High-frequency, high-altitude measurements of water chemistry provide insights into pro-cesses relating to the hydrology, climate, and geochemistry of mountain catchments.When such observations are combined with stream stage, temperature, snow, weather, and other surface hydroclimate measurements, they are particularly useful in allowing connections between climate, river discharge, river chemistry,and ecosystems to be discerned.Climate is the major source of variability in U.S. and global water resources. For example,large-scale variations in the global atmo-sphere and the Pacific Ocean are responsiblefor much of the variability in river discharge in Hawaii,Alaska, the U.S. Pacifi c Northwest,and the U.S. Southwest [Cayan and Peterson, 1989], and thus are closely linked to water and energy resources of the western United States [ Cayan et al.,2003].Long-term observations to defi ne these connections are limited at alpine elevations above 1,200–1,500 meters, partly because the observations are difficult to make, even though much of the water for lower elevation human needs is delivered (as snow) to thehigh elevations.A hydroclimate monitoring net-work in Yo semite National Park in the SierraNevada of California was started in 2000 to complement monitoring efforts in other parts of the Sierra Nevada and to better understand watershed responses to atmospheric forcing [ DiLeo et al., 2004].Hydroclimate linkages are known from long-term stream gauge and meteorological observa-tions, but these observations are largely from settings below 1,500 meters, and generally are not informed by river chemistry information.A first goal of the Yo semite monitoring networkwas to determine the degree of the associationbetween these continuing, long-term obser-vations and observations of stream water chemistry in natural, unmanaged watersheds.This was accomplished by monitoring river chemistry with continuous sensors at some of the long-term observation sites.Examples of river conductivity, nitrate, and dissolved silica, and watershed primary production are shown to illustrate how river chemistry and remotely sensed observations may be used to augment the information pro-vided by conventional hydrological and meteo-rological observations.Monitoring Results:The Value of Multiple ObservationsHydroclimatic Variability. Key monitoring sta-tions of the Yo semite hydroclimate monitoring network (note that there are several other in-struments in Yo semite as well) are shown in Fig-ure 1a. Geochemists recognize the importance of both geology and climate in alpine river chemistry research. In the impervious granitic watersheds of Central Sierra Nevada, geology (soil thickness) perhaps discerns hydrologic differences in alpine watersheds better than climate, and climate (large-scale snowpack and air temperature) defines hydrologic similarities in alpine watersheds better than geology. Thus,both lines of research are essential.This study, as part of a hydroclimate network,is starting with the climate connections to river chemistry. For this reason, it is important to sample the variables at the same rates, which is unusual.Examples of linkages in snow water content,air and water temperatures, river discharge,conductivity (a measure of total dissolved solids), nitrate, and dissolved silica concentra-tions are shown for 2003, a year with a strong peak in both snow melt discharge and rain events. Figures 1b–1h illustrate covariations between the stream chemistry and traditional hydrologic measures on hourly, daily, and longer timescales during spring and summer,when the system is undergoing large diurnal and seasonal changes.The steep decline in daily snow water equivalents (the depth of liq-uid water that is obtained when the snowpack is melted) is in response to the steep rise in daily air temperatures in mid-April, which in turn generates a rise in daily discharge with increasing amplitudes of the hourly snow melt discharge (SMD) deviations.Conductivity. In the Yo semite’s Merced River at Happy Isles, conductivity varies inversely as river discharge (i.e., as river discharge increases,or as time per unit discharge decreases, con-ductivity decreases), until river conductivity approaches the conductivity of melted snow (<10 μS cm-1), in mid-June (Figure 2a). By the time the lowest conductivities are achieved (Figure 2b), soil salts have been well fl ushed from the watershed above the gage.This rela-tively complete flushing within a span of a few months is possible because the soils above the gage were severely thinned and cleaned out by the most recent Pleistocene glaciation that ended about 10,000 years ago.Rain Pulse Response. High-resolution sam-pling captures the Merced River’s chemical response to rain.Dissolved silica concentrations roughly follow conductivity variations during SMD and the dry seasons, as the source of conductivity and dissolved silica is from dis-solution (weathering) of soil. However, although the lowest riverine conductivity values are similar to snowmelt, correspondingly low concentrations of dissolved silica [Brown and Skau,1975] are never observed.This is prob-ably because when the soil salts are fl ushed from the watershed, the fl ushing flows are in contact with silica including silicates, a major component of the soil and river sediment.Nitrate chemistry is different from dissolved silica and conductivity in that the atmosphere is the major source of nitrogen.Thus, a sum-mer storm (circle insets in Figures 1e–1g) produced a decrease (dilution response) in conductivity and dissolved silica and an in-crease in nitrate concentrations. Mechanisms for the increase of nitrate include nitrifi cation by lightning and wash-off of atmospheric de-position and mineralized organic matter from the dry summer landscape and soils.Chemical Signatures of Runoff From Snow-melt Versus Soil Water. The Merced River atHappy Isles and the Stanislaus River at Clark Fork (in the neighboring basin north of the Tuolumne River) have similar runoff rates (riverdischarge per unit area) and climate-driven daily and hourly SMD variations. However, the two watersheds differ geologically, with sparse soil in the Merced(lowmean soil thickness) compared with the Stanislaus (high mean soil thickness). The difference is that while all of the glaciations, including the most recent,“cleaned out” the Merced above the gage, the watershed above Clark Fork, on the Stanislaus, was