Analysis of the Mt. Crested Butte Ice Detector and Associated Measurements in Colorado during the Winter Season

Transcription

1 North American Weather Consultants, Inc. Analysis of the Mt. Crested Butte Ice Detector and Associated Measurements in Colorado during the Winter Season Prepared for Colorado Water Conservation Board and Gunnison County Lower Colorado River Basin States Upper Gunnison River Water Conservancy District by Stephanie Beall Don A. Griffith, CCM North American Weather Consultants, Inc S. Highland Dr., Suite B2 Sandy, Utah Report No. WM Project No September 2016

2 Analysis of the Mt. Crested Butte Ice Detector and Associated Measurements in Colorado during the Winter Season Prepared for Colorado Water Conservation Board and Gunnison County Lower Colorado River Basin States Upper Gunnison River Water Conservancy District by Stephanie D. Beall Don A. Griffith, CCM North American Weather Consultants, Inc S. Highland Dr., Suite B2 Sandy, Utah Report No. WM Project No September 2016

3 TABLE OF CONTENTS Section Page 1.0 Introduction and Background Mt. Crested Butte Sensor Suite Analysis Results for Periods of Observed Icing at Mt. Crested Butte Icing Relative to Synoptic-Scale Weather Situation and Storm Precipitation Periods Icing Relative to Synoptic-Scale Storm Situation Icing Relative to Storm Precipitation Periods Additional Analysis of Icing Occurrence According to Synoptic and Precipitation Categorizations Storm Temperature Structure during Icing Periods Icing and Wind Microwave Radiometer Specifics and Data Analyses Radiometer Liquid Water and Icing Meter Comparison Stability Derived from Radiometer Data Questions, Preliminary Conclusions, and Recommendations Questions Conclusions/Key Findings Recommendations...62 Acknowledgments...62 References...62 Appendix A Figure Page 1.1 Idealized Location of a Supercooled Water Accumulation Zone during Winter Storms in Relation to a Mountain Barrier Upper Gunnison River Basin Cloud Seeding Target Area and Generator Locations, Winter Season...3

4 Table of Contents (continued) Figure Page 1.3 Crested Butte Icing Meter and Radiometer Locations Rime Ice Accumulation, Brain Head Ski Area, Utah Mt. Crested Butte Icing Meter Site Close-up of Instrumentation at the Mt. Crested Butte Icing Meter Site Looking North from Icing Meter Site Looking East from Icing Meter Site Looking South from Icing Meter Site Looking West from Icing Meter Site Schematic of Goodrich Model 0871LH1 Freezing Rain Detector Goodrich Icing Rate Meter Precipitation Rate, Temperature and Icing Status from the Mt. Crested Butte Icing Site Synoptic Pattern Classification during Icing Periods, Season Synoptic Pattern Classification during Icing Periods, 2-Season Composite Icing Distribution with Respect to Precipitation Periods, Season Icing Distribution with Respect to Precipitation Periods, 2-Season Composite Temperature Distribution during Icing Periods for Season Winds Plots for Icing and Temperatures, Season Wind Plots for Icing and Temperatures, 2-season composite mb Wind Direction from Grand Junction during Seeded Events Microwave Radiometer Icing Meter Data from the Mt. Crested Butte Site for December 23, Radiometer Cross-Section Plot for December 23, Skew-T Diagram Generated from Radiometer Data, March 23, Nov , 2015 Radiometer Cross-Section Nov. 20, 2015 Radiometer Cross-Section Nov. 27, 2015 Radiometer Cross-Section Dec , 2015 Radiometer Cross-Section Dec , 2015 Radiometer Cross-Section Dec. 23, 2015 Radiometer Cross-Section Jan. 20, 2016 Radiometer Cross-Section Jan , 2016 Radiometer Cross-Section Feb , 2016 Radiometer Cross-Section Feb , 2016 Radiometer Cross-Section Mar. 6-7, 2016 Radiometer Cross-Section Mar , 2016 Radiometer Cross-Section Mar , 2016 Radiometer Cross-Section Mar , 2016 Radiometer Cross-Section Mar , 2016 Radiometer Cross-Section...53

5 Table of Contents (continued) Table Page 1-1 NAWC Winter Cloud Seeding Criteria Two-Season Total Icing Cycles by Synoptic/Precipitation Sub-Category Total Icing Cycles by Synoptic/Precipitation Sub-Category Seeding Information for Stability Cases during the Season Seeded Storm Events with Stability Comparison Results during the Season Radiometer and Gunnison Surface Site Comparison during Seeded Events...58

6 Analysis of the Mt. Crested Butte Ice Detector and Associated Measurements in Colorado during the Winter Season 1.0 Introduction and Background North American Weather Consultants (NAWC) has operated a winter cloud seeding program in the Upper Gunnison River Basin each winter since (Griffith, et al, 2011). A key ingredient in winter storms for them to be considered seedable is the presence of Supercooled Liquid Water (SLW). Supercooled means water is occurring in a liquid state at temperatures below freezing. In a clean laboratory environment water droplets can remain in a liquid phase down to temperatures of ~ C. In the atmosphere cloud droplets at below freezing temperatures can be changed into ice crystals (solid phase) due to a process known as nucleation. Natural particles in the atmosphere (e.g., certain dust particles or bacteria) can lead to the freezing (nucleation) of these cloud droplets. Natural nucleation increases with colder temperatures. Research conducted in the late 1940 s demonstrated that Silver Iodide (AgI) particles can serve as artificial freezing nuclei. In fact, AgI nuclei have been shown to be more active than their natural counterparts in the temperature range of approximately -5 0 C to C. The threshold activation temperature of AgI is ~ -5 0 C. In other words AgI can create artificial ice crystals that can grow into snowflakes at temperatures of ~ -5 0 C or colder. Given the above, numerous research programs have explored the different methods available to detect SLW in order to determine when cloud seeding with AgI nuclei might be productive in increasing natural precipitation. Methods that have been considered or used include: 1) in situ observations using aircraft, 2) modeling studies, 3) ground-based icing meters and 4) microwave radiometers. Fortunately NAWC has received funding from the Colorado Water Resources Board and the three Lower Colorado River Basin States to install, operate and analyze data from a suite of instruments including an icing rate meter located near the crest of Mt. Crested Butte for the past two winter seasons. NAWC recently published a peer reviewed paper that summarized some of the analysis conducted for the winter season (Griffith, et al, 2016). The above two funding agencies also provided funding to have a microwave radiometer installed then operated at a location east of Gunnison during the winter season. NAWC published a peer reviewed paper (Griffith, et al, 2013) that included an analysis of the location of SLW in wintertime storms in relation to mountain barriers. A consistent indication from various research programs indicated that SLW frequently occurred over the upwind slopes of mountain barriers at relatively low elevations with the top of the SLW only extending from 1500 to 3000 feet above the crest of the barrier. Figure 1.1 provides a stylized diagram of this SLW accumulation zone. 1

7 Figure 1.1 Idealized Location of a Supercooled Water Accumulation Zone during Winter Storms in Relation to a Mountain Barrier NAWC was given the responsibility of selecting the location of the microwave radiometer. Given the above this was somewhat of a challenge. Figure 1.2 provides a map of the intended target area for this cloud seeding program along with the locations of ground-based silver iodide generators used during the winter season s program. This is a more complex target area than most winter programs in which there is often only one mountain barrier of interest. Viewing the Gunnison target area there are 3-4 major mountain barriers of interest to the north, east south and west of the that the wind is blowing, for example a north wind would mean the wind is blowing from upper Gunnison Valley. In meteorology, wind directions are reported from the direction north to south. Since we expect SLW to develop SLW along the upwind slopes of mountain barriers (Griffith 2013), SLW may develop on the north slopes of the barriers to the north and south under northerly flow, along the west slopes of the barriers located to the east and west under westerly flow, and along the south slopes of the barriers to 2

8 the south and north under southerly flow. Upper-level winds blowing from the east seldom occur in winter storms that impact this area. A further complication is that SLW can be removed from an airmass if there is a separate barrier upwind of a target barrier. For example under southerly flow the San Juan Mountains may remove SLW that might be expected to form over the south slopes of the Elk Mountains. Since the Sawatch Range located east of the Gunnison Valley is such a high barrier and provides significant amounts of the Upper Gunnison River flow, we focused our attention on this barrier in choosing a location for the microwave radiometer. Consequently a site in or near the foothills of the western slope of this range should be within the expected SLW accumulation zone. We considered two additional logistical criteria; commercial power and internet service was required and a location on private land would preclude the need to acquire a use permit from a government agency like the U.S. Forest Service. Also, location at a residence would offer some deterrent to possible theft or vandalism. Taking all these considerations into account we decided to locate the radiometer at one of NAWC s manually operated cloud generator sites located east of Gunnison, site #20 in Figure 1.2. The Mt. Crested Butte icing meter site and radiometer locations are shown in Figure 1.3. Figure 1.2 Upper Gunnison River Basin Cloud Seeding Target Area and Ground Generator Locations, Winter Season 3

9 Figure 1.3 Crested Butte Icing Meter and Radiometer Locations. Black line represents the target area boundaries 4

10 SLW in Colorado often develops at lower altitudes above the windward slopes of mountain barriers during stormy weather, in many instances impinging on the higher mountain ridges, producing rime ice accumulation caused by freezing of supercooled cloud droplets on objects such as trees or structures (Super, 1999; Griffith, et al, 2013). Figure 1.4 provides a photograph of rime ice accumulation on a chair lift at the Brian Head Ski area located in southern Utah. In addition to the presence of sufficient amounts of SLW, there are several other factors of importance for successful cloud seeding operations. When SLW is present, silver iodide becomes an active seeding agent at temperatures below approximately -4 to -5 C, with its effectiveness increasing at lower temperatures. There is also a practical lower bound (usually near -15 C) below which significant amounts of SLW are uncommon, and the natural efficiency of the cloud system is such that seeding effects may be insignificant. Effective ground-based seeding also depends on the ability for silver iodide nuclei released at ground level to reach appropriate regions of the cloud in a timely manner, and for sufficient time after nucleation for ice particle growth within the cloud and fallout within the seeding target area. A well-mixed atmosphere at lower elevations, which is related to the vertical temperature profile, is a crucial factor for the transport of silver iodide nuclei. Wind speed, direction, and the location of the target area is a crucial factor with regard to seeded snow particle fallout. Finally, another key aspect is the ability of the project meteorologist to identify situations when all these factors are favorable for cloud seeding for a given target area, in real time. Although some detailed and technical analyses are included in this report, they were conducted with the goal of utilizing the available ice detector data to help the project meteorologist identify the likely development of good seeding situations. Table 1-1 summarizes the generalized cloud seeding criteria that NAWC utilizes in ground-based seeding programs. In addition to this, various meteorological factors are recognized as helping to optimize cloud seeding potential, as discussed in this report. An important point of emphasis is that each storm situation is unique, and the ice detector data are valuable both in real-time decision making as well as later analyses of storm events. 5

11 Figure 1.4 Rime Ice Accumulation, Brian Head Ski Area, Utah A high-mountain ice detector site was installed then operated by North American Weather Consultants during the winter season. Funding for the establishment of this site, its maintenance, and data analysis was provided by the Colorado Water Conservation Board and the three Lower Colorado River Basin States (Arizona, California and Nevada). NAWC subcontracted the acquisition and installation of the needed equipment to Meteorological Solutions, Inc., a consulting firm located in Salt Lake City, Utah. 6

12 Table 1-1 NAWC Winter Cloud Seeding Criteria 1) Cloud bases are below the mountain crest height. 2) Low-level winds would favor the movement of silver iodide nuclei from their release points into the intended target area. 3) No low-level atmospheric inversions or stable layers that would restrict the vertical movement of the silver iodide nuclei between the surface and at least the -5 C (23 F) level or colder. 4) The temperature at the mountain crest height should be -5 to -15 C, except in some strongly convective situations where seeding at warmer temperatures may be effective. 5) The temperature at the 700-mb level (approximately 10,000 feet elevation) should be -15 C or warmer. Two or three possible locations for the installation of the icing meter site were considered including: 1) Mt. Crested Butte, 2) a site at the top of Monarch Pass and 3) a site on the La Garita Mountains south of Gunnison. Previous work on both research and operational programs has indicated that SLW frequently occurs during portions of winter storms on the upwind slopes of mountain ranges. The Upper Gunnison River target area presents an interesting challenge in this regard in selecting a suitable location for an icing meter. There are several different mountain ranges that compose the target area. These ranges would be expected to produce SLW accumulations under different storm conditions basically being a function of wind direction and barrier orientation (Griffith, et al, 2013). Westerly winds would favor accumulations over the west slopes of the West Elk Range and the southern section of the Sawatch Range. Southerly flow would favor accumulations over the south slopes of the La Garita Range and the western end of the Sawatch Range. Northerly flow would favor accumulations over the northern slopes of the western end of the Sawatch Range. Recall that in meteorology wind directions are reported as the direction the wind is blowing from. All of the above areas offer seeding opportunities to affect the intended Upper Gunnison River target area. Potential Icing meter locations have certain logistical constraints; they must be accessible in winter and they cannot be located in wilderness areas. Experience in locating similar sites in Utah have indicated that selecting locations at or near the top of mountain ranges offer ideal locations to observe SLW passing over the mountain 7

13 barrier. Ski areas often are located in such locations which led to discussions with the Crested Butte Mountain Resort about the possibility of locating an icing site at or near the top of Mt. Crested Butte. The management of the resort was receptive to this approach, which led to the consideration of possible specific site locations. After consideration of several factors, a site was selected and arrangements were made for the site installation in the fall of There was an existing tower at this site (it is known as the Wind Tower site) which was used in the installation. NAWC subcontracted with Meteorological Solutions, Inc. of Salt Lake City to perform the installation work. The site includes more instrumentation than just the Goodrich icing meter, it includes temperature, wind speed, wind direction, precipitation, and solar insulation instrumentation. The site operates off of battery packs that are charged by solar panels (no commercial power available at the site). Figures 1.5 and 1.6 provide photos of the site. Perhaps of some importance is the fact that Mt. Crested Butte is a rather isolated prominent feature which is separated from the major mountain barriers to the north (Sawatch) and to the west (West Elks). This fact may impact the amount of icing that is observed in ways that may be difficult to quantify. The more typical situation is where there is a substantial mountain barrier is identified as the target area of a winter cloud seeding program. The conceptual model in this situation regarding the production of SLW in winter storms is that the lifting of a moist air mass over the barrier (an orographic effect) can generate SLW over the upwind slopes of the barrier. Mt. Crested Butte offers less of an obstacle to wind flow than the larger surrounding mountain ranges. The question might be whether winds flow around instead of over Mt. Crested Butte in which case the orographic effect would be less and possibly the production of SLW less. Of course some storms bring SLW with them without the need for orographic forcing to generate SLW. Another factor to consider is that the West Elk Mountains located west of Mt. Crested Butte will remove water vapor and may remove SLW from storms as they pass over these mountains leaving less moisture in the air mass as it passes over Mt. Crested Butte (i.e. perhaps less icing). This process is referred to as the rain shadow effect in meteorology. The fact that the icing meter site is at approximately the same elevation as the barrier summit of the West Elks may lessen this effect. These are complex issues; nonetheless as will be discussed in the following sections there was SLW observed last winter at the icing meter site on a number of occasions. The Mt. Crested Butte ice detector site is located on the north side of the summit of Mt. Crested Butte at an elevation of ~11,400 feet MSL. This site is at a rather unobstructed location in terms of near-by higher terrain. Figures 1.7 through 1.10 provide panoramic photos from the site that illustrate this point. NAWC considers it important to have such an unobstructed location since higher terrain located upwind of such a site may remove or block the SLW from reaching the icing meter site. Figure 7 8

14 indicates the potential for some blockage to the south due to Mt. Crested Butte with a crest elevation of 12,162 (~560 higher than the detector site). Mt. Crested Butte sits almost due south of the ice detector site so if any nearby terrain blockage should occur it would be from this direction; ~ 180. The icing observed at this site can be considered indicative of excess in-cloud supercooled liquid water (SLW), i.e., SLW that is not involved in the windward slope precipitation process. The ice detector represents a point measurement, and therefore the occurrence of SLW in the vicinity may be more frequent than indicated (such as in situations where SLW remains above the instrument elevation). Data from this site were collected from November 1, 2015 through May 31, This allowed for continuous monitoring of icing and associated observations during the period of seeding operations from November 15, 2015 through April 15,

15 Figure 1.5 Mt. Crested Butte Icing Meter Site 10

16 Figure 1.6 Close-Up of instrumentation at the Mt. Crested Butte Icing Meter Site 11

17 Figure 1.7 Looking North from Icing Meter Site 12

18 Figure 1.8 Looking East from Icing Meter Site 13

19 Figure 1.9 Looking South from Icing Meter Site 14

20 Figure 1.10 Looking West from Icing Meter Site 15

21 2.0 Mt. Crested Butte Sensor Suite The ice-detecting sensor is the key measurement at the Mt. Crested Butte site. A Goodrich Model 0871LH1 Freezing Rain Detector was used. Its sensing probe accumulates ice on a small vertical probe, via capture of supercooled liquid water droplets during stormy conditions at temperatures colder than freezing. An internal heater deices the probe when a predetermined mass of ice is bonded to the probe, and an icing cycle is recorded. The sensor then cools, and can once again begin to accumulate ice. A schematic of the sensor is shown in Figure 2.1. Figure 2.2 provides a photo of one of these units. The ice detector site was also equipped with instrumentation to measure various other meteorological parameters, as described in the following sub-section. Figure 2.1 Schematic of Goodrich Model 0871LH1 freezing rain detector (dimensions in inches). 16

22 Figure 2.2 Goodrich Icing Rate Meter Mt. Crested Butte Site ( N, W, 11,400 ft) The ice detection system at Crested Butte Ski Resort in southwestern Colorado is a stand-alone system. This system is composed of the following components. Goodrich Freezing Rain Detector, Model 0871LH1 CSI Temperature Sensor, Model 107-L10 R.M. Young Alpine Wind Speed and Direction Sensor, Model Thies Precipitation Rate Sensor, Model TC041-L Campbell Scientific Datalogger System, Model CR1000 Cellular Telephone Modem The data are recorded via an onsite Campbell Scientific datalogger. The data time resolution is 15 minutes. Real-time data access is available via a passwordprotected internet link using Vista Data Vision software hosted by Meteorological Solutions Inc. 17

23 3.0 Analysis Results for Periods of Observed Icing at Mt. Crested Butte Investigation of the various seasonal meteorological aspects of icing occurrences in light of other key winter storm characteristics can reveal important relationships that lead to improved seeding opportunity recognition and potentially improved cloud treatment. The following sections summarize the primary results from these analyses. Figure 3.1 provides an example of the output from the icing meter that is available to NAWC meteorologists in real-time during storm events. Icing is present on the icing meter when a blue spike occurs on the icing cycles and status panel. This panel also displays precipitation rate and temperature. Surface wind direction and speed can also be displayed. The temperature plot is also especially useful when making seeding decisions. This panel displays time on the horizontal axis and icing cycles on the vertical axis. NAWC meteorologists use this operationally when making decisions during seeding operations. 3.1 Icing Relative to Synoptic-Scale Weather Situations and Storm Precipitation Periods Icing Relative to Synoptic-Scale Storm Situations Synoptic-scale storm situations, such as pre- and post-frontal/upper trough air masses, differ by definition and their seedability can differ as well. The one season of data has been partitioned according to this distinction. The occurrence of icing cycles at the detector sites was examined with regard to the synoptic situation for storm events during this first season of data collection. The observed icing during significant events was divided into five categories, based on the synoptic-scale weather situation at each measurement site when the icing was observed: 1) Pre-frontal (warm sector of storm event, no cold frontal passage observed yet at the site) 2) After cold front and before main trough axis; the 500-mb level (approximately 18,000 feet MSL) is the standard used for defining the position of the trough axis. 3) Behind the main trough axis 4) Associated with a closed low 18

24 5) Associated with a zonal (westerly flow) pattern, or difficult to define the synoptic situation based on available data Figure 3.1 Precipitation Rate, Temperature and Icing Status from the Mt. Crested Butte Icing Site Results of the analysis are shown in Figure 3.2. For the past season ( ) of data at the Mt. Crested Butte site the post-trough synoptic category had the most icing periods with 36%. The pre-frontal category had the second highest number of occurrences with 32% and the least amount of icing occurring in post-frontal/pre-500 mb trough situations (4%). Note that this is for only one season of data so these percentages may not be indicative of a more climatological average of icing for each synoptic category. Figure 3.3 provides the same information as in Figure 3.2 but for two winter seasons of data ( and ). This figure indicates the most icing occurs in the pre-frontal category. 19

25 Number of Icing Cycles Number of Icing Cycles 40% 35% 30% 25% 20% 15% 10% 5% 0% 32% Pre-front 4% Between front/trough 36% 12% 16% Post-trough Closed low Zonal or undefined Synoptic Situation Figure 3.2 Synoptic Pattern Classification during Icing Periods, Season 45% 40% 35% 30% 25% 20% 15% 10% 5% 0% 40% Pre-front 14% Between front/trough 24% 7% 16% Post-trough Closed low Zonal or undefined Synoptic Situation Figure 3.3 Synoptic Pattern Classification during Icing Periods, 2-Season Composite 20

26 Number of Icing Cycles Some ambiguity exists in classification of certain synoptic situations, particularly for systems that evolve from one type into another (for example, from a classic midlatitude frontal system into a closed low). The closed low category in this analysis is reserved for systems that have a well-defined closed circulation during most of their time of impact, making it difficult to identify relevant cold front or trough passage times. Some situations (such as zonal or weak ridging conditions) during which icing occurs are not easily classified, and the associated icing is categorized as having an undefined synoptic situation. These analysis results are particularly valuable in highlighting when seeding opportunity occurs during storm sequences, helping to sharpen our operational procedures and convincing us of the real-time and post hoc value of the ice detector measurements. Since there is only two seasons of data at this point, the results could be misleading. However; as additional seasons of data are obtained the representativeness of the results will continue to improve Icing Relative to Storm Precipitation Periods The relationship of icing (riming) occurrence to area precipitation was investigated, via comparison of icing to time series of precipitation occurrence. This analysis was based on data from the on-site optical precipitation sensor. The threshold chosen to define precipitation periods in the analysis is 0.1 /hour of snowfall (roughly equivalent to 0.01 /hour of water content). Figures 3.4 and 3.5 provide the results of this analysis. 40% 35% 30% 25% 20% 15% 10% 5% 0% 35% 31% 16% 9% 10% Before pcp During pcp Between pcp After pcp No pcp Synoptic Situation Figure 3.4 Icing Distribution with Respect to Precipitation Periods of 0.01 /hr (0.1 /hr snowfall) or Greater during Storm Events, Season 21

27 Number of Icing Cycles 50% 45% 40% 35% 30% 25% 20% 15% 10% 5% 0% 44% 19% 21% 9% 7% Before pcp During pcp Between pcp After pcp No pcp Synoptic Situation Figure 3.5 Icing Distribution with Respect to Precipitation Periods of 0.01 /hr (0.1 /hr snowfall) or Greater during Storm Events, 2-Season Composite The precipitation vs. icing analyses is quite valuable, because it provides an approximation of relative storm precipitation efficiency. A result noted in Utah ice detector reports, and consistent with those data is that seeding opportunities can persist for a few to several hours after precipitation has ended in a given storm sequence. This has been suspected in the past, but these data provide confirmation and a better sense of the magnitude of the seeding opportunity. At the Mt. Crested Butte site most icing last season occurred between periods of precipitation (35%) with the second-most likely scenario when icing occurred after precipitation (22%). The late stages of a storm are frequently associated with colder air masses, which is important for seeding given the greater effectiveness of silver iodide at colder temperatures. Also, post-trough situations tend to be associated with a greater occurrence of lower- and mid-level moisture (necessary for riming), in contrast to pre-frontal situations when moisture is fairly often confined to the upper atmospheric levels in Colorado. Thus, these posttrough and post-precipitation periods can present very good seeding opportunities, especially if the air mass has become well-mixed at elevations below the crest height (i.e. no low-level stability issues). 22

28 3.1.3 Additional Analysis of Icing Occurrence According to Synoptic and Precipitation Categorizations Utilizing the five synoptic and the five precipitation categories as described in Sections and 3.1.2, respectively, allows a more complete analysis of icing occurrence according to each resulting sub-category. This provides additional information which can be useful for identifying general patterns which, in turn, can aid in seeding opportunity recognition. In order to classify the data in this way, both the ice detector data and corresponding high-resolution precipitation sensor data are necessary. Table 3-1 shows the observed icing totals in the main categories and subcategories. The total numbers of observed icing cycles (de-icing heat cycles) for both seasons of operation are shown. Table 3-1 Two-season Total Icing Cycles by Synoptic/Precipitation Sub-Category Pre-frontal Between Front/Trough Post 500-mb Trough Closed Low Undefined Category Totals Before Precip During Precip Between Precip After Precip No Precip Category Totals As shown in Table 3-1, the greatest amount of icing at Mt. Crested Butte for the two season period, occurred in the pre-frontal/between precipitation sub-category. There were 39 different storm periods in the last two winter seasons in which measured icing was classified as pre-frontal/between precipitation. However, it must be noted that a single case (or storm event) with lots of icing can occur in which one event accounts for most of the icing within a sub-category, possibly creating a bias due to a strong outlier. During the last two seasons, this did not seem to be an issue as most storm events had few icing cycles and these cycles were well distributed. The event with the most icing in one sub-category was on November 14-15, 2014, with 9 icing cycles measured within the pre-frontal/between precipitation category. However, with 23

29 26 icing cycles measured within the sub-category this event is likely not creating a large bias in the overall results. Table 3-2 shows the number of storm periods represented in each sub-category. As for storm events with icing, the greatest number of storm periods occurred in the between front/trough and between precipitation category with 7 storms. Table Total Icing Cycles by Synoptic/Precipitation Sub-Category Before Precip. During Precip. Between Precip. After Precip. No Precip. Category Totals Pre-frontal Between Front/Trough Post 500-mb Trough Closed Low Undefined Category Totals Based on two seasons of data in this sub-category analysis, it seems as if prefrontal situations present some of the best seeding potential in the Crested Butte area, with the most icing occurring in the between precipitation sub-category. Due to the small sample size of only two seasons, it is impossible to tell at this point if pre-frontal situations always offer the best seeding opportunity in the Crested Butte area. Icing detectors in Utah have indicated that the best seeding situations often occur in post-500 mb trough situations (Griffith, et al, 2013). More seasons of data are required to determine if this region of Colorado truly has better seeding situations in pre-frontal conditions or if last season was abnormal. Another potential reason for the difference may relate to the location of the icing detector itself. The Crested Butte ice detector is located to the north of the summit of Mt. Crested Butte and may experience terrain blocking or some other very small scale feature that may be causing differences in the results compared to a location at or near the summit (unfortunately, no access to the summit was available which prevented siting the equipment at this location). Results in Utah have indicated that the 500-mb trough passage is normally followed by subsidence and clearing skies with most cloud development and precipitation activity in the post- 24

30 trough synoptic category being due to orographic and/or convective processes, which can present good seeding opportunities (Griffith, et al, 2013). 3.2 Storm Temperature Structure during Icing Periods The temperature structure of the cloud bearing layer up to about 500-mb (~18,000 feet), and particularly below the 700-mb (~10,000 feet) level, is a major factor regarding the seedability of a given cloud system. Three key factors pertaining to the atmospheric temperature structure, especially in relation to use of ground-based seeding are: The height of the seeding material maximum (warm) nucleation activation temperature threshold (~ -5 0 C) relative to the mountain barrier height. The temperature range within the SLW layer, since that affects natural and induced ice particle habits (shapes) and their growth rates. The degree to which the atmosphere from the surface to the top of the SLW layer is thermodynamically stable or unstable. The temperature factors can be reasonably well assessed (accounted for) by monitoring of the mountain barrier summit temperature. The stability issue is more fully explored in the next sub-section. The maximum (warmest) nucleation activation temperature threshold for the fastacting silver iodide formulations acting as condensation-freezing nuclei, as used in this project, is approximately -5 0 C with the nucleation rate increasing exponentially at colder temperatures. For ground-based seeding releases, especially in stratiform clouds, it is thought that cloud systems with mountain summit temperatures warmer than -5 0 C offer little opportunity for enhancement using silver iodide. This is due to the relatively weak upward vertical air motion associated with stratiform-type cloud systems. Orographic (terrain-induced) lift can help loft ground-based releases, but consideration must be given to the prevailing wind velocities at varying heights and the barrier shape in terms of the time (distance) available for ice particle growth and eventual fallout into the intended target region. Research has suggested that seeding plumes from groundbased releases fairly commonly rise to 1,000 feet or more above the terrain, even in stratiform situations (Super, 1999). The more-convective cloud types loft the seeding material even higher into the SLW zone, allowing effective ground-based treatment in even warmer situations. Airborne seeding could potentially be more effective in warmer/stratiform circumstances if other conditions (e.g., winds and SLW concentrations) were favorable although a rather expensive alternative and not one 25

31 particularly suited to the Upper Gunnison River cloud seeding programs for various reasons among which would be terrain clearance considerations in obscured flight conditions (Instrument Flight Rules) rendering the ability to fly at effective altitudes upwind of the various target mountain barriers questionable.. Ground-based seeding with liquid CO2 or propane may expand the warm end of the seedable temperature range somewhat (perhaps up to -2 0 C) if certain cloud conditions and barrier configuration factors are satisfied although the growth rates of ice crystals created at -2 0 C are much slower that ice crystals created at -6 0 C which means that crystals formed at these temperatures would take longer to grow into snowflakes of sufficient size to begin to fall to the ground. Smaller snowflakes would also have less mass which is directly related to the water content of the snow. An exception to the above discussion occurs in springtime events when the atmosphere typically becomes more unstable during storm passages. This instability can cause the seeding material to rise to higher (colder heights) in the atmosphere than in the more stratiform clouds that typically occur in fall and winter storms. In other words some spring storms may be seedable even if the height of the -5 0 C level is above the target mountain barriers. Earlier research has indicated that the natural precipitation efficiency of some cloud systems can be quite high and that glaciogenic seeding (ice phase seeding like that caused by silver iodide) of those cloud systems will likely not yield appreciably more precipitation than is occurring naturally (Griffith, et al, 2013). Naturally high production of ice particles in these clouds is thought to produce near-optimum ice particle concentrations in the precipitation formation regions of the clouds. It is believed that a mountain-top temperature colder than approximately C is one indication that this naturally efficient situation likely exists. Another indication could be deep clouds with cold cloud top temperatures; e.g., C (Grant and Elliott, 1974). The data were analyzed to characterize the seasonal temperature characteristics from the cloud seeding perspective. The ice detector site is at ~ 11,400 feet near the summit height of the adjacent mountainous terrain. Figure 3.6 shows the distribution of the ice detector temperatures during each 15-minute observation period with sufficient icing to trigger the detectors de-icing heater. It should be noted that, since 2 0 C increments were most reasonable for displaying the data in these figures, statistics regarding the -5 to C temperature window were calculated separately and are summarized in the text. A majority (~75%) of the total measured icing during one season of data occurred within the favorable summit temperature window of -5 0 C to C. Approximately 9% of the icing periods were colder than C, and 16% occurred at temperatures warmer than -5 0 C. A significant amount of icing at temperatures colder than C would suggest that seeding opportunity is being missed at colder temperatures, according to the generalized seeding criteria used, but 26

32 < to to to to to to to -8-4 to -6-2 to -4 > -2 Percentage of Icing Cycles the observed low percentage in this category suggests that in general, little seeding opportunity exists at summit temperatures colder than C. 25% 23.8% 21.4% 20% 15% 10% 5% 0% 3.6% 2.4% 10.7% 2.4% 9.5% 15.5% 4.8% 4.8% 1.2% Temperature Range (C) Figure 3.6 Temperature Distribution during Icing Periods for Season NAWC frequently uses the 700-mb level (~ 10,000 feet MSL) to which seeding criteria can be indexed (refer to Table 1-1). The 700-mb level frequently approximates the average barrier heights of seeded mountain ranges in the western United States. This is not the case for the Upper Gunnison River target area which encompasses several higher mountain ranges. The barrier heights of these mountain ranges are closer to 11,000 12,000 MSL. Since the Mt. Crested Butte icing meter site sits at an elevation of ~11,400, its temperatures during storms observed at this location should be representative of the mean barrier crest heights. As indicated in Table 1-1 this elevation should be at the correct height to assess seeability as was done in the previous paragraph. The basic concept is if air masses are forced over the target mountain barriers then the temperatures would be indicative of where silver iodide nuclei have been shown to be effective in other programs. In other words, silver iodide nuclei can add to the natural ice nuclei background which is relatively low in the -5 0 C to C temperature range. 27

33 The data from this site suggests that little seeding opportunity is being missed at the cold end of the spectrum, although real-time monitoring of ice detector data may lead to recognition of some seedable storm periods at temperatures below C. At the warm end of the temperature spectrum, a small amount of icing (roughly around 16%) has been observed at temperatures warmer than -5 0 C. Regarding the warm end of the spectrum, it should also be emphasized that at least some of the icing occurrence at these warmer temperatures is likely seedable from valley or foothill based sites, primarily during spring storm situations with relatively deep atmospheric mixing. Wellmixed atmospheric conditions often allow some of the seeding material to be quickly carried to elevations well above the crest height, where temperatures are colder. 3.3 Icing and Wind For the season and the seasons combined, wind plots were produced for all icing events occurring at temperatures between 0 and C. This helps to provide additional insight regarding the types of wind patterns that are likely favorable for seeding operations (e.g. where should ground generators be located to be upwind of Mt. Crested Butte and in more general terms upwind of the other target barriers?). The wind directions were those from the Mt. Crested Butte site. Figure 3.7 contains wind plots for icing periods during season, while a two-season total of winds plots for icing and temperatures are presented in Figure

34 Figure 3.7 Wind Plots for Icing and Temperatures, Season 29

35 Figure 3.8 Wind Plots for Icing and Temperatures, 2-Season Composite Both figures illustrate that most of the icing cycles are observed in southwesterly wind situations (i.e. winds blowing towards the northeast). There are a few cases with an easterly component, but overall most icing situations have westerly component winds, this is expected as storms typically move from west to east and hence why most seeding generators are placed west of the intended target areas. The plot shows that the majority of icing occurred during southwesterly flow but some did occur in northwest flow. 30

36 In Figure 3.9, wind direction and speed at the 700-mb level were taken from the Grand Junction sounding while liquid water was being reported during seeding at the radiometer site (e.g. either 00Z or 12Z sounding times) for the winter season. This scatterplot shows a much different distribution than winds compared to icing meter cycles in Figure 3.8. Figure 3.9 shows a more evenly distributed wind field between a southerly wind direction to a northerly/northeasterly wind direction. The majority of these icing events fall into a westerly wind component. The wind plots for icing are surface winds, but seeing that the surface is around 750-mb, these plots can be comparable. Figure 3.9: 700-mb Wind Direction from Grand Junction during Seeded Events 31

37 4.0 Microwave Radiometer Specifics and Data Analyses A Radiometrics (headquartered in Boulder, Colorado) microwave radiometer, model MP-3000A, was installed in the fall of 2015 at N, W at an elevation of 8166 feet MSL. Commercial power and an internet connection were available at this site. Figure 19 provides a photo of this unit. This is a passive system that provides vertical profiles of temperature, relative humidity and liquid parameters in a continuous fashion through the entire atmosphere above the radiometer. Figure 4.1 Microwave Radiometer Sited at the #20 Manual Cloud Seeding Generator Location (refer to Figures 1.2 and 1.3) 32

38 4.1 Radiometer Liquid Water and Icing Meter Comparison As mentioned in the previous section, the radiometer stationed in the target area was able to record liquid water with respect to time and height. This coupled with temperature data can show where water is potentially supercooled. This then can be compared to the icing meter location at Mount Crested Butte to determine any similarities or differences. There were twenty-one seeded events during the operational season. A few events were excluded in this comparison due to bad/missing radiometer data. Figure 4.2 shows the information archived during seeded periods from the icing meter at Mount Crested Butte. The top panel is rain rate, the middle panel is temperature, and the bottom panel is icing cycles. Figure 4.3 is a radiometer cross section plot from the same time period on December 23, The panels that were compared are at the bottom of both images. The time scale on the icing meter is in local time whereas the radiometer is in Zulu time. The seeding period was on the 23 rd, and the only icing cycle on this date were around 1800 MST and 2400 MST. The radiometer shows liquid water occurring at that time with temperatures -5 0 C or colder. It seems that where the icing cycles were occurring was also where maxima s in liquid water appear on the radiometer cross section plots. This comparison was done for each of the seeded events mentioned in the above paragraphs. The rest of the images can be found in Appendix A. 33

39 Figure 4.2: Icing Meter Data from the Mt. Crested Butte Site for December 23,

40 Figure 4.3 Radiometer Cross-Section Plot for December 23, 2015 Seeding Event When comparing the two sites during seeding times of the chosen events, it seemed that the radiometer was showing much more liquid water than the icing meter was recording. This makes sense, as it is likely that the icing meter is under reporting the amount of icing occurring. One of the likely reasons for this occurring would be that the icing meter is a fixed-point measurement in the atmosphere, whereas the radiometer provides integrated values of the presence of SLW from the surface to the top of the atmosphere. The location of the icing meter verses the location of the radiometer may be another possibility affecting the difference in sensor readings. The height of liquid water on the radiometer cross section plot in Figure 4.3 seems to be occurring between the 400 and 500-mb level. This is likely not correct as water at this level would be completely frozen, which we know is not the case during seeding events knowing atmospheric temperatures. This seems to be a consistent trend on all of the radiometer cross-section plots. 35

41 4.2 Stability Derived from Radiometer Data The radiometer placed in the Gunnison target area was a very valuable tool when it comes to determining stability during seeding events. The closest atmospheric rawinsonde sounding location to the Gunnison River Basin is located in Grand Junction, which is miles away, depending what side of the target area one is being considered. This sounding can show the general atmospheric conditions for western Colorado, but the radiometer has the ability to create a sounding from the data it collects during its scans which are at a much closer distance to the downwind target area. It also provides continuous measurements instead of just two measurements a day, as those provided by rawinsondes. These radiometer derived soundings can be very useful in determining stability not only during real-time seeding events, but also after the fact in order to compare inversions and liquid water content to the actual seeded periods. Table 4-1 displays the seeded storm periods that were selected for the stability analyses. It is important to keep in mind when examining different seeding events in this table that all seeding events are not equal and staff meteorologists take many different variables into account on when to seed and how much seeding will be done. If a particular weather event is not forecast to be too impressive in terms of precipitation, or if stability is present, very few generators will be used. Table 4-1: Seeding Information for Stability Cases during the Storm No. Date(s) No. of CNGs Used Generator Hours 1 November November November December December December January January February February March March March March March

42 Cross-sections were created from the radiometer data collected during each seeding period. The software used to create these cross sections from the radiometer data is called The Rawindsonde Observation Program (RAOB). The RAOB program displays cross-sections with a number of times available where the radiometer was collecting data and can then be used to plot the data into a Skew-T diagram (a common tool used in meteorology). This diagram shows the user a vertical profile of the atmosphere (temperature, dew point and winds) at that moment in time. Figure 4.4 shows a Skew-T diagram taken from radiometer data collected on March 23, This particular diagram can also identify where atmospheric inversions are present. Figure 4.4 Skew-T diagram generated from Radiometer Data, March 23,

43 These cross-sections contain three panels that represent, from top to bottom, temperatures in 0 C, liquid water (g/m 3), and any inversions. Two vertical lines have been drawn across the liquid water panel to indicate seeding start and stop times. The black line represents when seeding started and the white line indicates when seeding concluded. The radiometer cross-section plots are in Coordinated Universal time (UTC). Figures 4.5 to 4.19 show each of seeded cases involved with the stability analysis. 38

44 Figure 4.5 Nov , 2015 Radiometer Cross-Section 39

45 Figure 4.6 Nov. 20, 2015 Radiometer Cross-Section 40

46 Figure 4.7 Nov. 27, 2015 Radiometer Cross-Section 41

47 Figure 4.8 Dec , 2015 Radiometer Cross-Section 42

48 Figure 4.9 Dec , 2015 Radiometer Cross-Section 43

49 Figure 4.10 Dec. 23, 2015 Radiometer Cross-Section 44

50 Figure 4.11 Jan. 20, 2016 Radiometer Cross-Section 45

51 Figure 4.12 Jan , 2016 Radiometer Cross-Section 46

52 Figure 4.13 Feb.14-15, 2016 Radiometer Cross-Section 47

53 Figure 4.14 Feb , 2016 Radiometer Cross-Section 48

54 Figure 4.15 Mar. 6-7, 2016 Radiometer Cross-Section 49

55 Figure 4.16 Mar , 2016 Radiometer Cross-Section 50

56 Figure 4.17 Mar , 2016 Radiometer Cross-Section 51

57 Figure 4.18 Mar , 2016 Radiometer Cross-Section 52

58 Figure 4.19 Mar , 2016 Radiometer Cross-Section Table 4-2 illustrates the seeding period, characteristics regarding any inversions that were detected as well as general comments about the seeding period and atmospheric conditions. Wind data were also collected for this figure. Sounding time was chosen from the available radiometer data during seeding and when liquid water was occurring. About 20% of the selected seeded events revealed no inversions. The majority of the seeding events did have some sort of an inversion, but these inversions were often just near the surface and very shallow and a number of these were weak 53

59 inversions which would likely not trap seeding material released at or above the elevation of the radiometer site. An item to note is that the elevation of the radiometer is 8166 feet MSL. The majority of the inversions detected from the radiometer were within a couple of hundred feet of this elevation. There were some elevated inversions but the temperature at the base of the lowest of these inversions was usually colder than -5 0 C so seeding with silver iodide should be effective in these situations. There were fewer inversions during the month of March. The top of all of the lower level inversions in Table 4-2 are in the range of 8500 to 8650 feet MSL which were within a few hundred feet of the surface. It is NAWC s understanding that a separate instrument was installed by Radiometrics at the radiometer site to record surface temperatures. Since there were frequent surface inversions indicated in the radiometer cross-section plots, as summarized in Table 4-2, NAWC acquired surface temperature and wind data from the Gunnison-Crested Butte Regional Airport to compare with the surface temperatures at radiometer site for the time periods provided in Table 4-2. This can then show how radiometer data compares to an actual surface station in the area. The surface site at Gunnison and the radiometer are at fairly similar elevations. The radiometer is around 8166 feet while the surface site is around 7677 feet. Table 4-3 shows temperature and wind data from the Gunnison surface state and how it compares to the surface temperature at the radiometer site. From Table 4-3, it seems that the radiometer surface temperature is, in most cases, colder than the surface temperature from the Gunnison surface site. This renders the surface temperature at the radiometer site questionable during the periods indicated to be associated with low-level (near surface) inversions since the coldest temperatures tend to pool at the lowest elevations under stagnant atmospheric conditions. Looking at the four cases with apparently reliable data from the radiometer that also indicated moderate to strong surface inversions (Nov , Dec. 12, Jan , and Jan ) three out of these four cases had colder surface temperatures at the radiometer site compared to the Gunnison site (the Jan case was the exception). These three periods were also associated with light surface winds at the Gunnison site which suggests a pooling of cold air in the lower elevations of the Gunnison valley. We conclude that low-level inversions would not have been a significant factor in the dispersion of seeding material from NAWC s ground based generators during storm events considered seedable during the winter season although there were a few events where seeding was not attempted from certain sites due to the presence of low-level inversions. In cases where low level stability is present during storm events, higher elevation sites were used. 54

60 Table 4-2: Seeded Storm Events with Stability Comparison Results during the Season Date Seeding Start (Z) Sounding Time Nov , (11-17) Base of Inversion (feet) Temperature at base of inversion Top of Inversion (feet) Inversion strength ( 0 C) 700-mb temperature ( 0 C) General Comments Strong inversion at seeding start but eventually dissipated. High amounts of liquid water present during seeding. Nov. 20, Very weak inversions present. Minimal seeding conducted due to light precipitation amounts and marginally warm temperatures. Nov. 27, (11-27) Small amounts of liquid water present. Minimal seeding conducted due to only marginal amounts of precipitation occurring on radar. Dec. 12, Stronger inversion developed later in seeding period. Large amounts of liquid water present. 55

61 Dec. 15, Discrepancies between radiometer sounding and cross section temperature. Dec. 23, Majority of stability before seeding period. Jan 20-21, (1-21) Jan , (1-31) Feb , (2-15) Feb , (2-23) Strong inversion that precluded seeding Weak inversion develops near end of seeding event. N/A N/A N/A N/A N/A N/A Large amounts of data missing from radiometer (2-23) Cold event, with the 5 0 C height near 750- mb. Moderate amounts of liquid water present during seeding. Very weak inversion present. Mar. 6-7, High amounts of liquid water during seeding. A few weak inversions were present. Mar , (3-15) 1700 (3-15) -2.4 Liquid water convective in appearance. Moderate amounts of liquid water present 56

62 during seeding. Mar , Liquid water convective in appearance. Somewhat warm event, little liquid water present. No inversions present. Mar , (3-25) Mar , (3-30) 1 No inversions were present -9.2 Liquid water convective in appearance. High amounts of liquid water during seeding. No inversions present Liquid water convective in appearance. Multiple periods of high liquid water occurrences. No inversions. 57

63 Date Table 4-3 Radiometer and Gunnison Surface Site Comparison During Seeded Events Surface temperature at Radiometer Site ( 0 C) Surface temperature at Gunnison ( 0 C) Surface Wind Speed (mph) Time (Zulu) 11/17/ /20/ /27/ /27/ /12/ /12/ /15/ /15/ /15/ /21/ N/A /22/ /22/ /20/ /31/ /14/ /14/ /23/ /7/ /15/ /23/ /25/ /30/ /30/ NAWC s recent peer reviewed paper (Griffith, et al, 2016) that addressed the low-level stability issue on this program indicated that: 1). The ice detector data paired with surface data and modeling indications suggests that in ~63-70% of icing periods, atmospheric stability would not significantly inhibit orographic lofting of lower elevation ground generator seed- ing plumes into the SLW zones in winter clouds associated with storms passing over the target area. 58

64 2). HYSPLIT computer model runs as well as plots indicating the amount of stability during each seeded storm event were compiled for the winter season. These predicted plumes for the seeded cases indicated successful transport of seeding material over the intended target area in a large majority of the cases. 5.0 Questions, Preliminary Conclusions and Recommendations Questions, some general conclusions, and recommendations are provided in this section. Additional seasons of data collection will help increase confidence in or possibly modify the preliminary conclusions. 5.1 Questions Perhaps of some importance is the fact that Mt. Crested Butte is a rather isolated prominent feature which is separated from the major mountain barriers to the north (Sawatch) and to the west (West Elks). This fact may impact the amount of icing that is observed in ways that may be difficult to quantify. The more typical situation is where there is a substantial mountain barrier identified as the target area of a winter cloud seeding program. The conceptual model in this situation regarding the production of supercooled liquid water (SLW), the target of seeding in winter storms is that the lifting of a moist air mass over the barrier (an orographic effect) can generate SLW over the upwind slopes of the barrier. Mt. Crested Butte offers less of an obstacle to wind flow than the larger surrounding mountain ranges. The question might be whether winds flow around instead of over Mt. Crested Butte in certain conditions in which the orographic effect would be less and possibly the production of SLW less. Of course some storms bring SLW with them without the need of orographic forcing to generate SLW. Another factor to consider is that the West Elk Mountains located west of Crested Butte Mountain will remove water vapor and may remove SLW from storms as they pass over these mountains leaving less moisture in the air mass as it passes over Mt. Crested Butte (i.e. perhaps less icing). This process is referred to as the rain shadow effect in meteorology. The fact that the icing meter site is at approximately the same elevation as the barrier summit of the West Elks may lessen this effect. These are complex issues; nonetheless SLW was observed last winter at the icing meter site on a number of occasions. The microwave radiometer consistently indicated supercooled liquid water (SLW) at high altitudes; frequently at or above 500mb where temperatures were ~-20 0 C or colder. NAWC had previously noticed this tendency when a Radiometrics microwave radiometer was deployed in the city of Gunnison for the months of March and April During the winter season, NAWC asked Radiometrics whether the height resolution of the SLW on the radiometer was accurate. The reply was perhaps 59

65 not. NAWC believes the quantities of SLW are probably accurate but the locations are not. Based upon numerous earlier research programs, as summarized in Griffith, et al, 2013, it is highly likely that the SLW is located at much lower altitudes nearer to the surface. This is an important issue; if the SLW was actually located at ~500mb then the ability of ground generator plumes reaching these levels would be very limited which would suggest airborne seeding would be required. However, NAWC peer reviewed evaluation of the Upper Gunnison River program indicates that the ground-based generator network has been effective in producing precipitation increases in the intended target areas (Griffith, et al, 2009). 5.2 Conclusions / Key Findings It needs to be emphasized in the following conclusions and key findings that all of these are based upon only two winter seasons of data. Additional seasons of data would lead to a more climatological representations of the conditions associated with icing at the Mt. Crested Butte site. Although some meteorological patterns show good correlation to the occurrence of seedable conditions, providing reasonable prediction capability, it has become clear that real-time monitoring of the ice detector data is of significant value in seeding opportunity recognition. Further, post-hoc analyses of the data can yield considerable insights into winter storm seedability. We consider the near ridge-top ice detector system to constitute an exceptional value to the seeding program. Icing at a given ground-based location is sensitive to the surrounding terrain and instrument exposure, as well as site-specific wind conditions. One implication is that lifting of the low-level (valley) air mass over significant mountain barriers tends to be much more efficient in certain areas, such as canyons and terrain catch areas. This is often an important consideration in terms of the location of ground-based seeding sites (Super, 1999). A comparison was made between icing meter indications of SLW versus those obsereved by the radiometer during seeded periods. It seems that where the icing cycles were occurring this is also where maxima s in liquid water appeared on the radiometer cross section plots. As expected, SLW was observed much more frequently by the radiometer. 60

66 Of five synoptic categorizations (pre-frontal, post-frontal/pre-500-mb trough, posttrough, closed low, and undefined), the post trough situation was associated with the greatest amount of icing at the Mt. Crested Butte site during the season with 36% of icing events. The second most common category was pre-frontal with 32%. This pattern seems to be different than analysis of icing data at two Utah locations that indicate the most significant icing occurs in post 500 mb trough conditions (Griffith, et al, 2013). Of five synoptic categorizations (pre-frontal, post-frontal/pre-500-mb trough, posttrough, closed low, and undefined), the pre-frontal was associated with the greatest amount of icing at the Mt. Crested Butte site during two seasons with 40% of icing events. The second most common category was post-trough with 24%. In relation to precipitation, icing was observed most frequently between precipitation periods with 35% of the icing in this category. Icing was observed after measured precipitation in 16% of cases with 31% of icing associated with non-precipitating cloud systems during this season of data collection. In relation to precipitation, during two seaons of data, icing was observed most frequently between precipitation periods with 44% of the icing in this category. Icing was observed after measured precipitation in 19% of cases with 21% of icing associated with non-precipitating cloud systems. Because the ice detector is a fixed-point measurement, there may be significant SLW in the vicinity at times (for example, at a higher elevation) that the icing sensor does not detect. Thus, these measurements are believed to be underestimates of the total SLW occurrence. A means is available to measure the total SLW column in the atmosphere above a given point using a portable microwave radiometer. The microwave radiometer indicated there were frequent periods of SLW although the heights of these SLW zones was questionable. The radiometer observations also indicated that there were rather frequent very shallow low-level inversions of varying strengths during potentially seedable storm periods. The base of these inversions was frequently at the surface or only ~ feet above the surface. Since it is our understanding that the ground level surface temperature at the radiometer site was derived from a separate instrument, these inversions could be influenced by this surface sensor not providing accurate.temperatures. We conclude that low-level inversions would not have been a significant factor in the dispersion of 61

67 seeding material from NAWC s ground based generators during storm events considered seedable during the winter season although there were a few events where seeding was not attempted from certain sites due to the presence of low-level inversions. In cases where low level stability was present during storm events, higher elevation sites were used. A recent NAWC peer reviewed paper (Griffith, et al, 2016) suggests that ~63-70% of icing periods, atmospheric stability would not significantly inhibit orographic lofting of lower elevation ground generator seeding plumes into the SLW zones in winter clouds associated with storms passing over the target area. 5.3 Recommendations Continue operation of the ice detector site network and analysis of site data. If funding is available, provide a portable microwave radiometer for additional winter seasons (consider either purchase or lease options) to provide additional information on SLW and low-level inversions during winter storms that impact the Upper Gunnison River cloud seeding target area.. Due to the complex nature of the Upper Gunnison River target area, it would be useful to establish additional strategically located ice detector site(s) as budgets will allow. Possible locations would include one at the summit of Monarch Pass and at an accessible, manned point along the ridgeline of the La Gardia Mountains located south of Gunnison. Acknowledgements The support and cooperation of the following agencies is gratefully acknowledged: The Crested Butte Mountain Resort, Colorado Water Conservation Board, Gunnison County, Upper Gunnison River Water Conservancy District, and the Three Lower Colorado River States (Arizona, California and Nevada). References Grant, L.O. and R.D. Elliott, 1974: The cloud seeding temperature window. American Meteorological Society, Journal of Applied Meteorology, 13, p

68 Griffith, D.A., D.P. Yorty, T.W. Weston and M.E. Solak, 2013: Winter Cloud Seeding Windows and Potential Influences of Targeted Mountain Barriers. Weather Modification Association, Journal of Weather Modification, 45, p Griffith, D.A., S.M. Ward, and D.P. Yorty, 2016: Analysis of Ice Detector Observations at Mount Crested Butte, Colorado During the Winter Season. Weather Modification Association, Journal of Weather Modification, 48, p Super, A.B., 1999: Summary of the NOAA/Utah Atmospheric Modification Program; Weather Modification Association, Journal of Weather Modification, 31 p Yorty, D.P., T.W. Weston, M.E. Solak and D.A. Griffith, 2012: Low-Level Atmospheric Stability during Icing Periods in Utah and Implications for Winter Ground-Based Cloud Seeding. Weather Modification Association Journal of Weather Modification, 44, p

69 November 16-17, 2015 APPENDIX A

70 November 20, 2015

71 November 27, 2015

72 December 11-12, 2015

73 December 14-15, 2015

74 December 21-22, 2015

75 December 23, 2015

76 January 8, 2016

77 January 20, 2016

78 January 30-31, 2016

79 February 14-15, 2016

80 February 22-23, 2016

81 March 6-7, 2016

82 March 14-15, 2016

83 March 22-23, 2016

84 March 25-26, 2016

85 March 29-30, 2016

86 April 5, 2016