AN OVERVIEW OF THE 2016 RAINFALL ENHANCEMENT ACTIVITIES IN TEXAS : A MORE INTENSIVE USE OF HYGROSCOPIC MATERIAL

The Texas Weather Modification Association oversees four rainfall enhancement projects which combined have completed their 21st season in 2016. These projects include the Panhandle Groundwater Conservation District, the South Texas Weather Modification Association, the Trans-Pecos Weather Modification Association, and the West Texas Weather Modification Association. Classic Thunderstorm Identification, Tracking, Analysis and Nowcasting evaluation of seeded and control clouds for each project are analyzed to determine the effect seeding operations had on different cloud variables. A more recent evaluation analysis examining the results of dual seeding (hygroscopic plus glaciogenic material) compared to seeding with only glaciogenic material. Excellent results were achieved during the 2016 season with average precipitation increases of 1.34 inches above seasonal value, and over 2 million acre-feet of addition precipitation from cloud seeding activities. An analysis of hygroscopic seeding continues to show its importance for effective seeding operations. Clouds seeded with both glaciogenic and hygroscopic material lasted longer and produced more precipitation than similar clouds seeded with only glaciogenic material. ABSTRACT Kendell laRoche1, aRquimedes Ruiz-columbié2, Jonathan Jennings3


INTRODUCTION
While multiple cloud seeding programs have been operating in Texas over the past few decades, four rainfall enhancement projects continue to remain active across the state.These active projects include the Panhandle Groundwater Conservation District (PGCD) based in Whitedeer, the South Texas Weather Modification Association (STWMA) based in Pleasanton, the Trans-Pecos Weather Modification Association (TPWMA) based in Barstow, and the West Texas Weather Modification Association (WTWMA) based in San Angelo (Figure 1).The WTWMA project began cloud seeding in 1995, STWMA in 1997, PGCD in 2000, and the TPWMA project in 2003.The objective of each project is to enhance the precipitation across each project area, in the safest and most cost-effective manner possible.During cloud seeding operations, all storm cells capable of being seeded are investigated.No randomized seeding is conducted across any of the current Texas rainfall enhancement projects.
Determining which clouds, or cells, to investigate is dependent on the seeding potential and where these cells develop.Cells with strong updrafts within, or moving into, the target area are viable candidates for seeding.All cells that show signs of potential cloud seeding are investigated and may be seeded to maximize development.These cells can either be identified by radar, satellite observations, or identified visually.
Certain criteria also warrant a suspension of cloud seeding activities.Cloud seeding operations will stop immediately if a targeted cell becomes tornado warned, flash flood warned, or has a flood warning from the National Weather Service.The STWMA project will stop operations if a Severe Thunderstorm Warning is issued for the targeted cell.Operations can resume once these warnings expire.Occasionally, vast areas within a project might receive tremendous amounts of rain over a period of several days, leading to a suspension of seeding operations for a few days.Project personal will remain in contact with local and county representatives to help determine when it is safe to resume operations.
Detailed evaluations are critical for determining the impact cloud seeding activities have on precipitation for each project.Evaluations analyze small clouds (precipitation mass less than 10,000 kilotons (kton)), large clouds (precipitation mass greater than 10,000 kton), and type B clouds (clouds seeded when they were already one hour old or older), and an analysis process called microregionalization. Micro-regionalization is used to analyze precipitation increases county-by-county within each project.This analysis has led to a better understanding of the results per county and immediately outside the project area.To determine if a change in precipitation occurred per county, the estimated additional precipitation from cloud seeding was divided by the seasonal value to find the percent increase.This method helps determine the percent change for all counties within each project.The initial seeding, extended seeding, and acre-feet increase values for each project were then summed up.Average values for additional precipitation, county seasonal precipitation, and percent increase for each project were also calculated.All project values were then summarized for a final 2016 summary for the state of Texas.Data related to the amount of acre-feet of water per project generated by cloud seeding has received more attention over the last few years.While this data on the generated water may not be critical to cloud seeding results, this data is very important and better understood by various customers and the water resource community.

Texas Weather in 2016
Wide areas of central, southern, and southeastern The PGCD target counties all ended the year at or above normal rainfall totals.There were no tropical systems to affect the Texas coast during 2016.The radar-derived yearly rainfall for Texas is shown in Figure 2.

EQUIPMENT USED
Each program uses aircraft to deliver the cloud seeding material to the target clouds.Only cloud base seeding is used with individual burn-in-place flares attached to racks on the aircraft wings.Flares are ignited when the pilot encounters a strong enough updraft (at least 300 feet per minute) which results in the flare material being ingested into the cell.The pilot will then fly back and forth along the leading edge of updraft igniting flares until the pilot is instructed to stop, the updraft dissipates, conditions become unsafe, the cell moves out of the target area, or the aircraft runs out of flares.

ANALYSIS
The 2016 season began on 7 March over the WTWMA target area, and ended on 11 November over the STWMA target area.Despite having the third latest start date of 2016, the STWMA project had the most seeding flights, flew the most hours, used the most flares, and flew the most days.The PGCD has the most recon flights, which were flights where cells were investigated but seeding did not take place (Table 1).There were 119 operational days across all projects, with the most active period for all target areas from May to September.There were 306 seeded cells across all active projects within 2016, including 179 small seeded clouds, 81 large clouds, and 46 type B clouds.There were 0 missed seeding opportunities with clouds whose lifespan was longer than one hour.

Small Clouds
When analyzing data related to small clouds, values in parentheses in the right most column are modeled values, values in right most column not in parentheses are raw values, and η is defined as cloud efficiency.Efficiency is determined from the quotient of precipitation mass divided by cloud mass.During the 2016 season, 1181 AgI and 100 hygroscopic flares were burned within small clouds.The effective AgI dose per cloud was approximately 65 ice-nuclei per liter, with excellent timing (95% of the seeding agents were ingested into the clouds during their first half-lifetime).The precipitation mass increase of 126% and cloud mass increase of 45% provide evidence that seeded clouds grew at the expense of the environmental moisture.The precipitation mass was larger than the cloud mass for both the seeded and control clouds, indicating these clouds used only a fraction of the environmental moisture for their own maintenance.The increases in lifetime (40%), area (39%), volume (42%), volume above 6 km (45%), and precipitation flux (49%) are notable; while there were only slight increases in maximum reflectivity (1%), and in top height (3%).The sub-sample of small seeded clouds was about 56% more efficient than the sub-sample of small control clouds (Table 2).Seeding operations on small clouds would last about 8 minutes.

Large Clouds
For 2016, 2066 AgI and 185 Hygroscopic flares were burned in large clouds.Fewer variables were examined for large seeded clouds because small clouds occur more frequently.The effective AgI dose per cloud was approximately 95 ice-nuclei per liter, with perfect timing (100% of the seeding agents were ingested into the clouds during their first half-lifetime).Large clouds were generally 23 minutes old when seeding operations took place, while seeding operations would usually last 36 minutes.Large seeded clouds usually lasted longer (25% increase), had a greater area (23% increase), greater volume (22% increase), and had greater precipitation mass (61% increase) than non-seeded large clouds (Table 3).

Type B Clouds
A total of 1001 AgI and 82 hygroscopic flares were burned in type B clouds during 2016, with an effective AgI dose of approximately 90 ice-nuclei per liter.For 2016, 87% of the seeding material went into the clouds in their first half-lifetime, for a quasi-excellent timing.On average, type B clouds were 118 minutes old when seeding operations began, and operations would last about 32 minutes.
The lifetime of seeded type B clouds had a slight increase (4%) while the precipitation mass had a 10% increase over control clouds (Table 4).

Micro-regionalization
Increases in precipitation mass were analyzed county by county per project to better understand the performance and corresponding results per county.These results also consider seeded cells just outside the target area.Initial seeding is the number of cells seeded only within that county.Extended seeding is the number of cells seeded within the initial county plus the number of seeded cells from other counties which moved into the initial county.Extended seeding helps consider seeding propagation within, and immediately outside, the target area.The results for each county were then summarized, and the results for each program are shown in Table 5.
Examining precipitation which fell from seeded cells outside the target area accounted for 251,300 ac-f of water, or about 10% of the total precipitation increase.During 2016, 4248 AgI and 367 hygroscopic flares were burned during all Texas cloud seeding projects.The average increase in precipitation for each project was about 1.34 inches or 6.6% greater than the annual rainfall.
The WTWMA project had the largest increase above seasonal averages (2.02 in.or 9.3%), while the STWMA had the smallest increase (0.96 in.or 3.6%).Combining the results from all four programs, cloud seeding brought an additional 2,586,800 ac-f of precipitation to Texas in 2016 (assuming the precipitation reached the ground) (Table 5).

Hygroscopic Material Evaluation at STWMA and WTWMA
From 2009 thru 2011 (Table 6), a limited number of hygroscopic flares were used at both the STWMA hygroscopic seeding became operational.From Table 10, the STWMA average increases for cloud lifetime, area, volume, precipitation flux and precipitation mass after dual seeding became active were greater than or equal to increases before dual seeding became active for small and large clouds, however this was not necessarily the case for type B clouds.While the average lifetime increase for type B clouds after dual seeding activation was 3.6% compared to 3.2%, area, volume, precipitation flux and precipitation mass all had greater increases before dual seeding activation.

Evaluations for WTWMA
The WTWMA variable increases from 2007 to 2016 for small, large, and type B clouds are shown in Tables 11 -13.When examining the average variable increases before and after the start of operational hygroscopic seeding, both small and large clouds at the WTWMA had greater increases after dual seeding became operational for cloud lifetime, area, volume, precipitation flux, and precipitation mass.Type B clouds had greater increases before dual seeding activated for the same variables (Table 14).

Small Cloud Material Evaluation at STWMA
From 2012 to 2016 separate evaluations using only AgI material were calculated on small clouds for both STWMA and WTWMA to determine the variable differences between dual seeding and pure glaciogenic seeding.
When comparing the modeled increases of dual seeded and pure glaciogenic seeded small clouds at STWMA, the increases for dual seeding were generally greater except for variables involving measurement of radar reflectivity (Table 15).To summarize the results from Table 15, dual seeding and pure glaciogenic seeding averages were calculated for every variable from 2012 -2016 to evaluate any variable differences (Table 16).Cloud lifetime, area, volume, volume above 6 km, precipitation flux, mass, cloud mass, and efficiency increases were greater for dual seeding than pure glaciogenic seeding.Only for storm top height, maximum dBz, and top height of max dBz were pure glaciogenic seeding increases slightly greater than dual seeding.

Small Cloud Material Evaluations at WTWMA
The WTWMA project also had greater cloud lifetime, area, volume, volume above 6 km, precipitation flux, mass, and cloud mass increases for dual seeding than for pure glaciogenic seeding, however storm top height had the same average percent increase for both methods of seeding.The maximum dBz and top height of max dBz variables both had greater increases for pure glaciogenic seeding.The cloud efficiency was nearly the same for pure glaciogenic and dual seeding (Table 17 - 18).

ANALYSIS
Because weather patterns and precipitation amounts vary by year, daily performance analysis is calculated to determine how effective seeding operations were conducted.Daily operations were rated based on timing, dosage amount, material deposit location, increases in precipitation mass, and number of missed opportunities (Table 19).
For 2016, there were 78 operational days with excellent performance, 28 days with very good performance, 12 days with good performance, and one day with fair performance.The overall annual results for cloud seeding are evaluated as excellent because 1: 95% of the seeding agents were ingested into the cell during the cell's first half-lifetime, 2: the average glaciogenic dose was 75 ice-nuclei per liter, 3: small cloud precipitation mass increase of 126%, and 4: there were no missed opportunities with a cell lifetime longer than 45 minutes.

CONCLUSION
Notable increases in cloud lifetime, precipitation flux and precipitation mass were determined for small, large, and type B clouds for all projects during the 2016 season.The most noticeable variable increases were for small clouds, while type B clouds had the least noticeable increases.Type B clouds were seeded beyond their first half life, thus there was a much shorter time period when these clouds could be seeded.Because type B clouds were already established by the time seeding began, this could explain why these clouds did not respond as well to seeding as the small or large clouds.The micro-regionalization analysis showed increases in precipitation for each project, with an average increase of 1.34 inches, or 6.6%, above the seasonal value.In addition, 2,586,800 ac-f of additional precipitation fell because of cloud seeding activities across all 4 project areas.The largest increase above the seasonal average occurred at the WTWMA project, while the STWMA had the smallest.
The use of hygroscopic seeding during operations has become more important in cloud seeding activities across Texas since 2012.Hygroscopic seeding uses small salt particles to broaden the initial cloud droplet spectrum and accelerate the coalescence process.These introduced cloud condensation nuclei (CCN) are larger (> 0.3 μm diameter) than the natural CCN and will activate sooner within a cloud.The large CCN then start growing more efficiently by collision and coalescence (Bruintjes et al. 2012).Hygroscopic flares add more of these large CCN to the environment that what naturally occur.Using hygroscopic and glaciogenic material when cloud seeding enhances the cloud efficiency by allowing both the warm and cold rain processes to work in conjunction with one another.
When examining cell lifetime, area, volume, precipitation flux, and precipitation mass increases  Variable differences between dual seeding and pure glaciogenic seeding were also determined using data from small clouds for both STWMA and WTWMA from 2012 to 2016.Dual seeding had greater increases for cell lifetime, area, volume, volume above 6 km, precipitation flux, precipitation mass, and cloud mass for both the STWMA and WTWMA projects.Both projects had greater increases for pure glaciogenic seeding for maximum dBz and the top height of maximum dBz variables.The STWMA project had a greater increase in cloud efficiency from dual seeding, while pure glaciogenic and dual seeding had roughly the same efficiency at the WTWMA project (Sections 3.5.3-3.5.4).Cells across the STWMA project area appear to be naturally more efficient than cells over the WTWMA area.Even when neglecting the use of hygroscopic flares, natural cells over the STWMA area generally had greater efficiency values than over the WTWMA area (Tables 15, 17).Differences in air masses (continental in western TX, continental/maritime in southern TX) and increased amount of dust around WTWMA could be related to the differences between the WTWMA and STWMA results and efficiencies.
STWMA and WTWMA projects before and after the use of dual seeding, small and large clouds had greater increases for each variable after dual seeding became operational.Type B clouds had greater increases before dual seeding, except for STWMA lifetime (Sections 3.5.1 -3.5.2).
Control clouds have similar characteristics and are no more than 100 km away so to be immersed within the same environmental mesoscale and macroscale conditions as the seeded clouds.The control clouds must be within the same environmental conditions for an accurate comparison between seeded and control clouds.TITAN can analyze various aspects of seeded and control clouds including storm lifetime, area, volume, precipitation flux, and precipitation mass.The variable ratio between seeded and control clouds is used to determine the effect cloud seeding had versus control clouds.The (Ruiz-Columbie et al. 2003)nalysis and evaluation process to compare seeded clouds with non-seeded (control) clouds, and estimate what differences occurred between the two.evaluationused for each project in 2016 has been used by past rainfall enhancement projects across Texas(Ruiz-Columbie et al. 2003); thus, is referred to as a classic evaluation.

Table 1 .
Start date, end date, flight data, and flare data for each project during 2016.

Table 2 .
Variables and average results from classic TITAN evaluations for 179 small seeded clouds versus 179 small control clouds (Ruiz-Columbie 2016).

Table 3 .
Variables and average results from classic TITAN evaluations for 81 large seeded clouds versus 81 large control clouds (Ruiz-Columbie 2016).

Table 4 .
Variables and average results from classic TITAN evaluations for 46 type B seeded clouds versus 46 type B control clouds (Ruiz-Columbie 2016).

Table 6 .
Total number of AgI and hygroscopic flares used at the STWMA and WTWMA projects per year.

Table 7 .
Variable increases (%) for small clouds at STWMA per year from 2007 to 2016.Dual seeding became operational in 2012.

Table 8 .
Variable increases (%) for large clouds at STWMA per year from 2007 to 2016.

Table 9 .
Variable increases (%) for type B clouds at STWMA per year from 2007 to 2016.

Table 11 .
Variable increases (%) for small clouds at WTWMA per year.Dual seeding became operational in 2012.

Table 12 .
Variable increases (%) for large clouds at WTWMA per year.

Table 13 .
Variable increases (%) for type B clouds at WTWMA per year.