Introduction
Atmospheric deposition accounts for significant nutrient and
pollutant input to high-elevation watersheds such as the Sierra Nevada
(Dolislager et al., 2006; Fain et al., 2011; McDaniel, 2013; Sickman et
al., 2003; TERC, 2011; Vicars and Sickman, 2011; Williams and Melack,
1991a, b). Sierra Nevada snowpack supplies the majority of water to
downstream communities as well as to some of the nation's largest
agricultural areas. Quantifying atmospheric deposition in alpine watersheds
is challenging because of large spatial variability in deposition rates
caused by complex terrain, precipitation gradients, and varied origins of
atmospheric constituents (i.e., local versus regional and global, natural
versus anthropogenic; Jassby et al., 1994; Rohrbough et al., 2003).
Single-site measurements, therefore, do not allow for accurate extrapolation
of nutrient or pollutant deposition in alpine regions and broader temporal
and spatial data are needed to assess the mass and dynamics of atmospheric
inputs.
In this study, we used multiple and repeated sampling of full-depth snowpack
cores (integrated snowpack sampling) across the Lake Tahoe Basin to quantify
atmospheric deposition loads and patterns from the first snowfall until the
end of melting. Snowpack acts as an integrating reservoir for water, solutes,
and particulates that deposit throughout winter and spring (Turk et
al., 2001). Wet deposition, in the form of snowfall and rain, directly
accumulates in the developing snowpack throughout the snow season (Kuhn,
2001). Additionally, during storm-free periods, snowpack also receives dry
deposition, which is often complicated to quantify since dry deposition
samplers can be biased due to different collection efficiencies compared to
natural surfaces (Jassby et al., 1994). Representing a natural surface that
covers the ground for several months of the year, snowpack sampling thereby
can provide accurate on-the-ground measurements of total (bulk: wet and dry)
deposition occurring in mountainous areas.
While the snowpack integrates wintertime atmospheric deposition input, it
also records chemical and physical transformations that occur during storage
such as elution during melt events, chemical transformations, and
volatilization. For example, ionic pulses of anions and cations occur upon
snowpack melt, whereby ions are thought to be mobilized in the following
order: SO42- > NO3- > Cl- > alkali
metals > alkaline earth metals > cations (other than
NH4+) > anions > NH4+ > H2O2 (Berg, 1992; Brooks
and Williams, 1999; Kuhn, 2001; Stottlemyer and Rutkowski, 1990; Williams and
Melack, 1991b). In addition, pollutants such as Hg and persistent organic
pollutants (POPs) as well as nutrients can undergo photochemical
transformations and be subject to substantial gaseous re-emission to the
atmosphere (Fain et al., 2011; Halsall, 2004; Lalonde et al., 2002; Poulain
et al., 2007). Specific examples include photochemical reduction and
re-emission of mercury (Hg) during snowpack storage as well as photolysis and
emission of nitrate (NO3-) from polar snow (Galbavy et al., 2007; Jacobi
and Hilker, 2007; Rothlisberger et al., 2002). In addition, microbial
activity in and under the seasonal snowpack can play an important role in
snowpack N dynamics (Brooks et al., 1996; Williams et al., 1996), even in
Arctic environments with low temperatures and minimal water content (Larose
et al., 2013). Therefore, snowpack sampling yields relevant temporal
atmospheric deposition patterns in conjunction with post-depositional
chemical losses or conversions.
Spatially, snowpack sampling can be an elegant tool to quantify gradients in
atmospheric deposition that are difficult to assess with other methods; for
example, the Sierra Nevada show strong orographic precipitation effects, with
the leeward side receiving significantly less precipitation than the windward
side (O'Hara et al., 2009). Such different precipitation patterns can cause
large differences in wet deposition across mountain ranges (Fain et
al., 2011; NADP, 2012). Assessing spatial deposition patterns using snowpack
sampling at multiple locations across a watershed should allow for better
characterization of basin-wide deposition patterns as well as assessment of
impacts of nearby urban areas versus regional and global sources of
atmospheric deposition (Brown et al., 2011; Kuhn, 2001; Morales-Baquero et
al., 2006; Vicars and Sickman, 2011).
The main goal of this study was to quantify N, P, and Hg concentrations and
loads in Sierra Nevada snowpack in order to characterize the magnitude,
origin, and fate of atmospheric deposition of nutrients and pollutants that
accumulate throughout the winter and spring in this mountain range. We
quantified chemical loading at seven sites in the Lake Tahoe Basin, along two
elevation transects, throughout the duration of two full snow seasons.
Sampling included biweekly snowpack cores (full profile; integrated snowpack
samples) representing an integrated load of constituents in the developing
snowpack collected throughout the 2011–2012 and 2012–2013 snow years. In
addition, volume-weighted wet deposition measured at two sites in 2013–2014
was compared to snowpack accumulation and detailed vertical snow pit profiles
in that year to compare snowpack accumulation to wet deposition and to
further study in-snowpack chemical dynamics. Finally, basin-wide loading
estimates (massarea-1) were calculated by spatially extrapolating
nutrient and pollutant measurements across the basin combined with a
satellite-based snowpack reconstruction model.
Materials and methods
Study Site
The Lake Tahoe watershed lies in the northern portion of the Sierra Nevada
range along the border of Nevada and California. Renowned for its intense
blue color and water clarity, this lake has become a national landmark and
tourism hotspot. Lake clarity measurements have decreased, however, from
approximately 30.5 to 21.3 m since the 1960s due to eutrophication from
increased input of N and P, as well as additional input of light-scattering
particulates (TERC, 2011). Directly west and upwind of the basin lies the
central valley of California and cities of Sacramento and San Francisco,
CA, which are believed to contribute significant amounts of
nutrients and pollutants to the basin through agricultural and industrial
emissions.
Including all drainages, the Lake Tahoe watershed has an area of 1310 km2
(Fig. 1). The lake is 19 km wide and 35 km long with a total surface area
of 495 km2. The lake lies at 1897 m above sea level and is on average
300 m deep. Surrounding the lake on all sides are mountains up to elevations
of 3068 m. At the lake's surface, summer temperatures reach on average
27 ∘C and wintertime lows reach -9 ∘C. Precipitation
patterns in the watershed are highly dependent on elevation with an average
annual precipitation of 0.76 m at lake level and an average of 2.03 m
falling at higher elevations in the surrounding mountains (Fram and Belitz,
2011). Extreme snow events in this area are common and often produce snowpack
depths greater than 4.5 m at high elevations. Rain shadow effects typically
lead to decreased snow loading on the downwind, eastern side of the basin.
Approximately two-thirds of Lake Tahoe Basin parent material is granitic and
one-third is volcanic (LTTMDL, 2008). Vegetation, consisting of mixed
coniferous forest and montane-subalpine species, cover approximately 80 %
of the basin (LTTMDL, 2010). Areas of dense urban development occur along the
shoreline at South Lake Tahoe, Tahoe City, and Incline Village. Large
portions of the northern and western shores are occupied by seasonal cabins,
while much of the eastern shore is undeveloped.
Sample collection
Integrated snowpack sampling: 2011–2012 and 2012–2013 snow years
During the 2011–2012 and 2012–2013 water years, full-snowpack biweekly core
samples were collected at seven sites in the basin starting from the first
measurable snowpack until the majority of spring melting occurred
(2011–2012: n=49; 2012–2013: n=56). This included mid-January
through mid-April in 2012 and December to early April in 2012–2013. The
seven sites were distributed along eastern and western elevation transects
(Fig. 1). Three of the sites were located at lake level (one remote site, two
sites in urban areas, elevation approximately 1900 m), two sites were at
mid-mountain elevation (approximately 2200 m), and two sites were at high
elevation close to the mountain ridges (elevation approximately 2500 m). To
minimize throughfall signals, we selected areas that were free of canopy
coverage and had minimal snowpack disturbance (i.e., away from congested
areas). Canopy effects on total snow accumulation are incorporated in the
snow water equivalent (SWE) reconstruction model. However, measurements of
deposition and chemical snowpack storage are based on canopy-free, open
locations, and do not include effects of forest cover.
Lake Tahoe watershed map with biweekly sampling sites located along
east and west basin elevation gradients for spatial and temporal sampling
campaigns in 2011–2012 and 2012–2013. Additional wet deposition and snow
pit profile samples were collected near the Homewood High and Mt. Rose sites
during the 2013–2014 snow year.
Samples were collected using a Mt. Rose Federal sampler and were immediately
transferred to Whirl-Pak® clean bags and a
cooler with blue ice packs. Samples were transported within four hours to the
Desert Research Institute in Reno, NV, for storage at -20 ∘C until
laboratory analysis could be completed. Depth and SWE
were measured for each core using the Mt. Rose Federal sampler. In cases of
low snow accumulation, multiple cores were collected and homogenized to
provide sufficient sample for all analyses. During collection, sterile gloves
were worn, and soil contact and contamination were avoided in order to
capture only constituents stored within the snowpack. While sampling, the
first core taken at each site was discarded in order to avoid carryover from
previous sampling. Between each sampling campaign, the Federal sampler was
cleaned with Milli-Q deionized water (< 18.2 MΩ) and a
chelating soap in accordance with trace metal sampling procedures (EPA,
2002). Field blanks were measured by rinsing the sampler with Milli-Q water
prior to each sampling campaign.
Wet deposition sampling and snow pit collection: 2013–2014 snow year
In order to differentiate between snowpack storage and wet deposition and
further assess dynamics in the snowpack, additional sampling of full snow pit
profiles and wet deposition was completed during the 2013–2014 snow year.
Biweekly wet deposition sample collection following National Atmospheric
Deposition Program protocol (http://nadp.sws.uiuc.edu/) was conducted
at the two high-altitude sites by N-Con dual port trace metal samplers (model
TM 00-127; N-Con Inc., Crawford, GA, USA). These samplers allowed for
collection of real-time wet deposition samples of both nutrients (N, P, and
S) and Hg without cross contamination. The sample trains consist of an NADP
standard (19-128) glass funnel, a glass anti-evaporation capillary, a glass
sample bottle (2 L) for collection of Hg, and a (19-130) polyethylene funnel
with connector and 1.5 L HDPE sample container for nutrients. The glass
sample bottle was pre-charged with 20 mL of deionized water and 0.5 mL of
12 M HCl (EMD Omnitrace HX0607) to act as a preservative for Hg. Sample
bottles were collected in the field and kept in a cooler during transport
back to the Desert Research Institute in Reno, NV. Sample bottles were then
weighed in the lab and decanted into 250 mL HDPE bottles for nutrient
samples and glass containers for Hg samples. All samples were stored in
refrigerators until processing.
Three snow pit analyses were conducted at the high-elevation sites, two near
the Mt. Rose site (1 March and 4 April 2014) and one at the Homewood High
site (28 February 2014). The snow pit measured a minimum of a 1.5 m2 and
was dug from the snow surface to the ground. Measuring sticks were placed on
either side of the pit face. A measurement of height, layer density, and
crystal form was noted. Snow samples were collected vertically every 10 cm
using a 1000 cm3 Kelly wedge cutter (model: RIP 1 Cutter; Snowmetrics,
Fort Collins, CO, USA). Prior to collection, the acid-washed wedge was
inserted into the snow adjacent to the sample wall two to three times at each
layer before sampling to avoid carry over. Duplicate samples were collected
at each height and analyzed separately. All samples were double-bagged in
Whirl-Pak clean bags and weighed for
density. Samples were then transferred to -20 ∘C storage at the
Desert Research Institute in Reno, NV, until analysis. Reported concentrations
and densities are averages of the duplicate samples.
Laboratory analysis
Samples were analyzed for nitrite (NO2--N), nitrate (NO3--N), total
ammonia nitrogen (TAN; NH3+ NH4+), total Kjeldahl nitrogen (TKN),
orthophosphate (o-PO4), total phosphorus (TP), total Hg (THg, no
filtration), and dissolved Hg (DHg, filtration). Prior to analysis, all
samples were removed from the freezer and placed in a dark cabinet at room
temperature for approximately 18 h to melt. Once fully melted, the samples
were thoroughly mixed and dispensed into various aliquots for each analysis.
Subsamples of NO2--N, NO3--N, TAN, SO42-, and o-PO4 were
filtered through 0.45 µm filters (Pall
Supor®) prior to analysis. Laboratory filter
blanks were approximately < 2 µgL-1 for NO2--N,
6 µgL-1 for NO3--N, 5 µgL-1 for TAN,
< 20 µgL-1 for SO42-, and 2 µgL-1
for o-PO4.
Orthophosphate and TP were measured according to EPA Standard Method (SM)
365.1 and SM 365.1/USGS I-4600-85, respectively (EPA, 1993; USGS, 1985).
Method detection limits (MDLs) for these techniques were 0.60 and
0.63 µgL-1, respectively. Both techniques employed
colorimetric measurement with ascorbic acid. Prior to measurement of TP,
samples were digested with persulfate. Absorbance was then measured through
flow injection analysis (FIA; Rapid Flow Analyzer 300 equipped with an
Astoria-Pacific 305D high-sensitivity photometer detector; Alpkem, College
Station, TX).
Nitrite, NO3--N, and TAN analyses followed EPA SM 353.2 and SM 353.1
(EPA, 1979, 1993). Nitrite and NO3--N MDLs were
0.84 µgL-1, and the TAN MDL was 0.77 µgL-1.
Nitrite and NO3--N were measured by automated colorimetric analysis with
cadmium reduction being applied for the nitrate samples. Each sample was then
measured by FIA (Rapid Flow Analyzer 300 equipped with an Astoria-Pacific
305D high-sensitivity photometer detector; Alpkem, College Station, TX). TAN
samples were analyzed using automated phenate colorimetric techniques. Total
Kjeldahl nitrogen was analyzed using automated phenate block digestion
according to EPA SM 351.2. The MDL for TKN was
11.3 µgL-1. Organic N (bulk) was calculated as the
difference between TKN and TAN. All nitrogen species were reported as
[ µgL-1]-N with total N calculated as the sum of organic N,
TAN, and NO3--N. All snow sample NO2- concentrations were below the
detection limit (DL).
Sulfate was determined using a chromatography system (ICS 2000 with
Chromeleon version 6.6 software and AS14A column; Dionex Inc., Sunnyville,
CA) by EPA SM 300.0 (EPA, 1979). The MDL for SO42- was 19 µgL-1.
Total Hg and DHg were measured using a water analyzer (model 2600; Tekran
Inc., Toronto, Canada) according to EPA SM 1631
revision E (EPA, 2002). For DHg samples, approximately 50 mL of sample was
filtered through a 0.45 µm filter (Acrodisc syringe filter with
Supor® Membrane; Pall Corporation, Port
Washington, NY), while for THg, 50 mL of sample was poured directly into a
vial for analysis. Laboratory filter blanks were below the DL of the system
(< 0.3 ngL-1). Samples were preserved with 10 % bromine
chloride (BrCl) solution for storage until analysis the next day. Before
analysis, excess BrCl was neutralized with pre-purified hydroxylamine
hydrochloride. During analysis, samples were automatically mixed with
stannous chloride (SnCl2) in a phase separator; reducing oxidized Hg to
elemental Hg. Elemental Hg is then loaded onto two sequential gold traps by
an argon carrier gas. The Hg is then released through thermal desorption and
detected using atomic fluorescence spectrometry. The Tekran model 2600 was
calibrated using a NIST SRM-3133 Hg standard (with concentrations of 0, 0.5,
1.0, 5.0, 10.0, 25.0, and 50.0 ngL-1 Hg). System reliability was
checked using ongoing precision recovery injections of 5 ngL-1
throughout each run and ranged between 87 and 112 % recovery. Reagent
blanks measured regularly throughout each run ensured no contamination of the
system. DLs calculated as 3 times the standard deviation of the calibration
blanks, averaged 0.3 ngL-1 for all runs. Particulate Hg was
calculated as the difference between THg and DHg.
Statistics
We performed analysis of variance (ANOVA) for all chemical species using the
following independent variables: (i) year (n=2, 2011–2012 and
2012–2013), (ii) site elevation (n=3; low-, mid- and high-elevation
site), (iii) location (n=2; eastern and western basin), and season (n=2; early season (December through February) and late season (March and
April)). ANOVAs attribute variance of dependent variables to these various
independent variables and test their significance against the residual
variance. All relationships were considered statistically significant when
p values were ≤ 0.05.
Integrated snowpack concentrations were calculated by weighting each 10 cm
snow pit layer by its density. Seasonal wet deposition was calculated by
weighting all wet deposition samples by their volume up to the date of
sampling. Linear regression analyses were performed to test for correlations
between snowpack chemical concentrations, SWE, and elevation. All error bars
in figures represent standard error.
Basin-wide modeling with SWE reconstruction
Basin-wide loads and distribution were assessed using chemical concentrations
and loads measured throughout the 2011–2012 and 2012–2013 snow seasons as
well as basin-wide mean peak SWE estimates from SWE reconstruction for the
Sierra Nevada from 2000 to 2011 (Rittger, 2012). SWE reconstruction uses
estimates of snow energy balance with areal snow cover depletion from MODIS
Snow Covered Area and Grain size (MODSCAG) (Rittger et al., 2011). MODSCAG
calculates fractional snow cover area and grain size from MODIS data (Painter
et al., 2009). Compared with previous methods, MODSCAG has proven to give
reliable depletion rates throughout the spring season, when snowmelt is
highest (Rittger et al., 2013). Finally, the spatially refined MODSCAG data
set was combined with energy balance and temperature data to give accurate
reconstructed estimates of SWE throughout the Sierra Nevada, and specifically
the Lake Tahoe Basin. At the time of our study, SWE reconstruction data were
only available for 2000 to 2011, with no information from our sampling
seasons, 2011–2012 or 2012–2013. The 2000–2011 data set includes both
high- and low-accumulation snow years and gives a reasonable representation
of average snowpack accumulation in the Lake Tahoe Basin. In order to give an
estimate of average annual snowpack chemical storage, we applied the decadal
average peak SWE for 2000–2011 to our data (Fig. 2a). Estimates made during
this study were to establish relationships to previous estimates of the Lake
Tahoe nutrient budget and were not meant to represent a completely accurate
distribution or load stored within the basin's snowpack each year.
(a) Decadal average (2000–2011) peak SWE for the Tahoe
Basin from SWE reconstruction for the Sierra Nevada, and basin-wide peak snowpack
chemical loading estimates for (b) nitrogen, (c) total
phosphorus, and (d) total Hg.
Snowpack sampling throughout the Lake Tahoe Basin during 2011–2012 and
2012–2013 allowed for assessment of spatial and temporal chemical deposition
patterns. Specifically, relationships to wet or dry deposition, in-basin or
out-of-basin sources, and early- or late-season increases were identified.
These deposition and source controls were then related to orographic
characteristics to estimate chemical concentrations throughout the basin in
unknown areas. A GIS land-use layer of the Tahoe Basin (LTTMDL, 2010) was
applied in order to separate urban and non-urban locations with similar
orographic characteristics for urban-influenced species (i.e., TP). These
scaled concentrations were then applied to SWE reconstruction estimates to
determine total snowpack chemical loading throughout the entire basin.
Snowpack sampling occurred in open areas free of canopy coverage, but it is
possible that tree and plant particulate matter was still incorporated in
the snowpack. Litterfall contributions represent a form of chemical recycling
and will cause an overestimate of atmospheric contributions made during this
study. Visual inspection of snow samples, however, showed low contributions
of plant detritus in samples, and due to consistent forest types present
across the basin we would expect any additional plant-derived inputs to be
random and unbiased across sites.
Results and discussion
Spatial and temporal trends of snow accumulation and SWE
In the Lake Tahoe Basin, approximately 70 % of annual precipitation falls
during the winter and spring as snow (Fram and Belitz, 2011). The 2011–2012,
2012–2013, and 2013–2014 winter seasons were marked by relatively low snow
accumulation. Peak basin average snowpack storage (1 April) for the central
Sierra Nevada during 2011–2012, 2012–2013, and 2013–2014 was approximately
50, 53, and 41 % of the historical average (1951–present), respectively
(CADWR, 2014). Although peak SWE was similar in each season, the temporal
trends in snow accumulation and spatial distributions differed (Fig. 3). In
2011–2012, the Lake Tahoe Basin experienced low snowpack accumulation until
the middle of January, when a series of storms led to solid snow cover
throughout the basin. January storms were followed by a hiatus until late
February and March, when a series of storms brought peak basin average SWE up
to approximately 625 mm. The 2012–2013 snow year started earlier, with late
December storms bringing nearly 750 mm of SWE. Similar snowpack loading and
timing occurred across the Lake Tahoe Basin at sites with similar elevations
(e.g., Mt. Rose/Squaw Valley, Marlette Lake/Rubicon). Early season storms
were dominated by northerly wind patterns contributing substantial snowfall
in the northeastern areas of the Lake Tahoe Basin and reducing the typical
pattern of lower snow accumulation on the eastern side of the basin due to
the rain shadow effect of the Sierra Nevada crest (e.g., 2012–2013
Mt. Rose/Squaw Valley SNOTEL data). These early storms were followed by three
dry months with very little accumulation for the rest of winter. The
2013–2014 snow year experienced the lowest snow accumulation of all three
study years, with minimal snowpack development occurring until late-season
storms in March and April brought peak SWE storage up to approximately
575 mm. Minimal snowpack development occurred at lower lake level elevations
(e.g., Tahoe City SNOTEL data) throughout the entire 2013–2014 season.
Snow water equivalent measured in 2011–2012, 2012–2013, and
2013–2014 at select SNOTEL sites (NRCS, 2013) throughout the Lake Tahoe
Basin.
Analysis of variance results for 2011–2012 and 2012–2013 snowpack
concentrations. Controlling factors of year (n=2; 2011–2012,
2012–2013), elevation (n=3; high, middle, low), location (n=2; east,
west), and season (n=2; early, late) were investigated.
ANOVA results
TAN
NO3--N
Org. N
SO42-
TP
THg
DHg
(µgL-1)
(ngL-1)
Factor
d.f.
p value
Year
1
< 0.01**
1
0.05**
0.26
0.27
0.1*
< 0.01**
Elevation
2
0.25
0.19
0.15
0.68
0.06*
< 0.01**
0.02**
East/west
1
0.03**
0.55
0.93
0.22
0.03**
0.23
0.46
Early/late season
1
< 0.01**
0.58
0.23
0.75
0.36
0.12
0.65
* p value < 0.10, ** p value < 0.05.
Nitrogen
Nitrate (NO3-N)
Snowpack NO3--N concentrations ranged from 20 to 138 µgL-1 (n=49 cores), 14 to 98 µgL-1 (n=56 cores), and
28 to 62 µgL-1 (n=3 integrated snow pits) during
2011–2012, 2012–2013, and 2013–2014, respectively. These values were
comparable with previous measurements at the Emerald Lake Watershed, a remote
watershed in the southern Sierra Nevada (Williams et al., 1995). During
2011–2012 and 2012–2013 (i.e., the 2 years with detailed spatial and
temporal sampling), no distinguishable temporal or spatial pattern was
observed in either snowpack NO3--N concentrations or loads (Fig. 4).
ANOVA results confirmed that snowpack NO3--N concentrations were not
statistically affected by elevation, location (i.e., east/west), or early-
versus late-season sampling (Table 1). Comparisons of wet deposition and
integrated average snow pit concentrations during the 2013–2014 snow year
showed that snowpack NO3--N concentrations were similar to
volume-weighted wet deposition up to the date of snowpack sampling (Fig. 5).
This result is similar to patterns observed by Clow et al. (2002) and
Williams and Melack (1991a) and may be indicative of wet deposition as the
main source of NO3--N deposition. For example, wintertime deposition of
NO3--N in the Rocky Mountains was found to be highly correlated with
precipitation, with little difference between snowpack and NADP precipitation
volume-weighted mean concentrations, suggesting mainly wet deposition inputs
(Clow et al., 2002). Similarly, a study at the Emerald Lake Watershed
identified that dry deposition of NO3- was not an important contributor
of total NO3- load in winter snowpack (Williams and Melack, 1991a). Our
study revealed that increased precipitation on the west side of the Tahoe
Basin during 2011–2012 led to correspondingly greater NO3- loading;
however, little difference was seen during 2012–2013, when precipitation totals
throughout the basin were more uniform.
Average snowpack core concentrations during 2011–2012 (left) and
2012–2013 (right) snow seasons along with average SWE estimated from six
SNOTEL sites located within the Tahoe Basin.
Comparison of seasonal average volume-weighted wet deposition
concentrations with integrated snow pit samples from the 2013–2014 snow
year.
Vertical snow pit profile patterns show large variability in NO3--N
concentrations with depth, e.g., decreasing concentrations below the top
30–40 cm (Fig. 6). This variability suggests pronounced in-snowpack
dynamics possibly driven by conversion, vertical transport, or elution. In
addition, several studies have shown significant wintertime dry deposition of
NO3--N, in particular close to highways and urban areas (Cape et
al., 2004; Dasch and Cadle, 1986; Kirchner et al., 2005). Therefore, the fact
that wet deposition concentrations were very similar to snowpack
concentrations could be merely a coincidence and may not allow us to infer
dry versus wet deposition of NO3--N.
Previous studies have observed parallel concentration declines of SO42-
and NO3--N during snowpack melt events due to similar early-season ionic
pulses that lead to preferential losses of nutrients and other ions (Bales et
al., 1989; Harrington and Bales, 1998; Tranter et al., 1986). In support of
such potential losses, Fig. 4 shows decreasing snowpack NO3-
concentrations in spring months, particularly in the second year, 2012–2013,
when sampling captured the beginning of the melt season. Preferential
mobilization of solutes during melt events has also been shown to cause
downward movement of solutes in the snowpack (Williams and Melack, 1991b).
Our vertical snow pit samples show highly variable distribution patterns with
depth (Fig. 6), which may indicate insufficient temporal resolution of pit
sampling to detect vertical translocation. Similar early elution
characteristics have been observed for NO3- and SO42- (Stottlemyer
and Rutkowski, 1990; Williams and Melack, 1991b), and comparing
volume-weighted seasonal wet deposition concentrations of SO42- and
snowpack SO42- concentrations showed no large elution losses either
(Fig. 5). Our results suggest that Tahoe Basin snowpack NO3- is subject
to multiple inputs and complex in-snowpack processes, and that potential
losses (such as during early ionic pulses) may be difficult to detect against
additional surface (e.g., dry) deposition processes without very detailed
time- and depth-resolved snowpack measurements.
Year 2013–2014 snow pit profiles for nitrogen and mercury species
concentrations, snow density, and crystal form. Crystal classifications are
based on the ICSI classification for seasonal snow on the ground (Fierz et
al., 2009).
Total ammonia nitrogen (TAN)
Snowpack concentrations of TAN ranged from 16 to 104 µgL-1 (n=49 cores), 10 to 77 µgL-1 (n=56 cores), and
28 to 85 µgL-1 (n=3 integrated snow pits) during
2011–2012, 2012–2013, and 2013–2014, respectively. Snowpack TAN
concentrations are within the range of previous measurements made in the
Emerald Lake Watershed of California, where the amount of TAN deposited
within the seasonal snowpack accounted for approximately 90 % of annual
loading (Williams et al., 1995).
Unlike NO3--N, TAN is known to deposit through both wet and dry pathways
during winter (Clow et al., 2002; Ingersoll et al., 2008). In our study,
strong evidence for an important role of TAN dry deposition can be inferred
from the fact that snowpack TAN concentrations doubled from the early
(December–February) to the late (March–April) season in both 2011–2012 and
2012–2013 (Fig. 4). ANOVA results confirmed significant differences in
snowpack TAN concentrations between early- and late-season snowpack sampling
(Table 1, p=0.01). Increased late-season TAN concentration in snowpack is
consistent with similar observations in the Rocky Mountains and the Stubai
Alps (Bowman, 1992; Kuhn, 2001). These increases were attributed to the onset
of agricultural production in upwind valleys, as well as increased dry
deposition due to decreased atmospheric stability and increased convection.
Importantly, the late-season increase in snowpack TAN occurred in both years,
even though no significant late-season snowfall occurred in 2012–2013
(Fig. 4). The patterns of increasing TAN concentration in late-season
snowpack with no significant snowfall agree with previous research showing
dry deposition as the significant source of TAN deposition in the Sierra
Nevada (Bytnerowicz and Fenn, 1996).
Large increases in NH3 emissions from winter to spring have been measured
upwind of the Sierra Nevada in the San Joaquin Valley, CA, and were attributed
to increased agricultural and livestock activities (Battye et al., 2003).
Further support of snowpack TAN sourcing in the San Joaquin Valley was
higher concentrations at west basin sites than east basin sites during both
2011–2012 and 2012–2013. ANOVA results revealed a significant difference
between the east and west basin snowpack TAN concentrations (Table 1, p=0.03). This increase is likely due to the west basin sites being closer in
proximity to San Joaquin Valley agricultural activity, allowing for increased
transport and deposition.
During the 2013–2014 snow year, TAN concentrations were consistently higher
(up to a factor of 1.7) in integrated snow
pit samples than in volume-weighted wet deposition (Fig. 5; p=0.08; note low replicate of n=3). This increase in TAN further
emphasizes the importance of dry deposition of TAN to Tahoe Basin snowpack.
During snowpack storage, TAN is known to elute relatively late during melt
events (Kuhn, 2001); however, other transformations such as microbial
conversion can lead to decreases and losses throughout the season. Snow pit
depth profile sampling shows a decrease in TAN concentrations with depth and
therefore age (Fig. 6). This decrease coincides with increases in organic N
suggesting microbial conversion of inorganic N to organic N. Despite these
possible losses, the increase we observe between wet deposition and snow pit
concentrations indicates that the additional input of TAN from dry deposition
is large enough to exceed transformations that occur during snowpack storage.
Late season deposition doubled TAN snowpack loads prior to end-of-season
melt. The fate of snowpack TAN has been studied extensively through both
watershed mass balance and tracer-based research. For example, less than
1 % of TAN stored in snowpack at Emerald Lake, CA, reached the
lake as TAN during melt and runoff (Williams and Melack, 1991b). During a
later study, however, snowmelt with isotopically labeled NH3 was retained
in the soils during melt, making it a possible contributor to future NO3-
stream pulses after nitrification (Williams et al., 1996). Current
predictions show an increase in total N emissions during the next
half-century in the western United States due to large increases in
agricultural and livestock NH3 emissions (Fenn et al., 2003). Such
increased emissions could result in significant additional deposition loads
of TAN to snowpack in the Sierra Nevada with the potential to alter ecosystem
nutrient dynamics.
Organic nitrogen
Integrated snowpack organic N concentrations ranged from BDL (below detection
limit) to 211 µgL-1 in 2011–2012 (n=49 cores), BDL to
253 µgL-1 in 2012–2013 (n=56 cores), and 120 to
260 µgL-1 in 2013–2014 (n=3 integrated snow pit). No
dominant spatial or temporal patterns were observed in snowpack organic N
concentrations or loads for either 2011–2012 or 2012–2013 (Fig. 4). ANOVA
results supported this finding with no significant effects of location,
elevation, or early/late season on organic N concentrations (Table 1). A
previous study found large variation in wintertime deposition of organic N
throughout the Rocky Mountain Range, accounting for 40, 3, and 50 % of
total N in wet deposition during January, February, and March, respectively
(Benedict et al., 2013). Deposition rates and patterns of organic N are
difficult to quantify due to the large number of compounds – including
gaseous, particulate, and dissolved phases – originating from local,
regional, and global sources and subject to biological and chemical
transformations (Cape et al., 2011; Neff et al., 2002).
Overall, snowpack core samples collected during the 2011–2012 and 2012–2013
seasons showed high fractions of organic N, accounting for 49±17 % of
total snowpack N on average. Inorganic forms, TAN and NO3--N, accounted
for 21±10 and 29±10 %, respectively (Fig. 7). Research at a
high-elevation catchment in the Colorado Front Range identified organic N as
an important component in both wintertime wet deposition and stream export
(Williams et al., 2001), while data from a 14-year study (WY1985-1998)
in the Southern Sierra Nevada report that dissolved organic
nitrogen (DON) accounted on average for 35 % of total N (NH4++ NO3-+ DON)
in winter precipitation (Sickman et al., 2001). Comparison of volume-weighted
wet deposition and integrated snow pit concentrations showed higher
concentrations (up to a factor of 2.5) of organic N in the snowpack (Fig. 5).
Two possible sources could cause higher concentrations of organic N in
snowpack compared to wet deposition: snowpack microbial conversion of
inorganic N to organic N and dry deposition of organic N during storm-free
periods (Clement et al., 2012; Jones, 1999; Williams et al., 2001). Our
measurements do not allow for differentiation between the two possible
sources of snowpack organic N; however snow pit profile sampling shows
coinciding decreases in inorganic N and increases in organic N with snow pit
depth and therefore age (Fig. 6). One Arctic snowpack study found that
microbially based N cycling was a dominant process explaining N species
availability at the base of the snowpack (Larose et al., 2013). We suggest
that microbial uptake of inorganic N may be a primary driver of the
increasing snowpack organic N levels during storage. Overall, we observed
that the dominant form of N in Sierra Nevada snowpack during our study was
organic N, and we propose that this large representation warrants detailed
studies in regard to the sources, cycling, and fate of organic N in the
Sierra Nevada.
Concentrations and loads of total N in snowpack are apparently dependent on
contributions of both inorganic and organic forms, with respective
differences in deposition pathways (wet versus dry deposition), potential
conversion processes (e.g., from inorganic to organic forms), and different
mobilization during elution sequences leading to large fluctuations in both
the concentration and spatiotemporal patterns of snowpack total N
throughout the season. Total N accumulation in Sierra Nevada snowpack shows
strong interannual variability as well as different representation of various
N species.
Snowpack total N distribution for 2011–2012 (left) and 2012–2013
(right).
Average snowpack total phosphorus concentrations at all lake-level
sites. The Incline and Thunderbird sites are located on the east side of the
basin in urban and remote settings, respectively, and the Homewood Low site
is located on the west side of the basin in an urban setting.
Phosphorus
Snowpack TP concentrations ranged from 3 to 109 µgL-1 in
2011–2012 (n=49 cores), 3 to 59 µgL-1 in 2012–2013 (n=56 cores), and 10 to 41 µgL-1 in 2013–2014 (n=3 integrated snow pits). Figure 8 shows that the urban site in Incline
Village at lake level had by far the highest snowpack TP concentrations,
ranging up to 6 times higher than any other snowpack concentration at similar
elevation (i.e., lake level). In comparison, the Thunderbird site, also at
lake level, located in a very remote setting just 10 km from Incline, had
much lower P concentrations. Sources such as fugitive dust from plowing,
forest and agriculture biomass burning, and diesel engine combustion have
been identified as major sources of particulate-phase atmospheric P in
California (Alexis et al., 2001). Specifically, in the Lake Tahoe Basin, road
dust has been identified as a primary contributor of P input into Lake Tahoe
(Dolislager et al., 2012), while another study found significant P emissions
from urban biomass burning (Zhang et al., 2013). Our patterns suggest that
urban areas in the Lake Tahoe Basin are a major source area for P deposition
to snowpack during winter and spring.
Local and regional emissions are also relevant at larger scales, as evident
in 2011–2012, when remote sites at eastern locations in the basin showed
higher TP concentrations than western sites. We propose that the large
concentration of urban source sites at lake level combined with the dominant
west to east wind pattern led to increased deposition on the east side of the
basin. During 2012–2013, no west-to-east increase in TP concentration was
observed; however, the strong influence of urban activity remained. It is
unlikely that sources of P in the basin changed between 2011 and 2012 and
between 2012 and 2013, and it is more likely that different deposition
patterns due to differences in snow accumulation, timing, and storm track
directions caused this change. Even though there was significantly higher P
deposition on the east side of the basin from urban influence, the relatively
remote west basin snowpack still had TP concentrations of
11.8 µgL-1 on average. Diffuse regional P sources to the
Tahoe Basin include both dust and aerosol inputs. Particulate matter
particles smaller than 10 µm in diameter (PM10) are capable
of long-range transport, while larger particles have higher deposition
velocities and decreased transport (Vicars et al., 2010). Specifically,
dust-derived inputs originate from geologic sources and erosion from both
agricultural and urban activity, while burning from both forest and domestic
fires contributes additional particulate matter in the form of ash and soot
(Raison et al., 1985). Differences in P deposition rates between the dry and
wet seasons as well as spatial patterns associated with wind direction and
soil erosion vulnerability have been observed in the southern Sierra Nevada;
Ontario, Canada; and the Mediterranean (Brown et al., 2011; Morales-Baquero
et al., 2006; Vicars and Sickman, 2011).
Comparison of volume-weighted wet deposition and integrated snow pit
concentrations showed higher levels of TP (up to a factor of 5.8) in snowpack
than wet deposition (Fig. 5). This increase further supports dry deposition
as a primary input of snowpack P. Finally, snowpack o-PO4, the most
bioavailable form of P (Dodds, 2003), accounted for 34±15 % of
snowpack TP, similar to previous work in the Lake Tahoe region that estimated
approximately 40 % of TP in atmospheric deposition was in a bioavailable
form (LTTMDL, 2010).
Low P levels in parent material make high-elevation watersheds of the Sierra
Nevada, sensitive to the effects of external P inputs (Melack and Stoddard,
1991; White et al., 1999). Further research, however, has shown that
extractable P levels of parent material strongly influence P adsorption. The
very high extractable P levels in granitic soils in the Sierra Nevada lead to
low P adsorption potentials, while the low extractable P levels and
sesquioxide content of volcanic soils in the Sierra Nevada increase
adsorption (Johnson et al., 1997). Approximately two-thirds of the Lake Tahoe
Basin parent material is granitic and one-third is volcanic (LTTMDL, 2008),
making soil adsorption potentials of atmospherically deposited P throughout
the watershed highly variable with location. Along with N, P levels directly
control algal production within aquatic ecosystems, and algal production is a
key reason for declining clarity in Lake Tahoe (Dolislager et al., 2006). In
particular, the high snowpack concentrations at urban locations near the lake
may cause a significant influx of P into Lake Tahoe during melt.
Mercury
Snowpack THg concentrations ranged from 0.81 to 7.58 ngL-1 in
2011–2012 (n=49 cores), 0.97 to 5.96 ngL-1 in 2012–2013
(n=56 cores), and 3.28 to 7.56 ngL-1 in 2013–2014 (n=3 integrated snow pits). The Tahoe Basin average snowpack core THg
concentration for 2011–2012 and 2012–2013 was 2.56±1.3 ngL-1. Observed THg concentrations are slightly lower, but
within range of the end-of-season average snowpack concentration measured
during a watershed Hg balance study in 2009 at Sagehen Creek, CA (i.e.,
3.3 ngL-1; Fain et al., 2011), a remote watershed located only
32 km north of the Tahoe Basin. Particulate Hg was the dominant form of Hg
within Tahoe snowpack, accounting for 76.1±8.7, 70.3±13.4, and
87.1±4.7 % of THg on average during 2011–2012, 2012–2013, and
2013–2014, respectively. The large percentage of particulate Hg in the
snowpack agrees with previous findings from a study in Canada that saw a
post-depositional increase in particulate associated Hg from approximately 50
to 70 % (Poulain et al., 2007). This study attributed particulate
throughfall and photochemically induced emission as the main causes of the
speciation shift and also noted strong differences in snowpack Hg
concentrations between open and forested areas which were attributed to
throughfall contributions from tree canopies as well as shading reducing
photochemical evasion.
Snowpack coring revealed no dominant temporal or spatial patterns in THg or
DHg deposition, with ANOVA results showing no significant effects of season
(i.e., early versus late) or location (i.e., east versus west; Table 1). The
lack of spatial trends suggests global background atmospheric pollution,
rather than specific point sources such as urban areas, to be the main source of
snowpack Hg in the Lake Tahoe Basin. Mercury's long atmospheric lifetime and
global circulation allow for diffuse deposition to this relatively remote
mountain region (Fain et al., 2011; Schroeder and Munthe, 1998), and the
majority of large snowfall events in the Sierra Nevada originate as
large-scale convection cells in the eastern Pacific and travel hundreds of
kilometers before reaching the Tahoe Basin (O'Hara et al., 2009). To our
knowledge, few point sources for Hg emission exist within the Lake Tahoe
Basin, although one study within the basin reported that significant amounts
of particulate Hg are emitted from wildfires (Zhang et al., 2013) and found
increased levels of particulate Hg in urban areas of the Lake Tahoe Basin.
Both THg and DHg concentrations in the snowpack significantly increased with
elevation in the basin (Table 1; p< 0.05). This finding is in contrast to
an expected “washout effect”, which causes declines in Hg precipitation
concentrations with storm duration and magnitude (Poissant and Pilote, 1998).
King and Simpson (2001) observed that approximately 85 % of photochemical
reactions occur in the top 10 cm of the snowpack. It is possible that the
increase in Hg concentration with elevation is due to decreased light
penetration relative to snowpack depth and reduced photochemical re-emission,
as increased elevation leads to the formation of a deeper, denser snowpack.
In support of this notion is a significant positive correlation between
integrated snowpack THg concentration and total SWE (slope: 0.002
[ngL-1SWE(mm)-1]; p value: < 0.05), as well as strong
elevation gradients in total snowpack Hg pools. In agreement, total snowpack
Hg loading was significantly higher in 2012–2013 than in 2011–2012
(Table 1; p< 0.01) in accordance with higher overall SWE. Evidence for
surface-based photochemical losses of Hg are lower concentrations of Hg in
upper snowpack layers (Fig. 6). Declines in Hg concentrations between
cumulative wet deposition and integrated snowpack content were mainly driven
by DHg, with up to 4.5 times lower concentrations observed in integrated snow
pit samples than volume-weighted wet deposition (Fig. 5). Aside from
photochemical losses, it is possible that vertical patterns are co-determined
by vertical movement and solute transport of Hg. Previous studies have
reported Hg pulses in runoff during snowmelt events (Schuster et al., 2008).
In addition, sorption processes could lead to conversion between DHg and
particulate Hg and changes snowpack Hg speciation. The combination of strong
precipitation gradients and increased THg concentration with SWE leads to
large spatial variability in the total snowpack Hg pools in mountainous
areas. A previous study noted relationships between soil Hg content and
elevation (Gunda and Scanlon, 2013), possibly attributable to precipitation
gradients, while another study found that soil Hg storage was positively
correlated with total precipitation across multiple study sites but attributed
these effects to ecological processes such as increased plant productivity
and carbon accumulation (Obrist et al., 2009, 2011).
Photochemical reduction and volatile re-emission of gaseous Hg during snowpack
storage has been widely studied and is known to account for losses of up to
50 % from that measured in initial deposition (Fain et al., 2007, 2011;
Lalonde et al., 2002; Mann et al., 2011; Poulain et al., 2007). However, at
the end of the season, we still observed substantial concentrations of Hg
left in the snowpack (e.g., ranging from ∼ 55 to ∼ 105 % of
volume-weighted wet deposition) that will be subject to melt and infiltration
into the watershed. In addition to the declines of DHg during storage, an
increase in particulate Hg was observed in two of the three comparisons of
snow pit and wet deposition samples (Fig. 5), and it is possible that
photochemical losses from snow are in part offset by gaseous dry deposition
and particulate throughfall during storm free periods.
A study at the nearby Sagehen Creek, CA, watershed quantified that
only 4 % of total annual Hg wet deposition was exported from the
watershed in stream water and identified soil uptake and storage as well as
photochemical re-emission as the major sinks of atmospherically deposited Hg
(Fain et al., 2011). While soil uptake serves as a buffer delaying the
transport of upland wet deposition to streams, sediment core analyses still
showed that upland watershed contributions (i.e., through soil erosion and
sediment flux) are significant contributors of Hg input to lakes even under
relatively low watershed-to-lake-area ratios as in the Lake Tahoe Basin
(extrapolated to 42 % contributions when using relationships presented by
Lorey and Driscoll, 1999). Snowpack-based Hg input to the watershed,
therefore, is expected to contribute to lake water quality through erosion
and sediment-based influx, albeit delayed in time and closely linked to soil
Hg pools and mobilization.
Basin-wide loading estimates
Declines in Lake Tahoe water quality have been observed during the last
50 years (Sahoo et al., 2010; Schuster and Grismer, 2004). Specifically,
Secchi depths, a measure of lake transparency, have decreased from
approximately 30.5 to 21.3 m since the 1960s (TERC, 2011). Eutrophication
from atmospheric and terrestrial nitrogen (N) and phosphorus (P) inputs and
light scattering by particulate inputs are the main causes of this decline
(Jassby et al., 2003; Swift et al., 2006). Most previous studies in the Lake
Tahoe Basin have focused on direct atmospheric deposition to the lake surface
(Dolislager et al., 2012; NADP, 2012), and little information is available on
snowpack-based loading for the surrounding upland watershed. The surrounding
land surface covers 814 km2 of the 1310 km2 Lake Tahoe
watershed. Direct atmospheric inputs to the lake surface are estimated to
contribute 55 and 15 % of total N and P, respectively (TERC, 2011).
Stream monitoring data show that, upon snowmelt, Lake Tahoe receives large
pulses of N and P (Goldman et al., 1989; Hatch et al., 1999), which together
control algal production within the basin's aquatic ecosystems contributing
to the decline in clarity in Lake Tahoe during the last 50 years (Dolislager
et al., 2006). Although much of snowpack-based chemical loads may not
directly enter Lake Tahoe upon melt, snowpack loads are important for
terrestrial chemical budgets. For example, nutrient-rich O-horizon runoff –
measuring as high as 87.2 mgL-1 NH4-N,
95.4 mgL-1 NO3-N, and 24.4 mgL-1 PO4-P –
has been observed in Lake Tahoe forests during snowmelt events due to
leaching from the forest litter layer (Miller et al., 2005). In order to
relate peak snowpack nutrient and pollutant loading to previous terrestrial
and lake chemical budgets, we here estimate average peak basin-wide snowpack
chemical storage using the peak SWE decadal average from 2000 to 2011
(Fig. 2a). While canopy effects on total snow accumulation are incorporated
in this estimate through the SWE reconstruction model, we did not include
forest canopy effects on deposition and chemical dynamics, as our snowpack
measurements were limited to open, canopy-free locations. Deposition and
snowpack dynamic processes in forests are known to show substantial
differences compared to canopy-free locations, including increased dry
deposition, throughfall deposition, or different photochemical processes
(Poulain et al., 2007; Tarnay et al., 2002). In order to be able to compare different locations across the basin, we chose non-forested sites. The estimated deposition loads, therefore, are based on
deposition and snowpack storage measured in canopy-free locations and could
be different when effects of canopies and other forest processes are
incorporated.
Nitrogen
Snowpack NO3--N loading was highly dependent on snow accumulation, but
concentrations showed no significant temporal or spatial trends throughout the Lake Tahoe Basin (Table 1). To
calculate basin-wide NO3--N loads, we therefore multiplied the 2-year
seasonal average concentration (47.1,µgL-1) by the decadal
average reconstructed SWE. Basin-wide NO3--N loading estimates
(massarea-1) thus reflect snowpack accumulation patterns (i.e.,
SWE) with the highest loading occurring on the west side of the basin at high
elevations, up to approximately 1 kgha-1, and decreasing toward
the east and with lower elevations due to lower SWE accumulation. Average
annual snowpack NO3--N storage for the Lake Tahoe Basin is estimated to
be 28.7 t.
Unlike NO3--N, snowpack TAN loading in the Lake Tahoe Basin showed strong
spatial and temporal trends. Late season deposition effectively doubled
snowpack TAN concentrations, with significantly higher concentrations on the
west side of the basin than the east side (Table 1). Due to these
relationships, we applied the March and April (peak SWE generally occurs
during March and April in the Lake Tahoe Basin) average snowpack TAN
concentration from the east and west basin sites to the reconstructed SWE
data (western sites: 57.9 µgL-1, eastern sites:
41.6 µgL-1) to scale up snowpack TAN loading to the whole
basin. Modeled estimates, therefore, show greater TAN accumulation on the
western side of the basin, with highest loading occurring at high elevations
in the west (up to approximately 1.2 kgha-1) due to the
combination of both large snow accumulation and proximity to upwind sources.
Our estimate of average annual basin-wide accumulation of TAN within the
basin's snowpack is 30.5 t.
Snowpack organic N concentrations throughout each sampling season were
variable and showed no clear temporal or spatial trends (Table 1). Applying
the average concentration of 88.7 µgL-1 from all snowpack
samples throughout both years produced an annual estimate of 54.1 t of
organic N stored within the basin's snowpack.
Average annual snowpack N storage for the Lake Tahoe watershed – calculated
as the sum of NO3--N, TAN, and organic N – totaled 113 t (Fig. 2b).
Inorganic and organic forms made up 52 and 48 % of total N, respectively.
TAN and NO3--N accounted for 27 and 25 % of total snowpack N,
respectively. Annual N loading estimates for Lake Tahoe (from terrestrial
runoff and direct atmospheric deposition) were previously estimated to be
397 tyr-1, with 218 tyr-1 originating from
atmospheric sources depositing directly on the lake's surface (LTTMDL, 2010).
With the caveat that estimation methods differed, snowpack N storage
estimates from our study represent approximately 28 % of the lake's total
N budget. Comparing our estimates to the 179 t of N that originates from
runoff and terrestrial sources, annual snowpack N storage would replenish
approximately 63 % of this flux.
Phosphorus
Snowpack P accumulation in the Lake Tahoe Basin was strongly related to
proximity to urban sources, as well as transport along the dominant westerly
winds throughout the basin. This dependence caused the highest
P concentration in the snowpack to occur in developed areas and higher
concentrations across east basin sites than remote west basin sites
(Table 1). Applying different P concentrations based on degree of
urbanization (see Sect. 2.5), highest P loading (up to approximately
0.4 kgha-1) therefore occurs at high elevations with significant
impacts of urban emissions (i.e., northeastern and southern locations
influenced by Incline Village, NV, and South Lake Tahoe, CA). The basin-wide
average TP storage estimated during this study of 0.11 kgha-1 is
more than double the average snowpack storage reported for the
Emerald Lake Watershed (0.04 kgha-1; Sickman et al., 2003) and reflects increased
urbanization within the Tahoe Basin. Homyak et al. (2014) estimate that
atmospheric deposition has contributed up to 31 % of P accumulation and
loss in soils and runoff since deglaciation of the Emerald Lake Watershed.
The higher snowpack loading rates estimated during this study indicate that
atmospheric deposition could be the primary supplier of excess P input to the
Tahoe Basin.
Overall, we estimate a peak P load of approximately 9.3 t of P stored
annually in Lake Tahoe Basin snowpack (Fig. 2c). Previous pollutant loading
studies for Lake Tahoe have estimated that approximately 46 t of P enters the
lake each year, with approximately 39 t of the annual budget originating from
land-based sources (LTTMDL, 2010). Annual snowpack TP storage estimates could therefore represent approximately 20 % of total P input into Lake
Tahoe each year.
Mercury
Similar to NO3--N, snowpack THg concentrations showed little temporal or
east-to-west variation (Table 1). However, THg concentrations were positively
related to total SWE (slope: 0.00201 [ngL-1mm-1]; p value:
0.016). Applying this relationship to reconstructed SWE data produced the
following THg distribution throughout the Lake Tahoe Basin (Fig. 2d): THg
loading throughout the basin followed strong elevation gradients, with the
uppermost areas of the basin receiving the highest concentrations and total
loading (up to approximately 125 mgha-1) due to increased snow
accumulation. Average annual snowpack THg concentration and loading for the
Lake Tahoe watershed was 3.6 ngL-1 and 30 mgha-1,
respectively, based on the decadal SWE accumulation average of 750 mm. We do
not have any previous data on Hg deposition to this basin, but these values
are comparable to the 3.3 ngL-1 average snowpack Hg
concentration and 13 mgha-1 peak snowpack loading from the
Sagehen Creek watershed in 2009, when snowpack accumulation was approximately
400 mm (Fain et al., 2011). The basin-wide estimate of THg stored within the
annual snowpack was 1166.2 g. Snow-based Hg fluxes estimated during this
study fall within range of measurements
(3.36–36 mgha-1yr-1) taken at seven national parks
throughout western North America during the Western Airborne Contaminants
Assessment Project (WACAP), which found fish Hg levels above the human
consumption threshold even at sites with relatively low Hg deposition
(Landers et al., 2008).
Conclusions
In summary, spatial and temporal pattern analyses suggest that
out-of-basin sources were important for Hg and TAN, while in-basin sources
controlled P deposition, with the highest concentrations measured near urban
areas, exceeding remote location concentrations by up to a factor of 6.
Snowpack NO3--N concentrations were relatively uniform throughout the
basin, indicating out-of-basin sourced wet deposition as a primary input;
however, high variability in snow pit vertical concentrations suggests
additional inputs and in-snowpack transport and conversion processes. Second,
increased NH3 emissions from the San Joaquin Valley and increased
atmospheric vertical mixing during the onset of spring likely led to dry-deposition-based increases in snowpack TAN during March and April,
effectively doubling snowpack TAN concentrations prior to melt. Third,
chemical speciation showed that organic N in the Lake Tahoe snowpack
accounted for 48 % of total N on average, with possible microbial
conversion leading to higher enhanced organic N levels in deeper, older
snowpack. Fourth, particulate Hg was the dominant form of Hg (78 % on
average) within Tahoe snowpack, and concentrations of both THg and DHg
increased with elevation and SWE likely due to decreased light penetration
and reduced photochemical re-emission in deeper snowpack. Finally, basin-wide
modeling estimates indicated that Lake Tahoe Basin snowpack acts as a
substantial reservoir in which atmospheric nutrients and pollutants
accumulated throughout winter and spring. Estimates of basin-wide annual
snowpack mass loading showed accumulation of N, P, and Hg yielding 113 t of
N, 9.3 t of P, and 1166.2 g of Hg. Further research should focus on
quantifying the relationship between snowmelt processes and stream and
groundwater input, and address the substantial amount of organic N stored
within the basin's snowpack.