Project at a Glance

Title: Comparative analyses of high-altitude lakes and catchments in the Sierra Nevada: susceptibility to acidification

Principal Investigator / Author(s): Melack, John

Contractor: UC Santa Barbara

Contract Number: a032-188

Research Program Area: Atmospheric Processes

Topic Areas: Acid Deposition, Chemistry & Reactivity, Ecosystem Impacts, Impacts


The atmospheric deposition and surface-water chemistry of eight seasonally snow-covered catchments in the Sierra Nevada was measured from 1983 through 1994 to assess watershed susceptibility to acidic atmospheric deposition. Catchments and years included the following: Emerald Lake, 1983-1994; Pear Lake, 1986-1993; Topaz Lake, 1986-1993; Ruby Lake, 1986-1994; Crystal Lake, 1986-1993; Spuller Lake, 1989-1994; Topaz Lake, 1986-1993 and the Marble Fork of the Kaweah River, 1992-1994. The catchments are located in the alpine and subalpine zones of the Sierra Nevada of California, and their geographic locations span a majority of the north-south extent of the range. Four of the watersheds are located along the eastern slope of the range (Ruby Lake, Crystal Lake, Spuller Lake, Lost Lake), and the remainder are situated along the western slope (Emerald Lake, Pear Lake, Topaz Lake and the Marble Fork of Kaweah River). Seven of the catchments are glacial cirques, ranging in size from 25 to 441 ha, that contain lakes. The upper Marble Fork drains the Tokopah Valley of Sequoia National Park, a glacially carved basin of 1,900 ha which includes, within its boundaries, the watersheds for Emerald, Pear and Topaz lakes along with several other small lakes and ponds.

Atmospheric deposition of water and solutes was determined by collections of non-winter precipitation (April through October) and by sampling the snowpack in the spring. Lakes and outflow streams in these catchments were sampled for chemistry year-round. Outflow samples were collected on a biweekly to daily basis during snowmelt runoff and lakes were sampled ca. six times per year. Outflow discharge was gauged continuously. All major solutes in atmospheric deposition and runoff (pH, ammonium, calcium, magnesium, sodium, potassium, chloride, nitrate, sulfate and acid neutralizing capacity (ANC)) were measured. A rigorous quality assurance - quality control protocol was followed. Meteorological conditions were monitored in two of the catchments (Emerald and Spuller).

The quantity and timing of snowmelt affected annual and interannual variability of surface-water chemistry in seasonally snow-covered catchments in the Sierra Nevada. General patterns of surface-water chemistry were identified using statistical analysis, however; there was considerable variation in these patterns among the watersheds. In most cases, pH decreased as runoff increased, reaching a minimum near the peak of snowmelt runoff. Several other pH patterns were observed: (1) pH increased as discharge increased, (2) pH reached a maximum at peak runoff, (3) pH changes were unrelated to changes in discharge and (4) pH remained fairly constant despite large changes in discharge.

Temporal changes for most other solutes demonstrated one of three different patterns: dilution, pulse/dilution or pulse/depletion. Acid anions such as nitrate and sulfate often increased in concentration in early snowmelt, with nitrate becoming depleted (i.e., analytically undetectable) and sulfate declining at peak runoff. In most catchments, nitrate peaks of between 5 and 15 µEq L-1 were common; in nitrogen limited lakes, e.g., Crystal and Lost, nitrate peaks during snowmelt were usually less than 2 µEq L-1. However, nitrate concentrations in the Topaz Lake catchment, were often highest in the winter, prior to snowmelt, and declined as runoff increased in the spring. Sulfate patterns were qualitatively similar to nitrate, but the magnitude of the changes was smaller. Differences in sulfate maxima and minima were less than 1 to 2 µEq L-1 in most cases. In catchments with considerable groundwater and sulfur bearing bedrock i.e., Spuller and Ruby lakes, sulfate declined by 10 to 20 µEq L-1 over the course of snowmelt.

Base cations and ANC most commonly exhibited a dilution pattern: concentrations declined as snowmelt runoff increased, with minima occurring near peak runoff. Depressions of ANC usually began in the early stages of melt. Outflow ANC, declined by 25% - 80% over the course of the spring; the average decline was about 50%. ANC minimums ranged from ca. 15 to 30 µEq L-1, Lost and Pear lakes had the lowest minima while ANC minima at Ruby and Crystal lakes were highest. In all catchments, ANC depression was greatest during years with deep snowpacks and high snowmelt runoff.

No long-term trends in pH or ANC were identified in surface waters of the Sierra Nevada during the period of 1983 through 1994. Trends were detected in other solutes which suggests that Sierran ecosystems are potentially sensitive to increased nutrient loading and climatic perturbations. Drought conditions in the Sierra Nevada probably were responsible for increasing the proportion of runoff derived from shallow groundwater in the Ruby Lake basin as evidenced by an increase in sulfate concentrations from ca. 6 to 12 µEq L-1 during the period of 1987 through 1994. Drought may also be partially responsible for increased retention of N in the Emerald Lake catchment. Long-term monitoring has revealed a 25 to 50% reduction in annual nitrate maxima and minima at Emerald Lake, with a concomitant shift of the lake's phytoplankton community from phosphorus limitation towards nitrogen limitation. These findings support recent evidence that N uptake in alpine catchments may increase due to climate warming and runs counter to the recent shift of Lake Tahoe to P limitation of phytoplankton. At Emerald Lake, decreases in nitrate represent a shift from Stage 2 to Stage 1 of nitrogen saturation and is a counter example to lakes and streams in the Front Range of the Rocky Mountains. Based on nitrate peaks during snowmelt and nitrate levels in snow, most of the catchments in our study are experiencing Stage 1 symptoms of N-saturation.

Dilution was the primary factor in ANC depression in surface waters during our study. The lakes could be divided into two classes based on their response to snowmelt: shallow, short residence time (i.e., rapidly flushed) lakes where acidification accounted for < lO% of the ANC decrease (Lost, Topaz, and Spuller); and lakes where acidification caused 25 to 35% of the depression due to larger lake volumes or lower snowmelt rates (Emerald, Pear, Ruby and Crystal). In lakes where acidification was important, nitrate and sulfate contributed equally during the first half of snowmelt, while sulfate dominated in the latter half.

The relationship between minimum and fall-overturn ANC for the lakes in this study was linear (r2 = 0.84), and the equation remained unchanged as additional data from earlier synoptic surveys were added. This linear model, applied to Western Lakes Survey data for the Sierra Nevada, estimated that no lakes are currently acidified by snowmelt. However, the model's confidence limits allow for the possibility that up to 1.8% (~38) of Sierra Nevada lakes undergo snowmelt ANC depressions slightly below 0 µEq L-1.

The hydrology of high elevation catchments in the Sierra Nevada is dominated by the accumulation and melting of the winter snowpack. The majority of catchment outflow occurs during the annual snowmelt period which begins as early as late March following dry winters in the western Sierra, but may not start until early May in the eastern Sierra Nevada when snowpacks are deep. Peak runoff occurred in early to mid June when winter snowfall was light and during late June and early July in wet years (e.g., 1993). At some catchments, peak outflow discharge following a wet winter was much greater than in dry years (e.g., Lost and Topaz lakes) while in others the range of peak flows was relatively small (e.g., Ruby and Crystal lakes).

Snowmelt from seasonally snow-covered catchments in the Sierra Nevada was often punctuated by periods of low discharge caused by spring snowstorms that cooled air temperatures and lowered the rate of snowmelt for several days at a stretch. Following peak discharge, runoff receded gradually during the summer and autumn. However, groundwater storage and release in the Ruby Lake basin are partly responsible for maintaining year-round outflow. In most catchments, outflow streams went dry by September when snow-cover was gone indicating that little water is stored in most Sierran watersheds. With the exception of the Crystal Lake basin, the catchments in the study lost little water via subsurface flow and were hydrologically tight. Winter runoff at all catchments was very low. At basins without groundwater inputs, winter streamflow was primarily the result of displacement of lake water by snowfall and, to a lesser extent, from winter snowmelt from south facing slopes.

Only a small percentage of the settled snowpack is lost to evaporation in the Sierra Nevada. The typical evaporative loss from snow varies from 80 to 100 mm of water, approximately six percent of the average maximum accumulation- in drought years the loss is less, but represents a greater percentage of the maximum accumulation, 9 to 13%. During peak snow years the loss decreases to about four percent, although the extended length of the snow season somewhat increases the actual amount lost. Most of the loss occurs during the period of snow accumulation when vapor pressure differences are usually favorable for evaporation and conditions of atmospheric instability are often found. Even so, low vapor pressure differences between the air and the snow surface, the result of cold temperatures and the prior transit of the over-passing air over extensive snow-covered distances (which increase humidity and decrease the capacity to absorb additional moisture), limit total evaporation to relatively small quantities.

During snowmelt, stable atmospheric conditions and reduced vapor pressure differences, due to the higher vapor content of the warming spring air, reduce evaporation from snow to negligible amounts. Often near the end of this period, snowpack evaporation losses are exceeded by gains from condensation. Total evaporation from the snowpack during snowmelt is typically around 15 mm of water: two to three percent of the maximum accumulation during drought years, less than one percent for above normal snow years.

Snow chemistry was dilute and similar among the eight study catchments. Samples from the spring snowpack had pH levels typically between 5.3 and 5.6, with an overall mean value of 5.42. After hydrogen-ion, the most abundant ions in solution were compounds of nitrogen, ammonium and nitrate, with mean concentrations of 2.7 and 2.4 µEq L-1, respectively. Sulfate concentrations in winter snow were slightly lower: 1.0 to 3.0 µEq L-1 (overall mean, 2.0 µEq L-1). Of the remaining solutes, only calcium and sodium were found in levels much above the detection limit (mean values: 1.7 and 1.3 µEq L-1, respectively). Organic anions (acetate and for-mate) were difficult to measure in snow and were usually found at low concentrations (mean concentrations ca. 0.5 µEq L-1). The solutes that showed the most variability among years and among sites were potassium and the organic anions. The most consistent solutes in snow during the study were hydrogen-ion and sulfate. Mean annual snowfall during the study was 1,027 mm of water equivalence.

Non-winter precipitation (i.e., rain and snow from ca. April through November) varied considerably with respect to chemistry and quantity. Solute concentrations ranged from near detection limits to tens of microequivalents per liter. Most of the variability was due to the timing of the precipitation: spring and autumn storms were the most dilute and summer rains were relatively enriched with solutes. Thus, the annual volume-weighted mean concentration of non-winter precipitation depended on the mix of samples obtained. The mean pH of non-winter precipitation in the study was 4.93. Ammonium and nitrate concentrations were eight to nine times greater in non-winter precipitation than in winter snowfall (mean values, 23.4 and 20.7 µEq L-1, respectively). Ammonium levels were usually higher than nitrate, with both nitrogen ions exceeding the mean concentration of sulfate (15.1 µEq L-1). The mean chloride level measured in non-winter precipitation was 4.2 µEq L-1 which was only slightly higher than the mean concentration in winter snowfall. After ammonium and hydrogen-ion the next most abundant cations in non-winter precipitation were calcium (mean, 10.4 µEq L-1) and sodium (4.6 µEq L-1). In contrast to winter snow, organic anions were abundant in non-winter precipitation. Mean values for acetate and formate were on the order of seven to nine µEq L-1. The average annual input of water during non-winter periods of our study was 117 mm.

Precipitation intercepted by Sierran catchments is greatly altered by geochemical processes before exiting the catchments as streamflow. High rates of nitrogen deposition were measured in the study catchments (mean annual nitrogen deposition was 95.6 Eq ha-1). Biological processes and other sinks within the watersheds consumed the large majority of these nutrient inputs. Ammonium was rarely found at detectable levels in outflow streams in our study. Nitrate concentrations were typically higher, but in most cases the input-output budgets showed a net retention of nitrate in the catchments. During most years, the majority of nitrogen deposition occurred during non-winter periods.

Hydrogen-ion deposition was also substantial in these catchments and winter snowfall was the main contributor. These inputs were effectively neutralized by the catchments and, on average, the basins consumed 87% of the hydrogen-ions deposited in them. Catchments in the eastern Sierra Nevada had significantly (p < 0.05) higher outflow ANC and neutralized a higher percentage of acid inputs than did basins in the Tokopah Valley or the Lost Lake watershed. These findings suggest that watersheds along the eastern slope of the Sierra Nevada may be less susceptible to harm from acid deposition than catchments along the western slope. Processes, principally mineral weathering and biological uptake of nitrogen, changed the chemical makeup of precipitation so that streamwaters exiting the catchments were a solution composed primarily of ANC (i.e., HC03- or bicarbonate), calcium and dissolved silica.

For questions regarding this research project, including available data and progress status, contact: Research Division staff at (916) 445-0753

Stay involved, sign up with CARB's Research Email Distribution List