Lake O ’ Hara alpine hydrological observatory : Hydrological and meteorological dataset , 2004-2017

The Lake O’Hara watershed in the Canadian Rockies has been the site of several hydrological investigations. It has been instrumented to a degree uncommon for many alpine study watersheds. Air temperature, relative humidity, wind, precipitation, radiation, and snow depth are measured at two meteorological stations near Lake O’Hara and in the higher elevation Opabin Plateau. Water levels at Lake O’Hara, Opabin Lake, and several stream gauging stations are recorded with pressure transducers and validated against manual measurements. Stage-discharge rating curves were determined at gauging 10 stations and used to calculate discharge from stream stage. The database includes additional data such as water chemistry (temperature, electrical conductivity, and stable isotope abundance) and snow survey (snow depth and density) for select years, as well as geospatial data (elevation and land cover). This dataset will be useful for future study of alpine regions, where substantial and long-term hydrological datasets are scarce due to difficult field conditions. The dataset can be accessed at: https://doi.org/10.20383/101.035 15

Refer to "temperature" as "Air temperature" throughout the manuscript. This will avoid confusions with water temperature. The text has been edited to make the distinction between air and water temperature clear (Lines 39 and 161).
I understand this is a "data manuscript" but, nevertheless, I would like to see a little bit more of data analysis. For example, more about inter-annual variability of the fluxes (e.g. precipitation, streamflow, air temperature), is there a year particularly interesting in the dataset that it is worth to look at in more detail, runoff ratios (easy to calculate and very informative) and mean hydrograph. We have expanded upon our analysis of runoff at the Lake O'Hara outlet and precipitation. In addition to the comparison of total discharge and precipitation, we have added the comparison of precipitation and late-summer flow, which represents the contribution of groundwater discharge (Lines 172-176). There was no correlation between the two variables, indicating that the groundwater discharge is likely controlled by the storage capacity and transmission characteristics of alpine aquifers. This has a major implication in our understanding of groundwater processes, but we refrained from presenting lengthy discussion of the importance of this finding.

Provide coordinates for the weather and hydrometric stations (not found in the .csv files).
We have added weather and gauging station coordinates to the supplement (Table S3) and to the metadata included within the dataset.

L78-79: If sensor height has changed this should be noted with the associated height and date, a table could be a good idea.
A record of sensor height measurements is provided with the supplement (Table S4). The Table 1 caption has been edited to refer to this record more clearly.
L94-97: The procedure used to fill precipitation gaps requires more details. For example, how much of the data this had to fill? And when exactly? This kind of details is very important for future hydrological applications. How much did the precipitation changed after wind undercatch correction (percentage)?
We have added a table in the supplement listing the date ranges of when precipitation was approximated and total precipitation approximated over each range (Table S1). After the wind undercatch correction was applied, precipitation increased by 8.6% and 0.3% at OPAWS and OHAWS respectively (Line 94). We have edited the figure to display the linear equation forced through the origin and the equation and coefficient of determination. We have rewritten the caption to describe which data are plotted in greater detail.
L121: For what years exactly water levels are not available? Years missing corrected water level data are 2004-2008at Lake O'Hara, 2004-2008and 2012 at Opabin Lake. This is shown now in Table S2 of the supplement.
L145-146: I would expect LiDAR to be much more accurate than any older and lower resolution DEM. We agree. We suspect that the difference is likely due to the inconsistency in the elevation datum used during our LiDAR survey compared to the older map. We have revised the sentence to make this point more explicit (Lines 149-151).

L159: Speculation, remove.
This is widely reported in the literature. We have kept the statement, but added a reference for support (Line 164).
L167: Did you perform a statistical test to investigate runoff change? Your period of analysis is probably too short to do this. We agree. We have deleted the sentence on the trend in runoff (Line 176). For the correlation analysis between precipitation and discharge in this paragraph, we have included the statistical level of significance (p) (Lines 171-172).

Figure 4: Missing a legend.
We believe this comment refers to Figure 4a and 4b showing the annual and winter time series. We indicate these by annotating the lines instead of including a legend box, which would clutter the figures.

Introduction
Mountains are an important source of water for downstream regions (Viviroli et al., 2007). The hydrology of mountains in mid-and high-latitude regions is dominated by the storage of water in the form of snowpack and glaciers, which provide melt water to headwater streams during the melt season (Barnett et al., 2005). Climate warming can influence these processes through changes in the timing of snow accumulation and melt, the transition towards more rain and less snow, and 20 depletion of glaciers and mountain permafrost (Bales et al., 2006). While it is straightforward to understand the effects of these changes on the timing and magnitude of spring freshet, their effects on the flow during low-flow periods is uncertain because groundwater can store and release melt waters, thereby buffering the effects of warming (Tague and Grant, 2009).
However, groundwater processes in alpine headwaters are not well understood due to the lack of long-term, field-based studies examining the interaction of surface water and groundwater in alpine zones. To address this gap in knowledge, an 25 alpine hydrological observatory was established in the watershed of Lake O'Hara in the Canadian Rocky Mountains (see Study Site below) in 2004. The observatory has supported a series of studies designed to identify important alpine aquifer units and characterize their hydrogeological functions (Roy and Hayashi, 2009;Langston et al., 2011;Muir et al., 2011;Hood and Hayashi, 2015). The observatory was integrated into a larger hydrological monitoring network under the Changing public through the CCRN data server. The objective of this paper is to describe the physiographical characteristics of the Lake O'Hara watershed, the history of the hydrological observatory, and the methods of data collection and processing for the potential users of the dataset.

Site description
The Lake O'Hara watershed in Yoho National Park has an area of 14 km 2 and ranges in elevation from 1996 to 3440 meters 35 above sea level (m.a.s.l.) ( Figure 1). The site is accessible via an 11-km dirt road up to Lake O'Hara. Raw mean annual precipitation measured within the watershed at the Opabin automatic weather station (OPAWS) was 1021 mm during 2005-2017. After adjustment for noise and wind induced undercatch (Kochendorfer et al., 2017), corrected annual precipitation at OPAWS was 1113 mm. Mean monthly air temperatures at OPAWS range from -9.0 (January) to 10.2 o C (July).

40
Bedrock throughout the watershed consists mainly of quartzite of the Cambrian Gog Group. At higher elevations, carbonate of the Cathedral, Mt. Whyte, Stephen and Eldon Formations may be found capping mountains (Lickorish and Simony, 1995;Price et al., 1980). Bedrock is typically exposed at higher elevations as part of ridges which ring the watershed (Figure 2a).
Overburden deposits (e.g. talus and moraine) are present throughout the region, and are found in association with steeper bedrock slopes and small glaciers (Figure 2b and 2c). These deposits play an important role in groundwater and surface 45 water exchange throughout the watershed (Langston et al., 2011;Muir et al., 2011;Roy et al., 2009). For example, 70-80% of flow in Upper Opabin Creek is provided from a large talus-moraine complex present in the Opabin plateau (Hood and Hayashi, 2015). Surface water bodies such as Opabin Lake, Hungabee Lake and the eponymous Lake O'Hara are found at lower elevations in the watershed. Previous study has indicated active interaction between lakes and groundwater (Roy and Hayashi, 2008). Groundwater input to Lake O'Hara was estimated to equal 35-74% of surface outflow from the lake in 2005 50 (Hood et al., 2006). Alpine meadows (Figure 2d) are generally found in close association with surface water bodies and are important sites for hydrological and ecological processes (McClymont et al., 2010). Tipping bucket rain gauges (Hydrological Services, CS700) and weighing cumulative precipitation gauges (Geonor,T200B) are installed at both stations. Both types of precipitation gauge are suitable for measuring liquid precipitation, however, only the weighing gauge is capable of measuring solid precipitation. Tipping bucket gauges are placed on the ground, away from the main weather station tripod. Weighing gauges are mounted on a freestanding base and are equipped with an Alter wind 85 shield to reduce wind-related snow undercatch. Precipitation and relative humidity sensors are calibrated yearly in order to ensure that they continue to accurately record measurements. However, relative humidity measurements fail to reach 100% in several years, possibly due to the instrumental accuracy (2% reported by the manufacturer) affecting the two-point calibration procedure using the saturated chemical solutions with known equilibrium humidity..

90
Meteorological measurements in the dataset are presented in raw and corrected form. Some effort has been made to remove erroneous data due to sensor malfunction or maintenance in the raw dataset, but the data have largely been spared from other corrections. The corrected dataset has been adjusted for noise and wind induced undercatch (Kochendorfer et al., 2017).
From 2011 to 2016, snow has been observed to accumulate at the top of the OHAWS weighing precipitation gauge and 95 persist over the span of several weeks in the wintertime. This accumulation effectively blocks the opening of the gauge (i.e. snow capping), resulting in a period of time where no precipitation is registered. We attempt to approximate this missing data at OHAWS from measurements recorded at OPAWS in the corrected dataset. Comparison of noise and undercatchcorrected daily precipitation between the two weather stations indicates a linear correlation (Figure 3). Using the trend line from Figure 3, precipitation at OHAWS was approximated and used to fill in gaps during periods of snow capping. Table S1  100 in the supplement lists the date ranges where precipitation was approximated, and total approximated precipitation.

Stream discharge
Gauging stations are located at the outlets of Lake O'Hara and Hungabee Lake and along reaches of several creeks ( Figure   1). To establish the discharge-stage rating curve, biweekly manual discharge measurements are taken at each station over field seasons which typically last from early June to late September. Average stream flow velocity and water depth are 105 measured in 10 or 25 cm wide segments across the width of a stream with a propeller flow meter (Global Water, FP101), and discharge is calculated by the area-velocity method (Dingman, 2002, p609).
Pressure transducers (In-Situ, Minitroll; In-Situ, LevelTroll; Solinst, Levelogger) are installed in stilling wells at each stream gauging station and automatically record water level and temperature in fifteen minute intervals. To verify the transducer 110 data, water levels are measured manually from a reference (staff gauge or stilling well) with discharge measurements. A power function is used to define the rating curve, which is calibrated each year to find a coefficient and exponent which minimize error between measured and computed discharge. With the rating curve, a near-continuous record of stream discharge can be calculated from the transducer data. Hourly and half-hourly averaged discharge is included within the dataset. Discharge is only measured from spring (typically early June) to fall (typically late September) of each year due to 115 winter freeze up in stream channels.

Lake water level
Pressure transducers housed in stilling wells are used at water level monitoring stations in Lake O'Hara and Opabin Lake ( Figure 1). Transducer water levels are compared against manual measurements of water level taken during field visits. At Lake O'Hara, manual measurements are taken as the distance from the top of the stilling well casing to the lake water level. 120 At Opabin Lake, manual measurements are taken as the distance from a rock bolt (securing the stilling well) to the lake water level. Within the dataset, both raw water level (depth of water above the transducer) and corrected water level are included. Corrected water level is normalized against the benchmark and is calculated as the distance between the benchmark and the water level. Therefore, corrected water level increases when lake water level (and raw water level) drops.
The same is true in reverse. Negative values of corrected water level indicate periods when lake water level is above the 125 5 datum. Corrected water levels are not available in several years due to inconsistent manual measurement and recording of lake water levels (Table S2). Water levels are only available from spring to fall of each year due to freeze up.

Snow surveys
Snow depth and density are measured annually in mid-April to capture the amount of peak accumulation. The extent of these snow surveys has varied over the years from a handful of transects nearby OHAWS and OPAWS to hundreds of 130 measurements spanning the entire Opabin plateau. Snow surveys are typically conducted by laying a measurement chain along the survey transect and measuring snow depth at a fixed interval with a probe or ruler. Snow density measurements are taken along the survey transect or from snowpits (see Hood and Hayashi, 2015 for details). Handheld Global Positioning System devices are used to locate depth and density measurement points. Snow survey data from 2006 to 2017 are included within the dataset. Before 2012, measurements near OHAWS were made inconsistently, but in more recent years, data from 135 both OPAWS and OHAWS are available ( Table S2 in the supplement).

Water sample collection and analysis
Stream water and rain samples are collected during biweekly site visits. Stream water is sampled at most gauging stations ( Figure 1). During sampling, water is filtered in the field with 0.45 µm disposable filters and stored in pre-rinsed polyethylene bottles. Electrical conductivity and temperature of stream water are measured during collection with handheld 140 meters (VWR, 2052-B; Omega Engineering, HH-25TC). Rain samples are collected from samplers deployed near both weather stations and Opabin Lake. Depth-integrated snow samples were collected from snow pits during snow survey campaigns in 2015 and 2016, and sub-samples of entire melt water were kept for isotope analysis. All water samples are stored at 4 °C until analysis. Oxygen-18 and deuterium isotope abundances are measured in all collected samples using an off-axis integrated-cavity spectrometer (Los Gatos Research, DLT-100). This dataset includes chemical and stable isotope 145 data collected from 2004-2008, 2013, and 2015-2016.

Spatial data
The dataset includes a 2-m resolution digital elevation model (DEM) of the Lake O'Hara region derived from the light detecting and ranging (LiDAR) data (Hopkinson et al., 2009). Some difference in elevation has been found between the LiDAR DEM provided and older, lower resolution topographic maps of the region, possibly resulting from inconsistency in 150 the elevation datum used during the LiDAR survey. However, the LiDAR DEM still represents topography within the  October 1 -September 30; e.g., HY2006 starts on October 1, 2005. Winter is defined as October 14 -April 30, because the average start date of snow accumulation is October 14, and the accumulation normally peaks in late April. Annual total precipitation at Opabin AWS ranged between 920 and 1394 mm with a mean of 1113 mm, while winter precipitation ranged between 475 and 785 mm with a mean of 612 mm (Figure 4a). The ratio of winter to total precipitation ranged between 0.46 160 and 0.65. Annual mean and winter mean air temperature ranged from -2.4 to -0.1 o C and from -8.9 to -5.4 o C, respectively ( Figure 4b). There was no noticeable trend in any of these parameters.
The timing of snow accumulation and melt are believed to respond sensitively to the climate warming (e.g. Barnett et al., 2005). We define the first day of accumulation at Opabin AWS as the day when snowpack starts to persist continuously, and 165 the first day of complete melt as the day when the snow-depth sensor indicates no snow. These dates varied widely with the first day of melt from May 29 to July 3 ( Figure 4c) and the first day of accumulation from September 28 to October 26 ( Figure 4b), but there was no noticeable trend.
The total discharge (i.e. watershed runoff) during June-September measured at Lake O'Hara outlet had a large variability 170 ( Figure 4e) and was positively correlated with total annual precipitation (r, correlation coefficient = 0.78; p, significance level = 0.003) and winter precipitation (r = 0.65, p = 0.02), as expected. Previous studies in the region have shown that latesummer flow in alpine streams is predominantly sourced by groundwater (e.g. Hood and Hayashi, 2015;Harrington et al., 2018). However, discharge in September (not plotted) was not correlated to either total annual precipitation (r = 0.01, p = 0.75) or Jun-Aug precipitation (r = 0.003, p = 0.86) suggesting that the groundwater discharge rate may be controlled by 175 factors other than precipitation, such as the storage capacity or transmissivity of aquifers.There was no noticeable trend in runoff.