Restoring Ecosystem Health in Ponderosa Pine Forests of the Southwest

Home , Journal of Forestry

By W. Wallace Covington, Peter Z. Fule, Margaret M. Moore, Stephen C. Hart, Thomas E. Kolb,Joy N. Mast, Stephen S. Sackett, and Michael R.Wagner

major stumbling ment, heavy livestock grazing, fire sup- A block to ecosys- pression, logging disturbances, and cli- tem management is the matic events favored dense ponderosa lack of healthy ecosys- pine regeneration, and the open park- tems to serve as points lands closed. of comparison. Aldo Various authors (Covington and Leopold (1934) recog- Moore 1994a, 1994b; Kolb et al. nized this problem and 1994) have described unhealthy char- suggested using ecolog- acteristics of the postsettlement pon- ical restoration tech- derosa pine ecosystems, including in- niques to reconstruct creases in tree density, forest floor reasonable approxima- depth, and fuel loading, with the fol- tions of naturally func- lowing consequences: (1) decreases in tioning ecosystems soil moisture and nutrient availability, (Callicott and Flader (2) decreases in growth and diversityof 1992). We have initi- both herbaceous and woody plants, (3) ated an integrated set increases in mortality in the oldest age of research projects de- class of trees, (4) decreases in stream signed to accomplish and spring flows, and (5) increases in this goal and in the fire severity and size. process determine eco- Our research was guided by the fol- system health attrib- lowing general hypotheses: utes for a southwest- that both restoration of ecosystem ern ponderosa pine structure and reintroduction of fire are ecosystem. necessary for restoring rates of decom- Previous research position, nutrient cycling, and net pri- has established that mary production to natural, presettle- forests of ponderosa ment levels; and pine (Pinusponderosa) that the rates of these processes in the Southwest were will be higher in an ecosystem that ap- much more open be- proximates the natural structure and fore Euro-American disturbance regime. A controlled burn was set by researchers settlement (Cooper 1960; Covington Specifically, we hypothesized that in thinned stands of southwestern and Moore 1994b). Until the 1870s, reestablishing presettlement stand ponderosa pine to reconstruct the light surface fires every two to five structure alone (thinning a majority of open, parklike landscape that once years, along with grass competition the postsettlement trees) would result characterized the ecosystem. and regular drought, maintained an in lower rates of decomposition, nutri- open and parklike landscape domi- ent cycling, and net primary produc- nated by grasses, forbs, and shrubs tion compared with thinning, forest with scattered groups of ponderosa floor fuel manipulation, and pre- pine trees. After Euro-American settle- scribed burning in combination; but

Reprinted from the Journal of Forestry,Vol. 95, No. 4, April 1997. Not for further reproduction. 23 that both treatments would result in storage changed over the past century The time has come for higher rates of these processes com- of fire exclusion in a ponderosa pared with controls. We further hy- pine-bunchgrass ecosystem? science to busy itself pothesized that without periodic burn- 2. What are the implications of ing to hold them in check, pine these changes for net primary pro- with the earth itself. seedling irruptions would recur on the duction, decomposition, nutrient cy- The first step is to thin-only treatment area. We addressed cling, and other important ecosystem three specific questions: characteristics? reconstruct a sample 1. How have ecosystem structure 3. Does partial restoration (restor- (by biomass component) and nutrient ing tree structure alone by thinning of what we had to postsettlement trees) differ from complete begin with. restoration (the same --Aldo Leopold, 1934 thinning, plus forest floor removal, loading with herbaceous fuels, inches dbh are available at five-year in- and prescribed burn- tervals from 1920 to the present (Avery ing) in its effects on et al. 1976; and unpublished data). ecosystem structure The area is gentle in topography and function? (slopes less than 5 percent) with a southwest aspect. Elevation is 7,200 to The Study 7,400 feet. The climate is cool and Our study area is a subhumid with early summer drought. virgin stand of pon- Mean annual precipitation is 22.3 derosa pine within the inches and mean annual temperature is Fort Valley Experi- 45.5°F. Approximately half the precip- mental Forest approxi- itation falls as snow. The average frost- mately seven miles free growing season is 94 days. Soils at northwest of Flagstaff, the study site are derived from basalt Arizona. Stand struc- and volcanic cinders and are a fine ture is an uneven-aged montmorillonitic complex of frigid forest with trees in Typic Argiborolls and Argiboralfs. small, uneven-aged The understory consists primarily groups (fig.1), typical of Arizona fescue (Festuca arizonica), of ponderosa pine in mountain muhly (Muhlenbergia mon- the Southwest (Cooper tana), mutton bluegrass (Poa fendleri- 1960; Schubert 1974; ana), pine dropseed (Blepharoneuron White 1985). Pole- tricholepis),bottlebrush squirreltail (Si- sized trees, approxi- tanion hysterix),and a variety of forbs. mately 4 to 14.5 inches Buckbrush (Ceanothus fendleri) is the dbh, are the predomi- only common shrub. nant size class, with The site has never been harvested; scattered groups of pre- the only major disturbances have been settlement trees 14.5 to livestock grazing between 1876 and 41 inches dbh and 1910 and fire exclusion since 1876. dense “doghair” thick- Dieterich (1980) determined that the Figure 1. Canopy maps of the study area within the ets of saplings up to 3.9 last natural fire in the vicinity occurred FortValley Experimental Forest, near Flagstaff, Arizona. inches dbh. The treat- in 1876, before that, fires occurred at Scatteredgroups of southwestern ponderosa pine ment plots are in an an average interval of about two years. predominated in the late 19th century (top), eight-acreportion of the before settlement suppressed natural wildfires.This G.A. Pearson Natural Rationale forTreatments suppression caused a far greater profusion of Area, which was decom- A fundamental issue is what treat- vegetation in dense“doghair” thickets of pine saplings missioned in 1987. In- ment or combination of treatments (middle). Researchers are now attempting, through dividual tree growth and will rapidly achieve some facsimile of a prescribed burnsand manualthinning,to restore the mortality records for all healthy ponderosa pine ecosystem.The ecosystemto itsoriginal condition (bottom). trees greater than six two leading candidates for ecological

24 April 1997 restoration of ponderosa pine ecosystems are prescribed burning and thinning from below (Covington and Moore 1994b; Arno et al. 1995). Previous research near our study area had shown that although prescribed burning alone (without thinning or manual fuel removal) could reduce surface fuel loads, stimulate nitrogen availability,and increase herbaceous productivity, it caused high mortality of the presettlement trees (60 percent mortality over a 20-year period) and lethal soil temperatures under presettlement tree canopies (Covington and Sackett 1984, 1990, 1992; Harrington and Sackett 1990; Sackett et al. 1993). Although some thinning of postsettlement ponderosa pine trees was accomplished by prescribed burning (Harrington and Sackett 1990), results were localized, unpredictable, and difficult to control. Furthermore, even under very conservative prescriptions (low air temperatures and low windspeed), re- burning can produce dangerous fires (Covington and Sackett, pers. observ., 1984-95) because the continuing high density of postsettlement trees provides a continuous fuel ladder and thus a high crown-fire potential. Clearly, research shows that prescribed burning in today’s unnatural structure will not restore natural conditions in ponderosa pine-bunchgrass ecosystems. Thus, some combination of thinning, manual fuel removal, and prescribed burning will be necessary to quickly restore these systems to natural conditions.

Design and Sampling The study area was divided into 15 plots, five each in the completerestora- ment and potentially presettlement determined. A random 5 percent sub- tion (3.5acres), partial restoration (3.3 trees-those with diameter at breast sample of postsettlement trees was also acres), and control (3.8 acres) treat- height >14.8 inches or yellowed bark cored and aged. ments. Permanently marked photo (White 1985)-were cored and cross- Herbaceous and fuel variables were points were established. All plots were dated (Stokes and Smiley 1968) to de- measured on 55 replicated circular completely stem-mapped in 1992, in- termine age. Cross-sections or incre- subplots (16.4 feet in diameter) located cluding all living and dead trees and ment cores were taken from all pon- within four overstory strata: (1) preset- large downed logs. Species, condition derosa pine snags, stumps, and tlement patches, (2) postsettlement class, and diameter at breast height downed trees that were potentially of patches to be thinned with some trees (4.5 feet) were recorded for all trees, presettlement age. Samples were cross- retained, (3) postsettlement patches to and all retained and control plot trees dated and, when possible, death date, be converted to openings,and (4) rem- were permanently tagged. All presettle- pith date, and diameter in 1876 were nant natural grassy openings. Herba-

Journal of Forestry 25 ceous species composition, basal cover, and production were measured on two 10.7 ft2 quadrats within each subplot. Forest floor fuels were averaged from two 16.4-foot planar transects on each subplot (Sackett 1980). Both partial and complete restoration areas were thinned in fall 1993 and the stumps cut to ground level. All presettlement trees were retained, as were all trees above 16 inches dbh. In addition, smaller-diameter postsettlement trees were retained near stumps, snags, and down presettlement logs in an attempt to recreate Figure 2. Mean volumetric soil moisture content of mineral soil (0- to 12-inch the clumped pattern of the presettle- depth) for control, partial, and complete restorationtreatments during the I995 ment forest. Realizing that many of growingseason. Soil moisture intreated areas isappreciably higher during the the spindly postsettlement trees would normalearly summer drought,creating more favorable conditions for the survival not survive, we left approximately of young trees and other vegetation. three postsettlement trees for every dead presettlement tree. Slash was removed in spring 1994 and an eight-foot electrified elk fence was constructed around the study area perimeter. O n the complete restoration treatment plots, the litter layer

(Oi) was raked aside, accumulated duff (Oa and Oe) was removed, and litter was scattered back across the forest floor. Coarse woody fuels of presettlement origin, such as the limbs and boles of fallen presettlement trees, were retained, but postsettlement woody fuels were removed. Average duff depth in the presettlement stra- tum dropped from 3.3 to 0.4 inches Figure 3. Mean mineralsoil temperature (3-inch depth) in control, partial, and even though litter depth was nearly complete restoration treatments during summer and fall I995.The higher mineral unchanged (0.5 to 0.6 inches). Be- soil temperaturesin treated areas are presumedto be the result of increased cause the presettlement fuel structure insolationat the soil surface and decreased insulationat the forest floor. was composed primarily of grasses, approximately 600 pounds per acre of ceous vegetation cover, and fuel load- 1,240.7 in 1992, larger trees decreased supplemental native grass fuels (har- ing, we are monitoring a broad range from 19.4 per acre in 1876 to 12.8 in vested from a nearby natural meadow of ecological attributes related to eco- 1992, despite protection from logging. in September 1994) was scattered system health. The irruption of smaller-diameter pine across the complete restoration plots. trees also created a continuous tree Soil and tree cambial thermocouples Results and Discussion canopy cover (fig. 1) at the expense of were installed in representative loca- Dendrochronological analysis re- herbaceous vegetation. In 1876 only tions in each complete treatment plot vealed that forest structure had 19 percent of the surface area was (Sackett and Haase 1992). The plots changed substantially in our study area under pine canopy, with the balance slated for complete restoration were since fire regime disruption (table 1). (81 percent) representing grassy open- then burned on October 12, 1994, Particularly striking is the population ings, whereas in 1992pine canopy cov- when the temperature was 58o to growth of ponderosa pine from 22.8 ered 93 percent of the area, with only 60oF, winds were five to 10 mph from trees per acre in 1876 to 1,253.5 trees 7 percent left as grassy openings. the east-southeast, and relative hu- per acre in 1992. Although small-di- The spatial pattern of presettlement midity was around 30 percent. ameter trees (<16 inches dbh) in- trees in the Pearson Natural Area is In addition to tree density, herba- creased from 3.4 per acre in 1876 to striking. These trees grew in distinct

26 April 1997 0.05- to 0.7-acre groups of two to 40 longest period, with surface and soil trees (White 1985; Fule et al. 1993). temperatures exceeding 200°F (peak Although canopy cover within a group temperature >400°F) for 24 to 48 can be quite high, when averaged with hours. Fuels were reduced by the the interspaces that have no tree burn to an average 4.3 tons per canopy cover, the result is a low canopy acre-a total reduction of 85 percent cover overall. from the pretreatment fuel loading. Thinning resulted in the removal of During the 1995 growing season, 901 trees per acre. A total of 5,500 soil moisture and temperature were board feet per acre (3,700 board feet consistently higher in treated areas per acre of 9- to 16-inch dbh trees plus than in the control (figs. 2 and 3). 1,800) board feet per acre of 5- to 8.9- Soil moisture in treated areas is inch dbh trees) was removed. Most of higher probably because the canopy the smaller-diameter trees (629 trees and forest floor no longer intercept as per acre in the 1- to 4.9-inch dbh much precipitation, and there is also class) were used as latillas for adobe less transpiration by postsettlement presettlement trees was higher on the home construction, but the 37 tons trees. Higher mineral soil tempera- thinned area (4.4 ml per 24 hours) per acre of thinning slash posed a tures are presumably the combined than on the control area (3 ml per 24 major problem. Because there was no result of increased insolation at the hours), as was foliar toughness (76 g market for this material, it was hauled soil surface and decreased insulation of tension on thinned trees com- to a borrow pit and burned; the re- at the forest floor. pared with 69.8 g of tension on con- moval required about 75 loads in 18- We hypothesize that those micro- trol trees), suggesting increased resis- wheel dump trucks. environmental differences between tance to bark beetles and foliage- Before the forest floor was raked, treated and control areas will result feeding insects. No changes in popu- fuel loads averaged 28.3 tons per acre. in higher rates of such soil processes lations of turpentine beetles (Den- Approximately 21.3 tons per acre of as fine root production, litter de- droctonus valens) were observed. duff was removed by raking, some of composition, and nitrogen mineral- Furthermore, herbaceous produc- which became garden mulch. Addition ization in treated stands. Further- tion has responded markedly to the of mown grass (0.3 tons per acre) on more, we hypothesize that the treatments (fig. 4) with the greatest re- the complete restoration treatment in- changes in the soil process rates will sponse to date in the grassy substra- creased forest floor fuel loading to 7.5 in turn increase herbaceous produc- tum. By 1995 the treated areas were tons per acre; burning consumed 3.2 tion, tree growth, and resistance to producing almost twice as much tons per acre, leaving 4.3 tons. insect attack. In fact, resin flow of herbaceous vegetation as the control. For the complete restoration treatment, fire intensity was low: flame lengths averaged approximately six inches with occasional flareups of one to two feet. Several large dead and down presettlement logs, however, burned intensely and were completely consumed. Cambial heating was also low or insignificant, showing minor heat pulses (mean increase = 44°F, n = 23) during the burn. Individual thermocouples on two trees registered higher temperatures (maximum increases 82° and 147°F), but only on one of the six probes per tree. Surface soil temperatures peaked in excess of 400°F during burning, but subsurface temperatures (one to four inches) ex- Figure 4. Total aboveground herbaceous production (pounds per acre per year) by ceeded 150°F on only one of five treatment and patches. In postsettlement retained patches, postsettlement trees soil temperature sampling sites. were retained to replace presettlement trees that had died since 1876. Control Temperatures under dead presettle- treatment has no patches where postsettlement trees were removed. Grassy ment-era logs remained high for the patchesare natural openings in the overstory.

Journal of Forestry 27 Conclusions and Outlook term mean. We plan to continue this can alleviate many of the symptoms of Preliminary results from our ecosys- project for the next 24 years with eight ecosystem pathology while simultane- tem restoration work are encouraging. burnings at three-year intervals, rimed ously increasing our understanding of The combination of thinning and to coincide with the natural spring and ecosystem structure and function. burning-the complete restoration summer burning season. Data on other We argue that healthy conditions treatment-has changed forest struc- attributes will be used to increase un- should be based on an understanding ture from fire behavior fuel model 9 derstanding of the restoration treat- of the environmental conditions that (Anderson 1982), in which crown fires ments on a range of ecosystem charac- shaped the evolution of native forest arecommon, to fuel model 2, in which teristics. For example, we hypothesize ecosystems, and we urge the use of surface fires occur but crown fires are that the complete restoration treat- multidisciplinary measures of ecosys- highly improbable. ment will increase soil moisture and tem condition for understanding eco- The reduction in tree competition mineralization and uptake of nitrogen, system health. has improved moisture availability and leading to increased photosynthesis, has likely increased insect resistance of growth efficiency, and resin produc- presetdement trees. Grasses, forbs, and tion by presettlement trees, and in- Literature Cited shrubs are responding favorably as creased cover and production by ANDERSON, H.E. 1982. Aids to determining fuel well, indicating a shift away from a net shrubs and herbaceous vegetation. models for estimating fire behavior. General Tech- nical Report INT-122. Ogden, UT: USDA For- primary productivity dominated by With favorable climatic events, we ex- est Service. pine trees toward a more diverse bal- pect pine seedlings to become estab- ARNO, S.F., M.G. HARRINGTON, C.E. FIEDLER, and ance across a broader variety of plants. lished on the partial restoration rreat- C.E. CARLSON. 1995. Restoring fire-dependent Ecosystem restoration research re- ment, repeating the pine irruptions ponderosa pine forests in western Montana. Restorationand Management Notes 13(1):32-36. quires a long-term, interdisciplinary that occurred after fire regime disrup- AVERY, C.C., F.R. LARSON, and G.H. SCHUBERT. commitment. We expect the ecosystem tion in the 1870s. 1976. Fifty-year records of virgin stand develop- attributes we are measuring to be in Realizing that the intensive re- ment in southwestern ponderosa pine. General transition for the next 10 to 20 years search-oriented treatments we have Technical Report RM-22. Fort Collins, CO: USDA Forest Service. before they stabilize around some long- used in this research project are im- BASKETVILLE, G.L. 1965. Estimation of dry weight practical on an operational scale, we of tree components and total standing crop in have joined with the Arizona Strip Dis- conifer stands. Ecology 46:867-69. trict of the Bureau of Land Manage- BEARE, M.H., C.L. NEELY, D.C. COLEMAN, and W.L. HARGROVE. 1991. Characterization of a ment to test practical ecosystem res- substrate-induced respiration method for mea- toration treatments on a larger scale in suring fungal, bacterial and total microbialmass the Mount Trumbull Resource Con- on plant residues. Agriculture, Ecosystems and servation Area, north of the Grand Environment 34:65-73. Canyon. There we are using an adap- BINKLEY, D., and S.C. HART. 1989. The components of nitrogen availability assessmentsin for- tive ecosystem management approach est soils. Advances in Soil Science 10:57-112. (Walters and Holling 1990) to restore BROWN, J.K. 1974. Handbook for inventorying more than 3,000 acres of ponderosa downed woody material. General Technical Re- pine to conditions approximating port INT-16. Ogden, UT: USDA Forest Service. CALDWELL, M.M., and R.A. VIRGINIA. 1989. Root those that existed before Euro-Ameri- systems. In Pland physiological ecology: Field can settlement. We are using more methods and instrumcntation, 367-98. New practical approaches, such as raking York Chapman and Hall. the forest floor away from the base of CALLICOTT, J.B., and S.L. FLADER, eds. 1992. The River of the Mother of God and other essays by the presettlement trees (instead of re- Aldo Leopold. Madison: University of Wisconsin. moving it completely) and monitoring Press. a subset of our basic ecosystem health COOK, C.W., and J. STUBBENDICK, eds. 1986. attributes. Range research: Basic problems and techniques. Denver: Society for Range Management. Because the treatment areas at COOPER, C.F. 1960. Changes in vegetation, struc- Mount Trumbull are large, we are now ture, and growth of southwestern pine forest able to measure some variables that op- since white settlement. Ecological Monographs erate on a larger scale, such as passerine 30(2):129-64. COVINGTON, W.W., and M.M. MOORE. 1994a. bird populations, community structure Changes in multiresource conditions in pon- of selected insect guilds (butterflies), derosa pine forests since Euro-American settle- and other indicators of landscape-scale ment. Journal of Forestry 92(1):39-47. ecosystem health. Our hope is that ------, 1994b. Post-settlement changes in natural fire regimes and forest structure: ecological through such a set of integrated, adap- restoration of old-growth ponderosa pine for- tive ecosystem restoration projects, we ests.Journal of Sustainable Forestry 2(2):153-81.

28 April 1997 COVINGTON, W.W., and S.S. SACKETT. 1984. The PSW-GTR-131. Albany, CA: USDA Forest WHITE, A.S. 1985.Presettlement regeneration pat- effect of a prescribed burn in southwestern pon- Service. terns in a southwestern ponderosa pine stand. derosa pine on organic matter and nutrients in SACKETT, S.S., S. HAASE, and M.G. HARRINGTON. Ecology 66(2):589-94. woody debris and forest floor. Forest Science 1993. Restoration of southwestern ponderosa 30(3):183-92. pine ecosystems with fire. In Sustainable ecologi------, 1990. Fire effects on ponderosa pine soils cal systems: Zmplementhg an ecological approach and their management implications. In Effects of to land management, eds. W.W. Covington and ABOUT THE AUTHORS fire management of southwestern natural resources, L.F. DeBano. General Technical Report RM- 105-11. General Technical Report RM-191. 247. Fort Collins, CO: USDA Forest Service. W. Wallace Covington is regends' professor, Fort Collins, CO: USDA Forest Service. SCHUBERT, G.H. 1974. Silviculture of southwestern College of Ecoosystem Sciences and Manage------, 1992. Spatial variation in soil mineral ni- pondnosa pine: The status-of-our-knowledge. Re- ment, Northern Arizona University, trogen following prescribed burning in pon- search Paper RM-123. Fort Collins, CO: USDA Flagstaff 86011; Peter Z. Fulp is research derosa pine. Forest Ecology and Management Forest Service. 54(2):175-91. SMITH, R.H. 1971. Red turpentine beetle. Forest specialist, Northern Arizona University; EDWARDS, N.T. 1982. The use of soda-lime for Pest Leaflet 55. Washington, DC: USDA Forest Margaret M. Moore is associate professor, measuring respiration rates in terrestrial ecosys- Service. Northern Arizona University; Stephen C. tems. Pedobiologia 28:321-30. STOKES, M.A., and T.L. SMILEY. 1968. An intro- Hart is assistant professor, Northern Ari- duction to tree-ring dating. Chicago: University FRITTS, H.C., and T.W. SWETNAM. 1989. Den- zona University; Thomas E. Kolb is assis- droecology: A tool for evaluating variations in of Chicago Press. past and present forest environments Advances WAGNER, M.R., and Z.Y. ZHANG. 1993. Host tant professor, Northern Arizona Univer- in Ecological Research 19:111-88. plant traits associated with resistance of pon- sity; Joy N. Mast is assistant professor, FULE, P.Z., W.W. COVINGTON, and M.M. MOORE. derosa pine to the pine sawfly Neodiprion fulvi- Northern Arizona University; Stephen S. 1993. Scaling sample plots to estimate patch reps. Canadian Journal of Forest Research 23(5): Sackett is research forester Forest Fire Lab- characteristics in southwestern ponderosa pine. 839-45. Paper presented at the Eighth Annual US Land- WALTERS, C.J., and C.S. HOLLING. 1990. Large- oratory, Pacific Southwest Research Station, scape Ecology Symposium, March 24-27, scale management experiments and learning by USDA Forest Service, Riverside, Califor- 1993, Oak Ridge, TN. doing. Ecology 71(6):2,060-68. nia; Michael R Wagner is professor, North- HARRINGTON, M.G., and S.S. SACKETT. 1990. WARING, RH., and G.B. PITMAN. 1985. Modify- ern Arizona University. Funding for this ing lodgepole pine stands to change susceptibil- Using fire as a management tool in southwest- research was provided by the National Sci- ern ponderosa pine. In Effects of fire management ity to mountain pine beetle attack. Ecology of southwestern natural resources. General Techni- 66(2):889-97. ence Foundation (Grant DEB-9322706). cal Report RM-191. Fort Collins, CO: USDA Forest Service. HART, S.C., E.A. PAUL, and M.K. FIRESTONE. 1992. Decomposition and nutrient dynamics in ponderosa pine needles in a Mediterranean-type climate. Canadian Journal of Forest Research 22:306-14. KOLB, T.E., M.R. WAGNET, and W.W. COVING- TON. 1994. Concepts of forest health. Journal of Forestry 92(7):10-15. KOZLOWSKI, T.T., P.J. KRAMER, and S.G. PAL- LARDY. 1991. The physiological ecology of woody plants. New York: Academic Press. LEOPOLD, A. 1934. The arboretum and the university. Parks and Recreation 18(2):59-60. LEVERENZ, J.W., and HALLGREN. 1991. Measur- ing photosynthesis and respiration of foliage. In Techniques and approaches in forest tree ecophysi- ology, eds. J.P Lassoie and T.M. Hinckley, 303-28. Boston: CRC Press. PARIZEK, R.R., and B.E. LANE. 1970, Soil-water sampling using a pan and deep pressure-vacuum lysimeters.Journal of Hydrology 11:l-21. RITCHIE, G.A., and T.M. HINCKLEY. 1975.The pm- sure chamber as an instrument for ecological research. Advances in Ecological Research 9: 165-254. RUNDEL, P.W., and W.M. JARRELL. 1989. Water in the environment. In Plant physiological ecology today-- field methods and instrumentation, eds. R.W. Pearcy, J. Ehletinger, H.A. Mooney, and R.W. Rundel, 29-56. New York: Chapman and Hall. SACKETT, S.S. 1980. Reducing natural ponderosa pine fuels using prescribed fire: Two case studies. Research Note RM-392. Fort Collins, CO: USDA Forest Service. SACKETT, S.S., and S. HAASE. 1992. Measuring soil and tree temperatures during prescrbied fires with thermocouple probes. General Technical Report

Journal of Forestry 29