley, W.F., M.A. Patterson, RJ. Neves, and A. D. Lemly. 2000. Effectsof sedimentation and turbidity on loticfood webs: A concise review for natural resource managers. Rev. Fisheries Sci. 8: 125-139. Huffman, E.L., L.H.MacDonald, and J.D. Stednick. 2001. Strength and persistence of fire-induced soilhydrophobicity under ponderosa and lodgepole pine, Colorado Front Range. Hydrological Processes 15: 2877-2892. Hutchinson, T.F., R.E.J. Boerner, S. Sutherland, E.K. Sutherland, M. Ortt, and L.R Iverson. 2005. Prescribed fireeffects on the herbaceous layer of mixed-oak forests. Can. J. For. Res. 35: 877-890.Johansen, M., T.E. Hakonson, and D.D. Breshears. 2001. Post-fire runoff and erosion fiom rainfall simulations:contrasting forestswith shrublands and grasslands. Hydrological Processes 15: 2953-2%5. Knoepp, J.D., and W.T.Swank. 1993. Site preparation burning to improve southern Appalachian pinehardwoodstands: nitrogen responses in soil, soil water, and streams. Can. J. For. Res. 23: 2263-2270. . Knoepp, J.D., DeBano, L.F., and D.G. Neary. 2005. Soil chemistry. Pp. 53-71 in D.G. Neary,K.C. Ryan, and L.F.Debano (eds.), Wildland Firein Ecosystems: Effects of fire on soil and Water. Gen. Tech. Rep. RMRS-GTR-42-vol. 4, Ogden, UT, USDA Forest Service, Rocky Mountain Research Station. Kunze, M.D., and J.D. Stednick. 2006. Streamflow and suspended sediment yield following the 2000 Bobcat fie,Colorado. Hydrological Processes 20: 1661-168 1. Letey, J. 200 1. Causes and consequences of fire-induced soil water repellency. Hydrological processes 15: 2867-2875. Lewis, S.A., J. Q. Wu,and P. R Robichaud. 2006. Assessing burn severity and comparing soil water repellency,Hayman Fire, Colorado. Hydrological Processes 20: 116. Miller, E.L., R.S. Beasley, E.R. Lawson. 1988. Forest harvest and site preparation effects on erosion andsedimentation in the Ouachita Mountains. Journal of Environmental Quality 17: 2 19-225. Minshall, G.W.,J.R. Brock, D.A. Andrews, and C.T.Robinson. 2001. Water quality, substratum and bioticresponses of five central Idaho (USA) streams during the first year following the Mortar Creek fire. InternationalJournal of Wildland Fire 10: 185-199. Moody, J.A., and D.A. Martin. 2001. Initial hydrologic and gmorphic response following a wildfire in theColorado Front Range. Earth Surhce Processes and Landforms 26: 1049-1070. Moody, J.A., and D.A. Kinner. 2006. Spatial structures of stream and hillslope drainage networks following gullyerosionafter wildfire. Earth Surface Processes and Landforms 3 1: 319-337. National Council for Air and Stream Improvement, INC. (NCASI). 2001. Patterns and processes of variation innitrogen and phosphorus concentration in forested streams. Technical Bulletin 836, Research Triangle Park, NC. Neary, D.G., and J.B. Currier. 1982. Impact of wildfire and watershed restoration on water quality in SouthCarolina's Blue Ridge Mountains. South.J. Appl. For. 6: 81-90. Neary, D.G., J.D. Landsberg, A.R. Tiedemann, P.F. Ffolliott. 2005. Water quality. Pp. 119-134 in D.G. Neary,K.C. Ryan, and L.F. Debano (eds.), Wildland Fire in Ecosystems: Effects of fire on soil and Water. Gen. Tech. Rep.RMRS-GTR-42-vol. 4, Ogden, UT,USDA Forest Service, Rocky Mountain Research Station. Neary, D.G. K.D. Ryan, L.F. Debano, J.D. Landsberg, J.K. Brown. 2005. Introduction. Pp. 1-17 in D.G. Neary,K.C. Ryan, and L.F. Debano (eds.), Wildland Fire in Ecosystems: Effects of fire on soil and Water. Gen. Tech. Rep.RMRS-GTR-42-vol. 4, Ogden, UT,USDA Forest Service, Rocky Mountain Research Station. Neary, D.G., and P.F. Ffolliott. 2005. The water resource: its importance, characteristics, and general responses tofire. Pp. 95-105 in D.G. Neary, K.C. Ryan, and L.F. Debano (eds.), Wildland Fire in Ecosystems: Effects of fire onsoil and Water. Gen.Tech. Rep. RMRS-GTR-42-vol. 4, Ogden, UT, USDA Forest Service, Rocky MountainResearch Station. N w , D.G., P.F. Ffolliott, and J.D. Landsberg. 2005 Fire and streamflow regimes. Pp. 107-119 in D.G. N e q ,K.C. Ryan, and L.F. Debano (eds.), Wildland Fire in Ecosystems: Effects of fire on soil and Water. Gen. Tech. Rep.RMRS-GTR-42-vol. 4, Ogden, UT, USDA Forest Service, Rocky Mountain Research Station. Phillips, M.J.,L.W. Swift, Jr., and D.R Blinn. 2000. Best management practices for riparian areas. Pp. 273-286 inE.S. Verry, J.W. Hornbeck, and A.C. Dolloff (eds.), Riparian management in forests of the cantinental easternUnited States,Boca Raton, FL, Lewis Publishers, CRC Press LLC. Reardon, J.R., KC. Ryan, L.F. Debano, and D.G. Neary. 2005. Wetlands and riparian systems. Pp. 149-169 inD.G. Neary, K.C. Ryan, and L.F. Debano feds.), Wildland Fire in Ecosystems: Effects of fire on soil and Water.Gen. Tech. Rep. MS-GTR-42-vol. 4, Ogden, UT,USDA Forest Service, Rocky Mountain Research Station. Richter, D.D., C.W. Ralston, and W.R. Harms. 1982. Prescribed fire: Effects on water quality and forest nutrientcycle. Science 215: 66 1-663. Riedel, M. S., J.M. Vose, and D.S. Leigh. 2003. The road to TMDL is paved with good intentions Totalmaximum daily loads for a wild and scenic river in the southern Appalachians. Pp. 356-366 in A. Saleh (ed.), TotalMaximum Daily Load (TMDL) Environmental Regulations41 Proceedings of the 8-1 2 November 2003 Conference,Albuquerque, New Mexico, USA. ASAE Publication Number 701PI 503. Robichaud, P.R., and T.S. Waldrop. 1994. A comparison of surface runoff and sediment yields fiom lowand high-severity site preparation burns. Water Resources Bulletin 30: 27-34. Robichaud, P.R. 2000. Fire effects on infiltration rates after prescribed fire in Northern Rocky Mountains, USA.Journal of Hydrology 23 1-232: 220-229. Robichaud, P.R, J.L. Beyers, and D.G. Neary. 2000. Evaluating the effectiveness of postfire rehabilitationtreatments. USDA Forest Service, Rocky Mountain Research Station, Gen. Tech. Rep. RMRS-GTR-63, FortCollins, CO. Robichaud, P.R 2005. Measurement of post-fire hillslope erosion to evaluate and model rehabilitation treatmenteffectiveness and recovery. International Journal of Wildland Fire 14: 475-485. Robichaud, P.R., T.R. Lillybridge, and J.W. Wagenbrenner. 2006. Effects of postfie seeding and fertilizing onhillslope erosion in north-central Washington, USA. Catena 67: 56-67. Shahlee, A.K., W.L. Nutter, L.A. Morris, and P.R. Robichaud. 1991. Erosion studies in burned forest sites ofGeorgia. Pp. 3.3 1-3.3 8 in S.S. Fan, and Y.-H.Kuo (eds.), Proceedings of the 5& Federal Interagency SedimentationConference, Federal Energy Regulating Commission, Washington, D.C. Shakesby, RA., and S.H.Doen. 2006. Wildfire as a hydrological and geornorphological agent. Earth-ScienceReviews 74: 269-307. Spencer, C.N, K.O. Gabel, and F.R. Hauer. 2003. Wildfire effects on stream food webs and nutrient dynamics inGlacier National Park, USA. For. Ecol. Manage. 178: 141-153. Swift, L.W., Jr., K.J.Elliott, R.D. Ottmar, and R.E.Vihanek 1993. Site preparation burning to improve southernAppalachian pine-hardwood stands: fire characteristicsand soil erosion, moisture, and temperature. Can. J. For.Res. 23: 2242-2254. Van Lear, D.H., J.E. Douglass, S.K.Fox, and M.K. Ausberger. 1985. Sediments and nutrient export in runoff fkomburned and harvested pine watersheds in the South Carolina Piedmont. Journal of Environmental Quality 14: 169-174. Van Lear, D.H., and S.J. Danielovich. 1988. Soil movement after broadcast burning in the southern Appalachians.South.J. Appl. For. 12: 49-53. Vose,J.M.,W.T. Swank, B.D. Clinton, J.D. Knoepp, and L.W. Swift, Jr. 1999. Using stand replacement fves torestore southern Appalachian pine-hardwood ecosystems: effects on mass, carbon, and nutrient pools. For. Ecol.Manage. 114: 215-226. Vose,J.M., S.H. Laseter, G.Sun, and S.G. McNulty. 2005. Stream nitrogen response to fire in the southeastern U.S.Pp. 577-584 in S. Zhu, K. Minami, and G. Xing (eds.), 3d International Nitrogen Conference, Oct. 12-16,2004,Nanjing China, Science Press USA, Inc. Wagenbrenner, J. W., L.H.MacDonald, and D. Rough. 2006. Effectiveness of three post-fire treatments in theColorado Front Range. Hydrological Processes 20: 2989-3006. Webster, J.R, S.W. Golladay, E.F.Benfield, D. J. D' Angelo, and G.T. Peters. 1990. Effects of forest disturbance onparticulate organic matter budgets of small streams.J. North Am. Benthol Soc. 9: 120-140. Webster, J.R S.W.Golladay, E.F. Benfield, J.L. Meyer, W.T.Swank,and J.B.Wallace. 1992. Catchmentdisturbance andstream response: An overview of stream research at Coweeta Hydrologic Laboratory. Pp. 23 1-253in P.J.Boon, P. Calow, and G.C. Petts (eds.), River conservation and management John Wiley & Sons, New York. Williams, M.R., and J.M Melack. 1997. Effects of prescribed burning and drought on the solute chemistry ofmixed-conifkr fmest streams of the Sierra Nevada, California. Biogeochemistry 39: 225-253. Second InteragencyConference onResearch in theWatersheds