Meteorological observations collected during the Storms and Precipitation Across the continental Divide Experiment (SPADE), April–June 2019

, Abstract. The continental divide along the spine of the Canadian Rockies in southwestern Canada is a critical headwater region for hydrological drainages to the Pacific, Arctic, and Atlantic oceans. Major flooding events are typically attributed to heavy precipitation on its eastern side due to upslope (easterly) flows. Precipitation can also occur on the western side of the divide when moisture originating from the Pacific Ocean encounters the west-facing 20 slopes of the Canadian Rockies. Often, storms propagating across the divide result in significant precipitation on both sides. Meteorological data over this critical region are sparse, with few stations located at high elevations. Given the importance of all these types of events, the Storms and Precipitation Across the continental Divide Experiment (SPADE) was initiated to enhance our knowledge of the atmospheric processes leading to storms and precipitation on either side of the continental divide. This was accomplished by installing specialized meteorological 25 instrumentation on both sides of the continental divide and carrying out manual observations during an intensive field campaign from 24 April–26 June 2019. On the eastern side, there were two field sites: (i) at Fortress Mountain Powerline (2076 m ASL) and (ii) at Fortress Junction Service, located in a high elevation 2 particles. The objective of the field campaign was to gather high temporal frequency meteorological data and to 40 compare the different conditions on either side of the divide to study the precipitation processes that can lead to catastrophic flooding in the region. Details on field sites, instrumentation used, and collection methods are discussed. Data from the study are publicly accessible from the Federated Research Data Repository at https://dx.doi.org/10.20383/101.0221 (Thériault et al., 2020). This dataset will be used to study atmospheric conditions associated with precipitation events documented simultaneously on either side of a continental divide. 45 This paper also provides a sample of the data gathered during a precipitation event.

to the Canadian Rockies. Moisture to the region can also originate from the Great Plains and the Gulf of Mexico, which is forced upslope on the eastern flanks of the Canadian Rockies. This is a mechanism that has previously led to extreme flooding, such as in southern Alberta and British Columbia in 2013 Kochtubajda et al., 2016;Liu et al., 2016).

60
To understand the contribution of these different moisture flows on precipitation across the Canadian Rockies, the Storms and Precipitation Across the continental Divide Experiment (SPADE) was initiated wherein precipitation events across the continental divide were investigated from 24 April-26 June 2019. Over the two-month period, 13 storms with varying atmospheric conditions and precipitation amounts and types were documented on both sides of the continental divide in southwestern Canada. This project enhances knowledge of both the large-and fine-scale Orographic Precipitation Experiment (FOPEX; Smith 2008) studied the changes in precipitation amount and elevation along an east-west transect from 2001 to 2005. Other projects around the world were also held to study cold season precipitation processes such as in the Swiss Alps (Steiner et al., 2003) as well as in China, where a 85 recent field study occurred in the Haituo Mountains north of Beijing (Ma et al., 2017). Nevertheless, none focused specifically on collecting high-resolution automatic and manual precipitation data simultaneously across a major continental divide using a combination of sophisticated instruments such as the Micro Rain Radars, laser disdrometers and microphotography, as well using Doppler Light Detection And Ranging instruments (LiDARs) to measure air flow at two elevations in mountainous terrain.

90
SPADE expands upon a previous study and field campaign by Thériault et al. (2018) in March and April 2015 that investigated precipitation characteristics and associated atmospheric driving mechanisms on the eastern slopes of the Canadian Rockies. Although that field campaign occurred during a period of above normal temperatures and less precipitation than normal, 17 precipitation events were observed. Precipitation events were categorized as either 95 upslope (easterly) or downslope (westerly) flow, and the precipitation events associated with downslope flows typically had deeper precipitation layers and greater instability (Thériault et al., 2018). It was also observed that liquid, solid, and mixed-phase precipitation could all occur within the same event. Of the solid precipitation particles that were observed, 62% of those particles were rimed, with rimed irregular particles being the most common particle type. Additionally, rimed and unrimed particles were observed simultaneously during precipitation events, 100 although the riming of particles was independent of whether flows were westerly or easterly (Thériault et al., 2018).
The SPADE domain spans both the western and eastern sides of the continental divide in the Canadian Rockies ( Fig.   1) from 117° W to 114° W longitude and 50° N to 52° N latitude. This region is subject to spatially variable weather conditions, with large orographic features acting as the major catalyst for precipitation. The highest peak in the 105 domain is Mount Assiniboine, at an elevation of 3612 m above sea level (ASL). The elevation of the continental divide varies from 1130 m to over 3600 m ASL. Meteorological data over the continental divide remain sparse and predominantly lie in the valleys where sites are accessible by road and AC power is more readily available (Vionnet et al., 2020). There are even fewer stations at higher elevations, hence the paucity of meteorological data over high elevation regions (Pepin et al., 2015;Hernández-Henríquez et al., 2018). Short, cool summers and long, cold winters 110 characterize the region's climate. Precipitation typically peaks in June (Liu et al., 2016) with 60% of the precipitation falling as snow in the mountains, which can occur in all months (DeBeer and Pomeroy, 2010).
Maximum snow water equivalent (SWE) occurs in May and decreases throughout June and early July . The three main atmospheric circulation types that induce snow to the Canadian Rocky Mountains were categorized by Moran et al. (2007) as northwesterly, westerly, and leeside (upslope). Westerly and northwesterly 115 flows often result in significant precipitation accumulation on the western side of the continental divide, while upslope (easterly) flows often result in heavy precipitation and greater accumulation on the eastern side of the continental divide. Upslope events typically occur from March-May (Sinclair and Marshall, 2009

125
Moved up [2]: To understand the contribution of these different moisture flows on precipitation across the Canadian Rockies, the Storms and Precipitation Across the continental Divide Experiment (SPADE) was initiated wherein precipitation events across the continental divide were 130 investigated from 24 April-26 June 2019. Over the twomonth period, 13 storms with varying atmospheric conditions and precipitation amounts and types were documented on both sides of the continental divide in southwestern Canada. This project enhances knowledge of both the large-and fine-135 scale atmospheric processes that contribute to storms and precipitation across the continental divide, and that may lead to major flooding events in western Canada. ¶

Moved (insertion) [1]
Deleted: studied 140 et al., Liu et al. 2016;Kochtubajda et al., 2016).  showed that flooding due to rain-on-snow events occurred more frequently in the late 19th and early 20th centuries than it does in the more recent period.

145
The objective of this paper is to describe the methods of data collection during the SPADE project. It fills in key gaps in the otherwise well-instrumented hydrometeorological measurements and long-standing research conducted at Fortress Mountain, a Canadian Rockies Hydrological Observatory (https://researchgroups.usask.ca/hydrology/science/research-facilities/crho.php#Overview). Section 2 outlines the study area and its climatology. Section 3 describes the instrumentation used and specific observations conducted during the project, and Sect. 4 provides examples of the meteorological variables and observations. Finally, Sect. 5 details the online location of the database, with concluding remarks in Sect. 6.

Site Description
The study region consisted of two main areas separated by the continental divide. Table 1 summarizes information on the SPADE field sites. On the eastern side, there were two main field sites and two secondary sites. The main 155 sites were: (i) Fortress Mountain Powerline (FMP) at 2076 m ASL and (ii) Fortress Junction Service (FJS), located in a high elevation valley (1580 m ASL) along Alberta Provincial Highway No. 40 (Fig. 1). FMP was chosen as it is the primary field site in the Canadian Rockies Hydrological Observatory (https://researchgroups.usask.ca/hydrology/science/research-facilities/crho.php#Overview), operated by the University of Saskatchewan Centre for Hydrology, and is already well instrumented with hydrometeorological equipment. It is the 160 site of long-standing research (Smith et al., 2017;Conway et al., 2018;Schirmer and Pomeroy, 2020). A maintained road from FJS up to FMP allowed researchers to collect meteorological data during a precipitation phase transition along the mountainside. The frequent phase transition between FJS and FMP justified having two main sites located at valley floor and at higher elevation. Select data were also gathered on an intermittent basis at two secondary sites but not simultaneously. These were (i) from the University of Calgary Biogeoscience Institute (BGI) at 1418 m ASL 165 located on the eastern side of the continental divide, off Alberta Highway No. 40, ~25 km north of FMP; and (ii) from the Storm Mountain Lodge (SML) at 1723 m ASL that was used temporarily on 7 June 2019 in anticipation of a forecasted significant precipitation event. SML is located approximately 5 km east of the continental divide, near the midway point between the Fortress Mountain and NMR sites, along BC Highway 93 (Fig. 1).

170
The Nipika Mountain Resort (NMR) site was located in a valley on the western side of the divide and was chosen to be comparable to FJS on the eastern side of the divide (Fig. 1). It approaches the latitude of the Fortress Mountain area and has an elevation of 1087 m ASL. NMR is easily accessible via an active logging road off the Banff-Windermere Highway (BC Highway 93). This area has a paucity of active meteorological stations and basic weather instruments were deployed on 21 September 2018, well before the start of the field experiment, to collect baseline 175 data.

385
Various meteorological equipment was deployed at each field location (Fig. 2). A complete list of all instruments used and their locations is given in Appendix A, Table A1. All data available in the repository are in UTC. The appendix also includes a table giving the uncertainty of each instrument (Table A2). The three main sites were nearly identically instrumented, but due to power constraints and limited equipment, NMR was not equipped with a Doppler LiDAR. Cameras were also set up at FMP and FJS to monitor conditions while observers were not on site.

390
The instruments are described in detail in Sect. 3.2-3.9. The FMP site is also known as the Fortress Mountain Snow Laboratory and operates within the regional Canadian Rockies Hydrological Observatory. Other projects and instrumentation at this site are described in Smith et al. (2017), Conway et al. (2018), and Schirmer and Pomeroy (2020). Only the MRR-2 files were processed and quality controlled using the Maahn and Kollias (2012) algorithm.
All other data files have not been processed nor quality controlled by the authors, and are the output of the 395 instrument or manufacturer's software.

Deleted: )
Deleted: The FMP site, also known as the Fortress Mountain

400
Snow Laboratory within the regional Canadian Rockies Hydrological Observatory, and some of the instruments located at FMP, but not used in the SPADE campaign, are

430
A Vaisala HMP155 air temperature and relative humidity sensor was deployed on the scaffolding at the FMP and FJS field sites (Fig. 3b) at a height of 195 cm and 168 cm above ground level (AGL), respectively. Data were recorded at 1 min resolution throughout the SPADE field campaign from 24 April-26 June 2019.

WXT520
A Vaisala WXT520 was deployed on the scaffolding at the FMP and FJS field sites (Fig. 3c) at a height of 265 cm and 238 cm AGL, respectively. The WXT520 measures six weather parameters in one sensor, including wind speed and direction, precipitation, atmospheric pressure, air temperature, and relative humidity (Vaisala, 2012). Data were 440 collected at 1 min resolution throughout the SPADE field campaign from 24 April-26 June 2019.

HOBO T/RH
The HOBO Pro V2 Temperature and Relative Humidity (T/RH) (U23-001) is a weatherproof data logger with built-445 in temperature and relative humidity sensors. The HOBO T/RH was temporarily deployed at SML during the 7 June 2019 precipitation event and data were collected at 5 min resolution throughout the storm. The instrument was housed in a temporary radiation shield attached to a wooden post at 120 cm AGL in a clearing and was level to the ground. This sensor was primarily used for the SML deployment to differentiate whether air temperature was below freezing during a storm event, not to explicitly record high accuracy temperature and humidity values.

Precipitation gauges
Several types of precipitation gauges were installed and used during the field campaign. At our three main field sites, we used shielded weighing gauges (OTT Pluvio and Geonor). These shielded-gauges are well-known for their accuracy and have been used interchangeably by Environment and Climate Change Canada (Milewska et al. 2018).
Tipping bucket rain gauges were installed at our "secondary" field sites (HOBO and Davis tipping bucket rain gauges) due to the remoteness of the locations, logistical and power constraints. The HOBO tipping bucket had been previously tested in the field and showed good accuracy when compared to the Geonor for rain. Additional efforts 485 were made to reduce wind induced undercatch by placing the gauges in sheltered areas and to reduce evaporative losses by removing the debris screens. No adjustment for wind undercatch of solid precipitation was performed on the archived data and they should be used with caution.

490
The HOBO tipping bucket rain gauge (TBRG) (RG3-M) measures liquid precipitation at a resolution of 0.2 mm tip -1 (Onset, 2019) and was recorded using an Onset HOBO data logger. Data were only collected when a tip occurred.
The TBRG was temporarily deployed at SML during the 7 June 2019 precipitation event. The instrument was attached to a wooden post at 134 cm AGL in a clearing and was level to the ground (Fig. 3d).  April 2019 at FJS (OTT Pluvio2) for the duration of the SPADE field campaign (Fig. 3f). No filtering/processing has been applied to these data.

Geonor
A three sensor Geonor T-200B all-weather precipitation gauge in a single Alter shield was installed at NMR on 21 September 2018 on a pedestal mounted into a heavy lumber base and the orifice was levelled with respect to the ground (Fig. 3g). The CR1000X data logger was configured to sample the period average frequency from each 520 vibrating wire sensor, this frequency is used to compute the average amount of liquid equivalent precipitation of the Deleted: -Deleted: . We acknowledge that this would affect the data and they deviations of the three sensor depths. No filtering/processing has been applied to these data.

Manual observations and microphotography
Manual weather observations were conducted at the FMP and NMR field sites during precipitation events 540 throughout the field campaign, and at SML during the 7 June 2019 precipitation event. Manual observers were on site during storms to report precipitation type and to take microphotographs of snow particles. Every 10 minutes, manual observations of precipitation type (solid, liquid, or mixed) were reported and microphotographs of snow particles were taken when conditions allowed. Microphotography equipment was situated at NMR, however, images were not taken because insufficient solid precipitation particles were observed. Microphotographs of snow particles 545 were taken following the methods established in Gibson and Stewart (2007) and Thériault et al. ( , 2018. Images at FMP were made with a Nikon D3200 Digital SLR camera and at SML with a Nikon Digital D80 Digital SLR camera, both fitted with a 60 mm macro lens and flash. The photography equipment was mounted vertically (Fig. 3i), and precipitation particles were gathered on a black velvet or felt covered collection pad that was placed outside for a period of time, typically 5 s to 2 min, depending on the rate of snowfall. The collection pad was then 550 brought inside an unheated fishing tent and a series of nine images were taken using a predefined method that ensured uniform results regardless of the observer. Due to the differing particle size and amount of light available for each photo, the camera settings varied throughout the project, but are noted in the metadata file. An image of the millimeter divisions on a ruler was captured periodically to provide a scale for each series of photographs.

Laser-optical disdrometer
Three OTT Parsivels, which are laser-optical disdrometers and present weather sensors, were deployed at FMP and FJS on 24 April 2019, and at NMR on 1 May 2019 for the duration of the field campaign (Fig. 3h). This instrument measures the size and speed of falling hydrometeors with the goal of classifying hydrometeor type and retrieving 560 precipitation particle size distribution. These raw data can also be used to determine the precipitation type, amount, intensity, and kinetic energy, as well as the equivalent radar reflectivity. OTT Parsivel laser-optical disdrometers function with two sensor heads facing each other, in which one head is a transmitter that emits radiation (at 650 nm wavelength at the red band) in a horizontal plane and the other head is a receiver that senses how much of that radiation is received. The instrument measures the size of the hydrometeor by determining the amount of radiation 565 that is blocked by the particle diameter. The velocity of the hydrometeor is estimated based on the time that a particular hydrometeor blocks the radiation between the transmitter and receiver. Each particle falling though the beam is classified into 32 fall speed and 32 diameter bins. An OTT Parsivel was installed at FMP at a 300 cm height and at FJS at 277 cm, and an OTT Parsivel² was installed at NMR at 256 cm. The OTT Parsivel² is similar in Deleted: The CR1000X data logger was configured to sample 570 the period average frequency from each vibrating wire sensor, convert the frequencies to depths, and calculate the average depth of the three sensors and the standard deviation across them.

600
Identical Halo Photonics Streamline XR scanning Doppler LiDARs were installed at FMP and FJS on 24 April 2019 for the duration of the field campaign (Fig. 3l). Doppler LiDARs have demonstrated their usefulness in complex, mountainous terrain such as those in the SPADE domain (Banta et al. 1997Darby et al., 1999;Fast and Darby, 2003;Mariani et al., 2018a, b). The Doppler LiDAR emits a pulsed laser and measures the Doppler shift and 605 the intensity of the signal backscattered by small sized targets (i.e., aerosols). The LiDARs operate at 1.5 µm using an 80 µJ pulsed laser at 10 kHz with a range resolution of 3 m (60 m overlapping range gates). This range gate length was chosen so that there would be a small vertical overlap between the LiDAR located at FJS and the LiDAR located at FMP. The LiDARs have full scanning capability, allowing them to conduct measurements at any elevation and azimuth, similar to most weather radars. The first measurement is 60 m from the LiDAR due to the 610 LiDAR's blind spot. Both LiDARs were subjected to identical quality control procedures based on their signal-tonoise ratio (SNR) within each range gate and filtering outliers and returns from clouds and rain droplets (Mariani et al., 2018a). The maximum range of the LiDAR is limited by the sensitivity (SNR) and a minimal value of 1.003.
This was used as a lower threshold based on the minimal value of 1.0008 from Päschke et al. (2015) and adjusted according to lower SNR values observed at FMP, due to a lower amount of aerosols. The LiDARs used identical 615 repeating 10 min scan sequences, performing vertical stare, constant 4° elevation 360° azimuth plan position

Overview
Over the course of the field campaign, 13 storms were observed, of which four were categorized as major (>20 mm of precipitation at FMP) (Fig. 4). The storms were caused by a variety of conditions and resulted in varying air temperatures, durations, amounts, and types of precipitation. The eastern side of the continental divide received more precipitation and was cooler than the western side. An example of some of the data collected during a storm are discussed in Sect. 4.2. MRRs, vertical bands of precipitation appear at both FMP and NMR (Fig. 5), however precipitation starts at FMP ~10 h before it begins at NMR. Observers were on site at both FMP and NMR throughout the storm to conduct field observations of weather conditions. These manual observations, combined with data from the laser-optical disdrometer, indicate that precipitation at NMR was mainly liquid, whereas at FMP, precipitation was mainly snow, snow pellets, or mixed phase (Fig. 6). At FMP, observers took microphotographs of snow particles and their 665 evolution throughout the course of the storm (Fig. 7). This approach allows for an analysis of the solid particle size distribution, in addition to providing another method for documenting the evolution of atmospheric conditions aloft.

Figure 4: Sub-hourly temperature (red line), dew point (blue line), and log-scale precipitation amount (bars) at (a) NMR and (b) FMP for the duration of the SPADE field campaign (24 April-26 June 2019). Precipitation is subdivided into rain (green bars), snow (blue bars), or mixed phase (red bars) based on manual observations. Grey bars indicate that
Observers at NMR were unable to take microphotographs due to the warm temperatures but observed that there was some mixed precipitation.

5 Data availability
The SPADE dataset is available from the Federated Research and Data Repository (FRDR) and can be accessed at https://dx.doi.org/10.20383/101.0221 (Thériault et al., 2020).

Final remarks
A valuable dataset was collected during the Storms and Precipitation Across the Continental Divide Experiment that 700 was held in April-June 2019 in the Canadian Rockies. SPADE was initiated to enhance our knowledge of the atmospheric processes leading to storms and precipitation across a large orographic feature by gathering meteorological data. This leads to a unique dataset to specifically address this critical issue of water redistribution and availability over North America. Furthermore, it augmented the large effort in monitoring hydrometeorological conditions in the Canadian Rockies.

705
A combination of manual and automatic measurements of precipitation and meteorological conditions at the surface and aloft was collected. These include information on the amount, type and characteristics of precipitation particles, including particle size and fall speed, as well as 3D wind fields at lower and higher elevations of Fortress Mountain.
Overall, data from the SPADE field campaign will contribute significantly to our understanding of precipitation 710 processes across the continental divide from the synoptic-scale conditions leading to precipitation to the fine-scale processes associated with precipitation trajectories near the surface.

730
The authors declare that they have no conflict of interest.

Disclaimer
Specific instrument manufacturers/models and suppliers mentioned in the manuscript and/or used in the field campaign is not a commercial endorsement of their products.    Precipitation intensity ± 1.0% or ± 6 (whichever is larger) Precipitation amount ±.1.0% or ± 0.1 (whichever is larger) OTT Pluvio 2 Precipitation intensity ± 1.0% or ± 6 (whichever is larger) Precipitation amount ± 1.0% or ± 0.1 (whichever is larger) Onset RG3-M Tipping Bucket Precipitation ±1.0%

Davis Tipping
Bucket 7852 Precipitation ± 4% of total or ± 1 tip of the bucket (0.2 mm)