This file was created by scanning the printed publication. Errors identified by the software have been corrected; however, some errors may remain. KI·lllnlnh.()lz Sllowd.J~ifts: Hydrologic Ilnplic(lt:i()llS (It (1 C())O)·(ldo rr)·eeiille Site Neil H. 8erg1 Abstract--The water equivalent of snow contained in (irifts downwind of kJUlT1ITlholz vl~getation at treeline on Niwot l,id~le in the Colorado Front Range Is estimated at 8,060 m:3annually (1830 m3 1<m-2 Iand area), or about 1% of the sumlTH~r streamflow from a 191-ha alpine basin adjacentto Niwot Rld~le. Most of this water would be lost to sublimation If the vl~getatlon did not capture the snow. Much of the watl~r becomes available aftl~r the date of pe?ll< discharge and contributes to late··season water supply and soil watet recharge. -------------------------,---_. Blowing and drifting snow are important factors in alpine environments. The upper reaches of the subalpine z.one become a deposition site for snow redistributed from the alpine. At the forest-alpine tundra ecotone, low-lying "krummholz." vegetation is the first impediment to snow-bearing winds. When snow is redistribute.d into drifts to t.he lee of krummholz tree islands, evaporation of wi.nd-driven snow is reduced. As snow drifts me.lt, the.y become water sources to natural ecosystems and, particularly in their late .. season residence after the primary snow cover has melted, for human use. Niwot Ridge, an east-west trending spur on the east slope of the Colorado Front Range, 35 km northwest of Boulder, Colorado (4003'20" N, 105035' W), ranges in elevation from 3450 to 3800 m msl. At increasing elevations on Niwot Ridge, the closed crown forest gives way to stunted growth forms near the upper limit of tree growth as environmental conditions increase in severity. Severely deformed "krummholz" exist as isolated clumps of one or more trees. "Flag" krummholz grow where environmental conditions allow one or several vertical branches to survive above the protection of the winter snowpack; below the pack, growth is dense and mat-like. "J\.fat" krummholz exist at slightly higher elevations under extreme conditions that do not allow survival of the vertical leaders (Daly 1984). On Niwot: RidgePicea engeimannii (Engelmann spruce), with admixture·s of Abies /asiocarpa (subalpine fir) and Pinus J7eJoilis (limber pine), are the primary mat krum .. mholz species. Krul1lmholz. occur over an elevation range. of approximately 200 m and form a discontinuous catchment zone for blowing snow spanning the length of the Front Range (Ives and Hansen-Bristow 1983). This paper investigates the hydrologic. role of krummholz snowdrifts and estimates the water equivalence and timing of release of snowmelt water from krummholz drifts on Niwot Ridge. Methods To e.stimate the water equivalent of krummholz snow drifts, four steps were followed: (1) mea.sure krummholz. and drift geometries; (2) relate krummholz geometry to drift shape to determine individual snowdrift volumes; (3) quantify krummholz occurrence on Niwot Ridge; and (4) determine water equivalent. J\.iat krummholz tend to be wedge-shaped with their apex height at the downwind edge (fig. 1). Width and height measurements were made at the apex of 171 randomlyselected mat krummholz during February and J\.iarch, 1975, near the ce.ntral portion of the Ridge. The length of the associated lee drifts and snow depth at the apex were also recorded. Techniques have been established to estimate drift volume, V. Tabler (1975:95) developed a snow retention mode.! for sagebrush (Artemisia spp.) in which the drift forming behind an isolated plant approximates the shape of a half.. ~one, with length 10H (H:= plant height), so that V = 5.2H. The areal frequency of mat krummholz. was determined by relating aerial photograph observations to a vegetation map of the Niwot Ridge forest-alpine tundra ecotone. HansenBristmv's (1981) vegetation map (1:10,000 scale) shows mat krummholz. in a discontinuo,!!s band approxi.mately 350 m wide over an area about 4.4 km L , along the eastern two-thirds of Niwot Ridge. Si.nce i.ndividual plants were not designated on this map, color aeri.al photographs ( circa 1:15,000 scale) taken October 2, 1974 were analyz.ed with a zoom stereoscope capable of 10 X magnification. Individual trees were identified 1Supervisory Hydrologist, Pacific Southwest Forest and Range Experiment Station, Forest Service, U.S. Department of Agriculture, P.O. Box 245, Berkeley, Calif. 94701 141 Figure 1.--Schematlc representation of a mat krummholz and lee snowdrift. on the photograph and compared to their location on the map to determine growth form. Shadow length helped identify flagged trees. Determination of snow water e.quivalent requires knowledge of the density of the drifts. Koerner (1969) found density in krummholz lee drifts on Niwot Ridge to vary generally between 0.40 g cm-3 and 0.47 g cm-3 and inc.rease through time into the ablation period. These measurements, based on approximate.!y 100 sampling points, are the basis for use of 0.45 g cm- 3 as the value for snow density at the maximum accumulation date. Water equivale.nt for individual snow drifts follows as the. product of krununholz frequency, drift volume, and density. Total snow drift water equivalent for the basin is calculated by applying a probability density function developed from the field measurements of apex snow depth (taken to equal H) to the total krummholz population on the Ridge by means of the "half-cone" volume equation. island lengths (table 1), averaging 3.75 m, were similar to those mapped by Koerner (1969). Mean tree width-to-height ratio was 2.4, and lee drift lengths averaged 8 times ma~mum drift depth (table 1). Relative variation of tree length was greater than for the other measured properties. The coefficient of variation ranged from 0.40 to 0.47 for the other four variables, but was 0.65 for tree length. The distribution of snow depths at the tree apex is not significantly different from normal (alpha = 0.05) (fig. 2). Four thousand, six hundre.d and ninety mat krummholz (approximately 670 km- 2) were estimated as occurring within the Niwot Ridge forest-alpine tundra ecotone. The volume of snow in the associated lee drifts was estimated at 17900 m3, with a snow water equivalent of 8,060 m3, or 1,830 m3 km- 2 (at snow density = 0.45 g cm-3). To put this value into perspective, in 1974, a slightly above-average winter precipitation year on Niwot Ridge, peak daily discharge from a 191-ha alpine basin immediately south of the Ridge was 35,200 m3 (Carroll 1974). Drift snow water equivalent therefore approximated onequarter of the peak annual single-day streamflow, or about 0.8% of the mid-June through late-October streamflow. II -- 12 ~ Results and Discussion Measured tree heights (table 1) were within the ranges (0.5 to 2 m) measured by Koerner (1969), and noted by HansenBristow (1981:38): "In this form the tree is dwarfed to a mat, usually no higher than 1.5 m, but up to 5 m long." Similarly, tree 10 10 10 40 10 10 70 10 80 100 110 120 110 140 110 110 170 Snow Depth {em} Figure 2.--Krummholz apex snow depths on Niwot Ridge, Colorado. Table 1.--Krummholz and lee snowdrift dimensions on Niwot Ridge, Colorado, February-March, 1975. Variable Krummholz width Krummholz height at apex Krummholz length Drift length Drift depth at krummholz apex Mean Median Error Assessment Sample Min. size (Qm) (run) (run) S.D. (Qrn) Max. (Qm) 269 255 123 755 55 170 112 375 605 115 320 560 45 245 282 295 2,000 1,750 12 60 100 131 171 171 76 75 34 168 8 171 The magnitude of errors associated with several components of this analysis needs explanation. Errors associated with the krummholz frequency estimate are probably the largest in this study. Even with the optical magnification available for the aerial photograph interpretation, trees less than 20 cm tall may not have been identified and the distinction between mat and flagged forms was not always completely reliable. 142 The half-cone equation (Tabler 1975) was formulated for an area with unidirectional winds, uniform terrain sloping less than 15%, sufficient snow for the drifts to reach equilibrium profiles, crown width approximate.ly twice crown height, and drifts approximately 10H in length. These conditions were generally satisfied within the krummholz communities sampled on Niwot Ridge. In this application, apex snow depth is the effe.ctive crown height, since the top portion of the krummholz is sparsely foliated and relati.vely ineffectual as an aerodynamic barrier. On the average, measured tree width was slightly more than twice tree height and drift length was slightly less than 10 times snow depth at the apex. The importance. of barrier hei.ght in determining lee drift volume is supported by numerous snmy fence studies (Tabler 1986) and by the moderate.ly high Pearson c.orrelation coefficients for associations between drift length (a proX}'for drift volume) and both krummholz height (r= 0.82) and apex snow depth (r= 0.78). The shape and porosity of krummholz and sagebrush differ; krummholz are wedge-shaped whereas sagebrush are spherical. Not enoughi.s known about the magnitude of these. variations to quantify the e·rror. Use of a constant value for snow density is a simplification. Snow deposited i.n the interstices of the vegetative mat, while relatively low in volume, i.s not accommodated, nor is the interaction between krununholz drifts and topographic snow traps. Snow firnification, drift erosion due to high velocity winds, variations in tree island aspect and the consequent variations in ablation rates due to differences in insolation a.re likewise not c.onsidere.d. These jnflue.nces are essentially impossible to quantify on a basin scale. The calculations de.scrIbed here are a conservative "first approximation" of drift snow water equivalent. probably only a few days after the date of peak stream discharge. This may be due to their relatively shallow depths. Drifts behind flag krummholz., on the other hand, can persist several weeks longer, often well into July. The 1974 measurements support this claim. ~faximum depths at terrain drifts were greater than 3.8 m on April 17, 1974, a date when the maximum depth recorded at a nearby flag krummholz drift was 3.7 m. Snow over 1.5 m deep persisted at the terrain drifts as late as mid-July 1974, several weeks after the June 24 peak streamflow discharge in nearby Green Lakes Valley. Although measure·ments of the flag krummholz drift were not made in July, given the near ~quality of snow depths on April 17, it is reasonable to hypothesize. that the flag krummholz drift had not completely melted in mid-July, 1974. The flag form of krummholz is generally larger than the mat form, so that deeper drifts form which take longer to melt. It may be that the optimal condition for both drift accumulation and melt delay occurs when krummholz, either flag or mat, are situated immediately upwind of terrain snow traps, as on ridge-terrace line.s (Koerner 1969). Alpine. snowfields gene.rally supply late season runoff to lowland areas, and the preliminary observations reporte,d in this paper suggest that drifts associated with krummholz contribute to runoff even later than the. average from alpine c.atchments. Although the quantity of water from the drifts is relatively small, about 1%. of the summer streamflow from a 191-ha alpine. basin adjacent to the study area (or one-quarter of the. peak single day discharge), a 7-to-10 day delay in release of snowmelt water from the larger krummholz drifts adds to the importance of this water source in an otherwise "summer dry" lowland environment. Much of this water would be lost to sublimation if the vegetation did not accumulate snowdri.fts. Timing of Water Availability Acknowledgements Even if the actual water equivalent is twice the estimated amount, the total water contained in the drifts is relatively small when compared to typical streamflow volumes. The importance of this water is not primarily in its magnitude, but may be in the timing of its availability. Estimation of the timing of water release. from the drifts on Niwot Ridge has relied on two data sources: fie.ld measurements of several mat krummholz lee drifts made throughout the 1965 ablation period (Koerner 1969), and of a flag krummholz and terrain drifts during 1974 (Berg 1977). Extrapolation of Koerner's data suggested drift disappearance by June 25,1965. Extreme snowpack ablation rates occurred in 1965 (Rennkk 1966), therefore the late-June melt out date was probably earlier than average. Even so, in an above-normal snowfall year, 1974, peak stream discharge. occurred on June 24, suggesting that krummholz drifts may melt out after the stream flow peak. C~asual observations over several years by Daly (per. comm.1987) suggest, however, that mat krummholz drifts melt relatively quickly, and disappear This study was supported by the Institute of Arctic and Alpine Research, Unive.rsity of Colorado, and the. Roc.ky Mountain Forest and Range Experiment Station, Forest Service, U.S. Department of Agriculture.. Katherine HansenBristow, Christopher Daly, Bruce McGurk, Richard Minnich, and Ronald Tabler made valuable review comments on an earlier draft of the manuscript. Literature Cited Berg, Neil H.1977. Prediction of natural snowdrift accumulation on alpine ridge sites. Ph.D. Dissertation, Departme.nt of Geography, University of Colorado, Boulder, CO. :310 p. Carroll, Thomas Ray. 1974. The water budget of an alpi.ne catchment in central Colorado. :M:.A. Thesis, Department of Geogra.phy, University of Colorado, Boulder, CO. 124 p. 143 Koerner, John Marvin. 1969. Krummholz influences on alpine snow accumulations. J\f.A. Thesis, Department of Geography, lJniversity of Colorado, Boulder, CO. 115 p. Rennick, Kenneth 8.1966. Floods of May-June, 1965 in eastcentral Wyoming. Open-file report, U.S. Ge.ological Survey in cooperation with the \\Tyoming State. Engineer. Table.r, Ronald D.1975. Estimating the transport and e.vaporation of blowing snow. In: Snow Management on the Great Plains; Symposium, 1975 July; Bismarck, ND. Proc. Great Plains Agric. Coune., Pub!. 73. 85-104. Tabler, Ronald D.1986. Snow Fence Handbook (Re.lease 1.0). Tabler and Associates. P.O. Box 576, Laramie, \\7Y 82070. Daly, Christopher. 1984. Snow distribution patterns in the alpine krummholz zone. Progress in Physical Geography. 3(2): 157-175. Hansen-Bristow, Katherine Jane. 1981. Environmental controls influencing the altitude ~tnd form of the forest-alpine tundra ecotone, Colorado Front Range. Ph.D. Dissertation, De.partment of Geography, Unive.rsity of Colorado, Boulde.r, CO. 245 p. Ives, Jack D., and Katherine J. Hansen-Bristow.1983. Stability and instability of natural and modified upper timberline landscapes in the. Colorado Rocky J\fountains, USA. l\fountain Research and D~velopment. 3(2): 149-155. I 144