Controls of streamflow generation in small catchments across the snow-rain transition in the Southern Sierra Nevada, California

1Revised Manuscript for Hydrological Processes December 14, 2011 Controls of streamflow generation in small catchments across the snow- rain transition in the Southern Sierra Nevada, California Fengjing Liu1, 2, Carolyn Hunsaker3, Roger Bales1 10 11 12 Sierra Nevada Research Institute, University of California, Merced, CA Department of Agriculture and Environmental Science and Cooperative Research Program, 13 14 Lincoln University of Missouri, Jefferson City, MO Pacific Southwest Research Station, USDA Forest Service, Fresno, CA 15 16 17Corresponding address† 18 Fengjing Liu 19 Department of Agriculture and Environmental Science and 20 Cooperative Research Programs 21 Lincoln University 22 904 Chestnut Street 23 Jefferson City, MO 65101 24 Phone: 573/681-5390 25 Email: liuf@lincolnu.edu 26Abstract Processes controlling streamflow generation were determined using geochemical 27tracers for water years 2004-2007 at eight headwater catchments at the Kings River 28Experimental Watersheds (KREW) in the Southern Sierra Nevada Four catchments are snow 29dominated and four receive a mix of rain and snow Results of diagnostic tools of mixing models 30indicate that Ca2+, Mg2+, K+ and Cl- behaved conservatively in streamflow at all catchments, 31reflecting mixing of three endmembers Using endmember mixing analysis, the endmembers 32were determined to be snowmelt runoff (including rain on snow), subsurface flow, and fall storm 33runoff In seven of the eight catchments, streamflow was dominated by subsurface flow, with an 34average relative contribution (% of streamflow discharge) greater than 60% Snowmelt runoff 35contributed less than 40% and fall storm runoff less than 6% on average Streamflow peaked 2-4 36weeks earlier at mixed rain-snow than snow-dominated catchments, but relative endmember 37contributions were not significantly different between the two groups of catchments Both soil 38water in the unsaturated zone and regional groundwater were not significant contributors to 39streamflow The contributions of snowmelt runoff and subsurface flow, when expressed as 40discharge, were linearly correlated with streamflow discharge (R of 0.85-0.99) These results 41suggest that subsurface flow is generated from the soil-bedrock interface through preferential 42pathways and is not very sensitive to snow-rain proportions Thus a declining of the snow-rain 43ratio under a warming climate should not systematically affect the streamflow pathways at these 44catchments 45 46Key words: Hydrologic pathways, snow-rain transition, endmember mixing analysis, Southern 47Sierra Nevada 48Running Title: Controls of streamflow pathways in catchments of snow-rain transition 2 49INTRODUCTION 50 Precipitation has been changing in volume, intensity, and form (e.g., rain and snow) 51throughout many regions of the world due to climate warming [Dore, 2005; IPCC, 2007] In the 52mountains of the Western United States, trends toward less precipitation falling as snow [e.g., 53Mote et al., 2005; Knowles et al., 2006; Cayan et al., 2001] and the melting of snow earlier in the 54year [e.g., Stewart et al., 2004; Bales et al., 2006; Rauscher et al., 2008] are expected to 55continue April snow depth at index sites has decreased by 20-40% since the 1950s at moderate 56elevations (1500 – 2200 m) of Sierra Nevada [Mote et al., 2005] Observations and modeling 57results have shown that less snow and earlier snowmelt lead to a shift in peak river runoff toward 58late winter and early spring, away from summer when water demand is highest [e.g., Barnett et 59al., 2005; Stewart et al., 2005] However, it is still unclear how the decline in snow relative to 60rain systematically affects subsurface water storage and streamflow generation [e.g., Stewart et 61al., 2005; Kundzewicz et al., 2007] This hydrologic insight is critical for water resources 62management and has important implications for water supplies at local to global scales 63 Mechanisms of streamflow generation have been well studied for both rain-dominated 64and snow-dominated catchments across a wide range of climate, geology and vegetation Many 65studies have shown that shallow subsurface flow, including lateral flow, lateral subsurface flow, 66through flow, and interflow, is usually one of the important pathways in streamflow generation in 67small, forested catchments regardless of snow or rain dominance [e.g., Beighley et al., 2005; 68Redding and Devito, 2010; Hogan and Blum, 2003; Tromp-van Meerveld and McDonnell, 2006] 69For example, a few studies from an 870-m ponderosa pine hillslope at Los Alamos have 70indicated that lateral subsurface flow is an important flow process that controls snowmelt runoff 71at hillslope scales in semiarid environments [Wilcox et al., 1997; Newman et al., 1998; Newman 3 72et al., 2004] The importance of this process at catchment scales in semiarid regions with a 73seasonal snow cover has also been recognized [McNamara et al., 2005; Liu et al., 2008a, 2008b; 74Frisbee et al., 2011] However, notably lacking from these studies is a direct comparison to 75examine how the response of subsurface flow to snow and rain differs across catchments in the 76same region with similar geology, vegetation and soils A mixed snow-rain versus snow77dominated catchments in a given area may imply less in-catchment seasonal storage, shorter in78catchment residence times, and earlier seasonal change of soil storage [Bales et al., 2011; 79Hunsaker et al., in press] These discrepancies may cause differences in the processes that 80control streamflow generation in those catchments 81 The objectives of the study reported here were to quantitatively determine the dominant 82processes controlling streamflow across snow- and rain-dominated, headwater catchments and to 83understand how changes in the snow-rain proportion affect streamflow generation 84 85METHODS 86Research Area 87 This study was conducted in eight forested catchments that make up the Kings River 88Experimental Watersheds (KREW), a watershed-level, integrated ecosystem project for long89term research on nested headwater streams in the southern Sierra Nevada (Figure 1) KREW is 90operated by the Forest Service’s Pacific Southwest Research Station Four catchments are 91located at the Providence site and four catchments are located at the Bull site within the Sierra 92National Forest, northeast of Fresno, California (Figure 1) The four catchments at the 93Providence site range in size from 0.49 to 1.32 km and in elevation from 1479 to 2113 m, while 4 94the four catchments at the Bull site range in size from 0.53 to 2.28 km and in elevation from 952055 to 2490 m (Table 1) 96 Annual precipitation was 75-90% snow in the high-elevation Bull catchments (water 97years 2004 to 2007) and was up to 80% rain in the lower-elevation Providence catchments 98[Hunsaker et al., in press] For the same period, mean air temperatures were 7.8 and 6.8 oC at 99the lower-Providence and upper-Bull meteorological stations, respectively Soils are well 100drained, mixed, frigid Dystric Xeropsamment, formed from decomposed granite [Dahlgren et 101al., 1997], including Shaver and Gerle-Cagwin soils at Providence and colder, Cagwin soils at 102Bull [Sierra National Forest, 1983] Litter depth and depth to bedrock vary across the study area, 103but all soils have similar texture and water percolation rate [Bales et al., 2011] The Providence 104catchments are largely mixed-conifer forest, with some chaparral, barren and meadow The Bull 105catchments also are mainly mixed conifer forests, with a higher proportion of red fir at higher 106elevations 107 The study area and its vicinity are made up of granitic, metamorphic, and volcanic rocks, 108with some glacial-till materials Clay mineralogy is dominated by hydroxyl-Al interlayered 109vermiculite and gibbsite, as a result of weathering of feldspar and plagioclase under intense 110leaching environment [Dahlgren et al., 1997] This weathering process may cause much higher 111cationic concentrations (e.g., Ca2+, Mg2+ and Na+) in subsurface water than in rainwater and 112snowmelt This weathering environment is also very effective at removing Si in spite of the cold 113soil temperatures, resulting in Si-depleted minerals Quantitative pit and surface soil samples 114indicated that the higher-elevation Bull watersheds had significantly greater C, N and B contents 115in soils but lower extractable P, Ca2+, Mg2+ and Na+ contents than the lower-elevation Providence 116watersheds [Johnson et al., 2010] 5 117 118Sample collection and analysis 119 Streamflow samples were collected biweekly at the outlets of the eight catchments from 120fall 2003 to fall 2007 (Figure 1) Samples were either grabbed by hand or collected by automated 121ISCO samplers to increase sampling frequency during a storm The ISCO samplers were 122triggered when streamflow discharge exceeded a certain value and provided samples several 123hours apart during storm events 124 Soil water was collected from Prenart samplers at two depths, 13 and 26 cm Prenart soil 125samplers are suction-cup lysimeters that are made of porous teflon mixed with silica flour or 126stainless steel powder (for more information, see http://www.prenart.dk/sampler.php) Each pair 127of samplers was placed symmetrically at 2, 4, and m away from the tree trunk but under the 128canopy and one in the open at each depth Prenart samplers were deployed at all Providence 129catchments 130 Snowmelt was collected using plastic sampling bottles placed at four meteorological 131stations (Figure 1) Each bottle has a funnel to gather snow and allow meltwater to flow into the 132bottle Bottles were placed before a significant storm came and collected right after the storm 133ended 134 Samples were also collected in 2008 and 2009 from piezometers, a spring and 135groundwater wells in several locations (Figure 1) Groundwater was collected from drinking136water wells at Glenn Meadow, Dinkey Creek, the Pacific Gas and Energy (PG & E) work center, 137and the Blue Canyon Work Center to times in August 2008 and October 2009 A sample 138collected from a tank (used for supplying drinking water to local residents) near Dinkey Creek 6 139was actually from a nearby well Samples were taken once in October 2009 from a spring at 140B201 and two 1.5-m depth soil piezometers at B201 and P301 meadows 141 Samples collected from wells, the spring and piezometers in 2008-2009 were analyzed 142for major cations (Ca2+, Mg2+, Na+, K+) and anions (Cl-, NO3-, SO42-) using a Dionex 2000 Ion 143Chromatograph (IC) at the Environmental Analytical Laboratory of the University of California, 144Merced Analytical precision (1σ standard deviation) for all ions was less than 1% and detection 145limit less than µeq L-1 All other samples were analyzed for major cations and anions using IC 146at Pacific Southwest Research Station, Riverside, CA Precision is also less than 1% of ionic 147concentrations Acid neutralizing capacity (ANC) was calculated as the difference between the 148total concentrations of cations and anions, all in µeq L-1 149 150Endmember mixing analysis and diagnostic tools of mixing models 151 Contributions of endmembers to streamflow were determined using tracer-based 152endmember mixing analysis (EMMA) in combination with the diagnostic tools of mixing models 153(DTMM), following Liu et al [2008a] DTMM, developed by Hooper [2003], is used (i) to 154identify solutes that undergo chemical processes within and en route to streams and that behave 155conservatively upon mixing of various sources of water (endmembers) and (ii) to determine the 156number of endmembers needed for mixing of conservative solutes DTMM distinguishes 157whether solute correlations are controlled by chemical equilibria, which are nonlinear (solute 158concentrations are associated to each other with polynomial functions) and mixing, which is 159linear (solute concentrations are associated to each other by a linear function under one or more 160dimensions in U-space) Principal component analysis (PCA) is used to test which solutes are not 161associated linearly and which ones are, and under how many U-space dimensions Those solutes 7 162with polynomial relationships indicate the predominance of a chemical reaction for their 163formation Those with linear relationships suggest that their concentrations in streamflow are a 164mixture of various endmembers The number of U-space dimensions for linear expressions of 165conservative solutes is one less than the number of endmembers needed for the mixing 166 EMMA was then used with the determined conservative tracers to identify endmembers 167and quantify the contributions of endmembers to streamflow following Christophersen and 168Hooper [1992] and Liu et al [2004] PCA was performed again to extract eigenvectors using a 169correlation matrix (not the original ionic concentrations) of conservative tracers (not all solutes 170as used in DTMM) that were determined using DTMM above The PCA scores were used to 171solve for endmember contributions, a procedure mathematically the same as using two tracers for 172a three-component mixing model [e.g., Rice and Hornberger, 1998] 173 Three criteria were used to identify eligible endmembers from potential ones, following 174Liu et al [2008a] First, eligible endmembers must form a convex polygon (e.g., a triangle in the 175case of three endmembers) to bound most, if not all, streamflow samples Second, the distance of 176all eligible endmembers between original compositions (S-space) and U-space orthogonal 177projections should be reasonably short for all tracers used in the analysis The threshold values 178are not available in literature, but in the past studies of Liu et al (2004; 2008a; 2008b) 179endmembers with the projected to measured ionic concentrations less than 50-60% for all tracers 180worked very well, except for fresh snow with very low ionic concentrations Third, streamflow 181chemistry must be well recreated for conservative tracers using the results of EMMA 182 183RESULTS 184Hydrology and meteorology 8 185 Annual precipitation measured at four meteorological stations along an elevation gradient 186from 1750 to 2463 m was essentially the same (Figure 2a) Annual mean precipitation was 95, 187178, 198, and 76 cm in water years 2004, 2005, 2006 and 2007, respectively Precipitation 188primarily occurred from December to March, as seen from a sharp increase of cumulative daily 189precipitation (Figure 2a) Less than 10% of the annual precipitation occurred after April but 190before the fall wet season each year 191 Snow started accumulating in November or December and attained a maximum depth in 192early spring at all stations (Figure 2b) The maximum depth occurred at Upper Bull 193meteorological station, with 266, 380, 397, and 210 cm on March in 2004, March 24 in 2005, 194April in 2006 and February 28 in 2007, respectively The maximum depth at other stations was 195less than 70% of those at the Upper Bull station Snow depth declined almost monotonically as 196snow started melting The snowpack was depleted 2-3 months after the maximum accumulation 197at the Upper Bull, but 4-6 weeks earlier at the other stations 198 After maximum snow accumulation, streamflow runoff increased rapidly at all 199catchments, particularly in the relatively wet years 2005 and 2006 (Figure 2c) After snowpack 200depletion, cumulative runoff increased slightly with time The annual runoff was much higher in 2012005 and 2006 than 2004 and 2007 at all catchments The annual runoff also varied significantly 202among catchments, usually higher at B203 and B204 and lower at P303 and D102 For example, 203annual runoff was 38 and 12 cm at B203 and P303 in 2004 and 130 and 60 cm in 2005 at the 204same catchments, respectively Streamflow discharge peaked 2-4 weeks earlier at the Providence 205than Bull catchments during the snowmelt period from the maximum snow depth to snow 206depletion (bottom panels of Figure 3) Isolated streamflow peaks also occurred during winter and 9 207spring before the maximum snow accumulation, with some of peak discharges even higher than 208those during the snowmelt period 209 210Ionic concentrations 211 Mean ionic concentrations of Ca2+, Mg2+, Na+ and K+ in streamflow were significantly 212higher at Providence than at Bull catchments (Table 2) Mean concentrations of Ca 2+ were greater 213than 200 µeq L-1 (1σ > 50 µeq L-1) at Providence catchments, while those were less than 160 µeq 214L-1 (1σ < 35 µeq L-1) at Bull catchments Mean Cl- concentrations were slightly higher at 215Providence than at Bull catchments, but SO42- concentrations were slightly lower 216 The temporal variation of ionic concentrations generally followed the opposite pattern of 217streamflow discharge, with lower concentrations at higher flows during snowmelt and higher 218concentrations at low discharges, as demonstrated by Ca 2+, K+ and Cl- in Figure However, 219isolated peaks of high ionic concentrations, particularly those of K + and Cl-, occurred following a 220transient increase in streamflow discharge during late summer and fall or even in winter (Figure 2213) To the contrary, rain storms also occurred in spring, but there was a lack of isolated peaks in 222ionic concentrations during those events At Providence, ionic concentrations were generally 223lowest at P301 and highest at P304; at Bull, ionic concentrations were lowest at B203 and 224highest at T003, particularly for Ca2+ (Figure 3) 225 Mean ionic concentrations in snowmelt were much lower than in streamflow at both 226Providence and Bull catchments, but those in soil water were higher than in streamflow for all 227ions except Na+ and SO42- (Table 2) The mean concentration of Ca2+ in snowmelt was 27 µeq L-1, 228about 10% of that in streamflow at the Providence catchments, while that of soil water was 398 229µeq L-1, at least 20% higher than that in streamflow The mean concentration of Na + in soil water 10 10 298 A streamflow sample collected in fall with highest ionic concentrations was selected to 299characterize fall storm runoff, namely the one collected on October 5, 2006 at P304 for 300Providence catchments and on October 17, 2004 at B201 for Bull catchments The U- and S301space distance was less than and 10% for those two samples, respectively (Table 3) 302 Subsurface flow was characterized by streamflow samples collected during low 303discharges, following Liu et al [2008a; 2008b] The selection of streamflow samples for 304subsurface flow at each catchment in Table was made based on their geometrical positions in 305Figures and and the spatial endmember distances The distances are usually less than 10% for 306all ions, except Cl- (Table 3) Similar to ionic concentrations in snowmelt, the large distances for 307Cl- is likely caused by its relative low concentrations (Table 2) 308 309Endmember contributions 310 Relative contributions of snowmelt runoff (% of streamflow) were less than 35% on 311average from 2004 to 2007 at the Providence catchments, with higher values (mean ± 1σ 312standard deviation) at P301 (33±16) and P303 (35±19) than at P304 (24±15) and D102 (26±14) 313(Table 4) The mean contributions of subsurface flow varied between 60 and 70% at all of the 314Providence catchments, with 1σ standard deviations ranging from 17 to 20% Fall storm runoff 315contributed less than 10% at all catchments The Student’s t-tests (two-sample assuming unequal 316variances) showed that the mean contributions are not significantly different at P301 and P303 317and at P304 and D102, respectively, for both snowmelt runoff and subsurface flow (p > 0.05 for 318two tails) 319 Snowmelt runoff contributed less than 40% of streamflow on average from 2004 to 2007 320at all Bull catchments except B201, with 32(±21), 38(±17), 32(±16) at B203, B204 and T003, 14 14 321respectively (Table 4) The mean contribution of snowmelt runoff was 56(±18)% at B201 The 322mean contributions of subsurface flow varied between 60 and 66% for B203, B204 and T003, 323with 1σ standard deviations ranging from 16 to 20%, but 42(±18)% for B201 Fall storm runoff 324contributed less than 4% on average at all catchments The Student’s t-tests showed that the 325mean contributions are not significantly different for B203 and T003 (p > 0.05 for two tails) for 326both snowmelt runoff and subsurface flow More interestingly, the mean relative contributions of 327snowmelt runoff and subsurface flow were not significantly different (p > 0.05) between some 328catchments across Providence and Bull sites, including any pairs of B203, T003, P301 and P303 329and between B204 and P303 330 331Correlation of endmember contributions with streamflow 332 Contributions of snowmelt runoff and subsurface flow, by discharge rather than percents 333of streamflow discharge, were linearly correlated with streamflow discharge at both Providence 334and Bull catchments (Figures and 9) The R values were 0.85-0.99 and 0.91-0.96 (p < 0.001) 335for snowmelt runoff and subsurface flow, respectively The regression slopes varied from 0.45 to 3360.75 for snowmelt runoff and from 0.25 to 0.53 for subsurface flow The intercepts were all 337negative for snowmelt runoff, with a magnitude equal to or less than 12, and all positive for 338subsurface flow, with a value less than 7.1 There were a few outliers for subsurface flow in 339some of the catchments (Figure 9) These outliers mainly occurred on three days, December 31 340in 2005, January in 2006, and February 28 in 2006 It was presumably a result of analytical 341errors on ionic concentrations rather than instrumental errors on streamflow discharge, as there 342were no obvious outliers for snowmelt runoff 343 15 15 344DISCUSSION 345Mechanism of streamflow generation across snow-rain transition 346 Streamflow is dominated by subsurface flow in these small, snow-rain transition zone 347catchments, with a mean relative contribution greater than 60% from 2004 to 2007, except for 348B201 (Table 4) The importance of subsurface flow in streamflow generation is consistent with 349many studies from small, forested catchments in northern America with both snow- and rain350dominated hydrologic systems [e.g., Wilcox et al., 1997; Newman et al., 1998; Newman et al., 3512004; Tromp-van Meerveld and McDonnell, 2006; Hogan and Blum, 2003] However, the 352contributions of subsurface flow to streamflow reported here are generally higher than those in 353the other reports For example, subsurface flow accounted for only 40% of streamflow at the 354Hubbard Brook Experimental Forest [Hogan and Blum, 2003] Lateral subsurface flow occurred 355at an aspen-forested hillslope on the Boreal Plain only when precipitation is greater than 15 mm 356[Redding and Devito, 2008] 357 The contributions of subsurface flow are highly correlated to streamflow discharge 358(Figure 9), indicating that subsurface water is rapidly delivered to streams in response to rainfall 359and snowmelt events Soils at these catchments are dominated by sandy and loamy-sand textural 360classes, which are coarsely structured, with a sand fraction averaging above 0.70 [Bales et al., 3612011] This coarse soil structure may have contributed to the relatively high importance of 362subsurface flow The mean depth of soils from 87 soil pits is 75 and 77 cm for the Providence 363and Bull catchments, respectively [Johnson et al., 2010] Thus, subsurface flow is most likely 364generated from the shallow soil-bedrock interface through preferential pathways, consistent with 365the mechanism reported by others for the Mediterranean and humid climates (e.g., Fiori et al., 3662007; Tromp-van Meerveld and McDonnell, 2006; and Redding and Devito, 2010; Graham and 16 16 367Lin, 2011] On the other hand, soil-pit excavations indicate that soil thickness can vary from less 368than 50 cm to greater than 150 cm across short distances [Bales et al., 2011] The bedrock is also 369highly weathered and consists of unconsolidated deep regolith where hard bedrock is not 370typically encountered within a 150-cm depth [Bales et al., 2011] These observations suggest 371that bedrock depressions or hollows may exist at these catchments and may play a significant 372role in regulating subsurface flow, as proposed for the Panola Mountain Research Watershed by 373Freer et al [1997] Infiltrated water through preferential pathways such as macropores may have 374filled those hollows and then released together with preevent water already stored in the hollows 375to streams, a “fill and spill” mechanism suggested for the same Panola Mountain watershed by 376Tromp-van Meerveld and McDonnell [2006] We hypothesize that bedrock geomorphology at 377B201 is significantly different from the other catchments and it is this difference that caused the 378lower percentage of subsurface flow at B201 However, it is upon further data on bedrock 379geomorphology to confirm it 380 Soil water in the unsaturated zone does not play a significant role in streamflow 381generation This argument is line with the fact that soil water was not identified as a contributing 382endmember and its ionic concentrations were significantly different from those in streamflow, 383piezometer and springs (Table 2) Soil water chemistry of more than 800 samples was tested in 384EMMA by treating each individual sample as a potential endmember, but none of them qualified 385as a contributing endmember based on their positions in the mixing diagrams and projected 386endmember distances (data not presented) The calculation of water balance at watershed scale 387by Bales et al [2011] showed that soil moisture above m in depth is basically all consumed for 388evapotranspiration, consistent with Barnard et al [2010] for the H J Andrews Experimental 389Forest in Oregon The volumetric water content of soils down to m in depth varied between 20 17 17 390and 30% during the snowmelt periods and below 10% after those periods The high ionic 391concentrations in soil water were apparently caused by evapotranspiration This argument is also 392consistent with a stable isotope study in the H J Andrews Experimental Forest in Oregon that 393water entering the stream comes from preferential pathways and soil water consumed by trees 394does not participate in translator flow and significantly contribute to the stream [Brooks et al., 3952009] 396 Also, different from a large watershed in the southern Rocky Mountains reported by 397Frisbee et al [2011], these small catchments in southern Sierra Nevada were not fed by regional 398groundwater flow Deep groundwater such as those sampled in the Glenn Meadow and the Blue 399Canyon well was not a contributor to streamflow in these catchments (Figure 5) 400 401Will a change in snow-rain ratio affects streamflow generation? 402 Streamflow generation is apparently not controlled by the proportion of snow versus rain 403at these catchments, as the mean relative contributions to streamflow are even not significantly 404different between most higher-elevation Bull and lower-elevation Providence catchments for 405both snowmelt runoff and subsurface flow Thus even if snow-rain ratio at the higher-elevation 406Bull catchments decreases in the future due to climate warming, the mechanism of streamflow 407generation at the Bull catchments is not expected to change systematically or significantly Note 408that the difference between this statement and statements regarding the timing of streamflow 409peaks in spring with declining snow in mountains Less snow and earlier snowmelt have led to a 410shift in peak river runoff toward late winter and early spring in the U.S West [e.g., Barnett et al., 4112005; Stewart et al., 2005] and at these eight catchments [Hunsaker et al., in press] Our 412argument here in this report means that even if streamflow peaks earlier in future, the percent 18 18 413contributions of subsurface flow relative to snowmelt runoff (again which includes rain on snow) 414will not change significantly 415 This result, however, is derived from these small, forested catchments located in a 416relatively narrow elevation range in Sierra Nevada It is unclear how declines in the snow-rain 417proportion affect deep groundwater recharge, or mountain-block recharge as defined by Wilson 418and Guan [2004] As catchment size increases significantly, regional, long groundwater 419flowpaths may become increasingly important, as demonstrated by Frisbee et al [2011], and 420streamflow regimes may change significantly with a change in the snow-rain ratio 421 Also note that rain on snow was accounted the same as snowmelt by the definition of 422snowmelt runoff in this study because they are chemically undistinguished One might think that 423since rain on snow and snowmelt cannot be separated in the EMMA, the EMMA results would 424not be valid to be used to examine changes in streamflow generation associated with snow-rain 425ratio In fact, this is not right If rain on snow indeed generates different pathways from 426snowmelt, the responses of streamflow chemistry should be different Say if rain on snow would 427generate more subsurface flow than snowmelt, ionic concentrations in streamflow samples with 428rain on snow should be higher than those with snowmelt because ionic concentrations in 429subsurface water are higher With higher ionic concentrations in streamflow with rain on snow, 430more subsurface flow contributions would be detected in EMMA even if the same chemical 431signatures are used for both rain on snow and snowmelt 432 433Evaluation of EMMA results 434 The selection of endmembers for both the Providence and Bull catchments and numerical 435solutions of EMMA were quantitatively evaluated by recreating streamflow chemistry using the 19 19 436products of relative contributions of EMMA and ionic concentrations of endmembers (Figure 43710), following Christophersen and Hooper [1992] and Liu et al [2004] This recreation is not 438self-recurring because the relative contributions of EMMA were determined by correlations of 439ionic concentrations rather than ionic concentrations themselves [Liu et al., 2008a] The 440concentrations of Ca2+ and Mg2+ were very well recreated for both Providence and Bull 441catchments, with a slope near 1.0 and R2 greater than 0.92 between measured and recreated 442values The concentrations of K+ and Cl- were also very well recreated for Bull catchments, but 443R2 values were much lower for Providence catchments The lower R values of K+ and Cl- at 444Providence catchments were caused by a number of outliers (Figure 10) ANC was reasonably 445well recreated for both Providence and Bull catchments, with a slope near 1.0 and R > 0.92, 446even though it was not used in EMMA Recreation of Na + was also reasonably good, with R of 4470.83 and 0.71 for Providence and Bull catchments, respectively Na + did not behave completely 448conservatively (Figure 4) and thus an ideal recreation was not expected Nonetheless, the 449recreation of streamflow chemistry enhanced the confidence of the endmember selection 450 A single streamflow sample collected in October at P304 and B201 was used to 451characterize fall storm runoff for all Providence and Bull catchments, respectively (Figures and 4526) The sensitivity of EMMA results to ionic concentrations of fall storm runoff was tested by 453replacing this sample by a stream sample collected in October at their own catchment (results not 454presented) It turns out that the EMMA results for all three endmembers are not very sensitive to 455ionic concentrations of fall storm runoff The differences in relative contributions are less than 4561% for snowmelt runoff and subsurface flow and less than 2% for fall storm runoff 457 The geometrical position of soil water in the mixing diagram suggests that it could be a 458contributing endmember to streamflow instead of fall storm runoff at the Providence catchments 20 20 459(Figure 5) Using soil water rather than fall storm runoff in EMMA, the recreation of streamflow 460chemistry was very poor for K+ and Cl- at the Providence catchments, with a R2 value of 0.55 and 4610.44, respectively This result is consistent with the large spatial distance projected for Cl - using 462streamflow eigenvectors (Table 3) 463 464CONCLUSIONS 465 Streamflow is dominated by subsurface flow in small, forested, headwater catchments 466across the snow-rain transition elevations in the southern Sierra Nevada Subsurface water is 467primarily formed by relatively rapid infiltration of snowmelt and rainwater to the interface of 468lower soil horizons and bedrock through preferential pathways, stored in hollows and delivered 469to streams by “fill and spill” mechanism Both near-surface runoff and subsurface flow are very 470responsive to snowmelt and rainstorm events and are strongly linearly correlated with 471streamflow discharge With changes in snow-rain proportions, systematic changes in streamflow 472pathways are not expected under current land use and vegetation coverage in the southern Sierra 473Nevada headwater catchments 474 475ACKNOWLEDGEMENTS 476 Authors appreciate sample collection of T Whitaker and other staff at Pacific Southwest 477Research Station, Forest Service, U.S Department of Agriculture (USDA) Fresno, California 478Research was supported by the Forest Service, USDA, by Calfed (Carolyn, please add the name 479and number of your Calfed project here!), and by the National Science Foundation (NSF), 480through the Southern Sierra Critical Zone Observatory (EAR-0725097) and grant BES-0610112 481to the University of California, Merced Research was also supported in part by an Evans-Allen 21 21 482project (225140) and a Capacity Building Grant (2011-02451) funded through National Institute 483of Food and Agriculture (NIFA), USDA 484REFERENCES 485Bales, R C., N P Molotch, T H Painter, M D Dettinger, R Rice, and J Dozier (2006), 486 Mountain hydrology of the western United States, Water Resources Research, 42, 487 W08432, doi:10.1029/2005WR004387 488Bales, R C., J Hopmans, T O’Geen, M Meadows, P Hartsough, P Kirchner, C Hunsaker, D 489 Beaudette (2011), Soil moisture response to snowmelt and rainfall in a Sierra Nevada 490 mixed-conifer forest, Vadose Zone Journal, 10, 786-799 491Barnard, H R., C B Graham, W J Van Verselveld, J R Brooks, B J Bond, and J J 492 McDonnell (2010), Mechanistic assessment of hillslope transpiration controls of diel 493 subsurface flow: a steady-state irrigation approach, Ecohydrology, 3, 133-142 494Barnett, T P., J C Adam, and D P Lettenmaier (2005), Potential impacts of a warming climate 495 on water availability in snow-dominated regions, Nature 438, doi:10.1038, 17 November 496 2005 497Beighley, R E., T Dune, and J M Melack (2005), Understanding and modeling basin 498 hydrology: interpreting the hydrogeological signature, Hydrologic Processes, 19, 1333- 499 1353, DOI: 10.1002/hyp.5567 500Brooks, J R., H R Barnard, R Coulombe, and J J McDonnell (2009), Ecohydrologic 501 separation of water between trees and streams in a Mediterranean climate, Nature 502 Geoscience, DOI: 10.1038/NGE0722 503Cayan, D R., S A Kammerdiener, M D Dettinger, J M Caprio, and D H Peterson (2001), 504 22 Changes in the onset of spring in the western United States, Bulletin of the American 22 505 Meteorological Society, 82 (3), 399-415 506Christophersen, N and R P Hooper (1992), Multivariate analysis of streamflow chemical data: 507 the use of principal components analysis for the end-member mixing problem, Water 508 Resources Research, 28(1), 99-107 509Dahlgren, R A., J L Boettinger, G L Huntington, and R G Amundson (1997), Soil 510 development along an elevational transect in the western Sierra Nevada, California, 511 Geoderma, 78, 207-236 512Dore, M H I (2005), Climate change and changes in global precipitation patterns: What we 513 know?, Environmental International, 31, 1167–1181 514Fiori, A., M Rommanelli, D J Cavalli, and D Russo (2007), Numerical experiments of 515 streamflow generation in steep catchments, Journal of Hydrology, 339, 183-192 516Freer, J., J J McDonnell, K J Beven, D D Brammer, D Burn, R P Hooper, C Kendall 517 (1997), Topographic controls on subsurface storm flow at the hillslope scale for two 518 hydrologically distinct small catchments, Hydrological Processes, 11(9), 1347-1352 519Frisbee, M D., F M Phillips, A R Campbell, F Liu, and S A Sanchez (2011), Streamflow 520 generation in a large, alpine watershed in the southern Rocky Mountains of Colorado: Is 521 streamflow generation simply the aggregation of hillslope runoff responses?, Water 522 Resouces Research, 47, W06512, doi:10.1029/2010WR009391 523Graham, C B and H S Lin (2011), Controls and frequency of preferential flow occurrence: A 524 175-event analysis, Vadose Zone Journal, 10 (3), 816-831 525Hogan, J F., and J D Blum (2003), Tracing hydrologic flow paths in a small forested watershed Sr/86Sr, [Ca]/[Sr], [Ba]/[Sr] and δ18O, Water Resources Research, 87 526 using variations in 527 39(10), 1282, doi:10.1029/2002WR001856 23 23 528Hooper, R P (2003), Diagnostic tools for mixing models of streamflow chemistry, Water 529 Resources Research, 39(3), 1055, doi: 10.1029/2002WR001528 530Hunsaker, C., T Whitaker, and R C Bales (in press), Water yield and runoff timing across the 531 rain-snow transition in California’s southern Sierra Nevada, Journal of the American Water 532 Resources Association 533Intergovernmental Panel on Climate Change (IPCC) (2007), Climate Change 2007: The 534 Physical Science Basis Contribution of Working Group I to the Fourth Assessment Report 535 of the Intergovernmental Panel on Climate Change, edited by S Solomon et al., 536 Cambridge Univ Press, New York 537Johnson, D W., C T Hunsaker, D W Glass, B M Rau, and B A Roath (2010), Carbon and 538 nutrient contents in soils from the Kings River Experimental Watersheds, Sierra Nevada 539 Mountains, California, Geoderma, doi:10.1016/j.geoderma.2010.10.019 540Knowles, N., M D Dettinger, and D R Cayan (2006), Trends in snowfall versus rainfall in the 541 Western United States, Journal of Climate, 19, 4545-4559 542Kundzewicz, Z W., L J Mata, N W Arnell, P Döll, P Kabat, B Jiménez, K A Miller, T Oki, 543 Z Sen and I A Shiklomanov (2007), Freshwater resources and their management, 544 Climate Change 2007: Impacts, Adaptation and Vulnerability, Contribution of Working 545 Group II to the Fourth Assessment Report of the Intergovernmental Panel on Climate 546 Change, M.L Parry, O.F Canziani, J.P Palutikof, P.J van der Linden and C.E Hanson, 547 Eds., Cambridge University Press, Cambridge, UK, 173-210 548Liu, F., R C Bales, M H Conklin, and M E Conrad (2008a), Streamflow generation from 549 snowmelt in semi-arid, forested and seasonally snow-covered catchments, Valles Caldera, 550 New Mexico, Water Resources Research, 44, W12443, doi:10.1029/2007WR006278 24 24 551Liu, F., R Parmenter, P Brooks, M Conklin, and R C Bales (2008b), Intra- and inter-annual 552 variation of streamflow pathways in semiarid, forested catchments, Valles Caldera, New 553 Mexico, Ecohydrology, 1(3), 239-252, doi: 10.1002/eco.22 554Liu, F., M Williams, and N Caine (2004), Source waters and flow paths in an alpine catchment, 555 Colorado Front Range, USA, Water Resources Research, Vol 40, W09401, doi: 556 10.1029/2004WR003076 557McNamara, J P., D Chandler, M Seyfried, and S Achet (2005), Soil moisture states, lateral 558 flow, and streamflow generation in a semi-arid, snowmelt-driven catchment, Hydrological 559 Processes, 19, 4023-4038 560Mote, P W., A M Hamlet, M P Clark, and D P Lettenmaier (2005), Declining mountain 561 snowpack in western North America, Bull Ameri Meteor Soc., 86, 1-39 562Newman, B D., A R Campbell, and B P Wilcox (1998), Lateral subsurface flow pathways in a 563 semiarid ponderosa pine hillslope, Water Resources Research, 34(12), 3485-3496 564Newman, B D., B P Wilcox, and R C Graham (2004), Snowmelt-driven macropore flow and 565 soil saturation in a semiarid forest, Hydrological Processes, 18, 1035-1042 566Rauscher, S A., J S Pal, N S Diffenbaugh, and M M Benedetti (2008), Future changes in 567 snowmelt-driven runoff timing over the western US, Geophysical Research Letters, 35, 568 L16703, doi:10.1029/2008GL034424 569Redding, T and K Devito (2008), Lateral and vertical flow thresholds for aspen forested 570 hillslopes on the Western Boreal Plain, Alberta, Canada, Hydrological Processes, 23, 4287– 571 34300 572Redding, T and K Devito (2010), Mechanisms and pathways of lateral flow on aspen-forested, 573 25 Luvisolic soils, Western Boreal Plains, Alberta, Canada, Hydrological Processes, 24, 2995– 25 574 3010, DOI: 10.1002/hyp.7710 575Rice, K C and G M Hornberger (1998), Comparison of hydrochemical tracers to estimate 576 source contributions to peak flow in a small, forested, headwater catchment, Water 577 Resources Research, 34(7), 1755-1766 578Sierra National Forest (1983), Soil Survey of Sierra National Forest Area, California, U.S 579 Department of Agriculture Report, Clovis, CA, 413pp 580Stewart, I T., D R Cayan, and M D Dettinger (2004), Changes in snowmelt runoff timing in 581 western north America under a ‘business as usual’ climate change scenario, Climatic 582 Change, 62: 217–232 583Stewart, I T., D R Cayan, and M D Dettinger (2005) Changes toward earlier streamflow 584 timing across western North America, Journal of Climate, 18, 1136-1155 585Tromp-van Meerveld, H J., and J J McDonnell (2006), Threshold relations in subsurface 586 stormflow: A 147-storm analysis of the Panola hillslope, Water Resources Research, 587 42, W02410, doi:10.1029/2004WR003778 588Wilcox, B P., B D Newman, D Brandes, D W Davenport, and K Reid (1997), Runoff from a 589 semiarid ponderosa pine hillslope in New Mexico, Water Resources Research, 33(10), 590 2301-2314 591Wilson, J L and H D Guan (2004), Mountain-block hydrology and mountain-block recharge, 592 In Groundwater Recharge in A Desert Environment: The Southwestern United States, 593 edited by Phillips, F M., J F Hogan, and B R Scanlon, AGU, Washington, DC 594 595 596 26 26 597Figure Captions 598Figure Study area, showing locations of eight catchments, stream gauges (stream sampling 599 sites), meteorological stations (snowmelt sampling sites), piezometer, wells and spring 600 Most groundwater wells are located outside the map and indicated by directions and 601 distances 602Figure Hydrometeological data from water years 2004 to 2007 for (a) daily cumulative 603 precipitation at four meteorological stations, (b) daily snow depth at the same stations, 604 with peak and snow depletion dates at the Upper Bull station indicated, and (c) daily 605 cumulative runoff at eight catchments Grey highlighted periods are from peak snow 606 depth to snow depletion at the Upper Bull station 607Figure Concentrations of selected ions from water year 2004 to 2007 in streamflow at eight 608 catchments at Providence and Bull, along with streamflow discharge, all in common 609 logarithmic scales Data points appear too intense and thus symbols are not used for 610 clarity Date-marked peaks were discussed in text Grey highlighted periods are from 611 peak snow depth to snow depletion at the Upper Bull station 612Figure Distribution of residuals between projected and measured concentrations with 613 measured concentrations in streamflow for all solutes used in diagnostic tools of mixing 614 models in 1-D and 2-D mixing spaces 615Figure Mixing diagrams using the first U-space projections, along with potential and 616 selected endmembers and the triangle they form at Providence catchments The ranges of 617 ordinate and abscissa are focused on stream samples, with a smaller panel on right 618 showing at full ranges at P303 as an example Abbreviations are: BCW for Black Canyon 619 well, DCW for a well near Dinkey Creek, DCT for another well near Dinkey Creek (but 27 27 620 sampled from a tank), GMW for Glenn meadow well, and HELM for PG & E well For 621 clarity of the figures, the errors are not shown for soil water and Dinkey Creek well in the 622 enlarged figures 623Figure The same as Figure 5, but for Bull catchments 624Figure Relative (percent) endmember contributions to streamflow from water year 2004 to 625 2007 at Providence and Bull catchments Grey highlighted periods are from peak snow 626 depth to snow depletion at the Upper Bull station 627Figure Correlation between snowmelt runoff and streamflow discharge at each catchment 628Figure Correlation between subsurface flow and streamflow discharge at each catchment 629 Dark red squares are outliers 630Figure 10 Recreation of ionic concentrations in streamflow based on relative contributions of 631 endmembers determined by EMMA and ionic concentrations in endmembers Note that 632 ANC and Na+ were not used in EMMA 28 28 ... Hydrologic pathways, snow-rain transition, endmember mixing analysis, Southern 4 7Sierra Nevada 48Running Title: Controls of streamflow pathways in catchments of snow-rain transition 2 49INTRODUCTION... 46 6across the snow-rain transition elevations in the southern Sierra Nevada Subsurface water is 467primarily formed by relatively rapid infiltration of snowmelt and rainwater to the interface of. .. timing of streamflow 409peaks in spring with declining snow in mountains Less snow and earlier snowmelt have led to a 410shift in peak river runoff toward late winter and early spring in the