A Student Perspective of Working in the National Weather Service

During my 2016-17 winter break, I had the opportunity to work at the National Weather Service field office in Wichita, Kansas.  Working with Chris Robbins, M.S. here at iWeatherNet had already shape much of my knowledge of the National Weather Service, since he worked for 15 years within the agency, including NOAA’s National Severe Storms Laboratory, the National Hurricane Center, and the NWS in Fort Worth.  As a very strong advocate of the National Weather Service, Chris has encouraged me to seek student opportunities within the agency in preparation for a potential NWS career after graduation. 

Image of the remote sensing weather station and satellite at the NWS Wichita forecast office.

I have been fortunate to have already gained experience while working in the private sector of meteorology throughout my undergraduate career. Seeking the opportunity to work with the National Weather Service in Wichita was a capstone to my goals of rounding out my professional experience. Over the course of my month long winter break, I spent the entire time in Wichita working three days a week at the NWS forecast office. 

Wrapping up a forecast analysis with other NWS meteorologists shortly before they issued the  area forecast discussion on the afternoon of Monday, December 19th, 2016.

One of many goals was to understand how the National Weather Service field offices operate, particularly when it comes to forecast communications between emergency management, other NWS offices, and media. Over the past few years, the National Weather Service has increased their reach to the public through the usage of social media and mobile apps, and this has served as a crucial role in disseminating forecast and hazardous weather warning information.

During high impact weather events, such as the April 14th, 2012 tornado outbreak or the May 8th, 2009 derecho event, the National Weather Service is in constant communications with local media outlets and emergency managers. While this is not always seen on the public end, media outlets have broadcasted live discussions between NWS meteorologists and their meteorologists discussing the potential for hazardous weather — this has seen an increase since the rise of social media.

The Operational Forecasting

National Weather Service field offices produce forecasts usually four to five times per day, known as an Area Forecast Discussion (AFD). These AFDs are compiled with a short term and long term discussion, as well as aviation, and perhaps other various forecast entities if they are warranted (such as fire weather, marine and tidal forecasts, etc). Depending on the ongoing weather pattern, often determines the number of AFDs that are issued, as well as extra details such as mesoscale (severe weather) discussions, fire weather discussions, etc. The field offices also communicate with each other on a daily basis in order to unify their forecast products to increase forecast accuracy and consistency.

Additionally, the majority of National Weather Service field offices launch radiosondes, or weather balloons, twice a day, every day of the week, 365 days a year, at 1200 UTC and 0000 UTC. The NWS Wichita field office does not launch radiosondes; however, the neighboring offices in Dodge City and Topeka, Kansas, as well as the National Weather Service office in Norman, Oklahoma, do. 

Forecasting the weather for the upcoming week on December 17th, 2016, across the county warning area for the National Weather Service in Wichita, Kansas.

My first week at the office was a busy one, as it was about a week before the Christmas holiday. There appeared to be, at the time, a risk for thunderstorms across the State of Kansas. This was a great opportunity to see how convective forecasts are compiled and disseminated to the public, as travel densities increase during any given holiday of the year.

The operational model guidance a week before the holidays did, indeed, verify, as a squall line traversed across the State of Kansas on December 25th. The majority of the state received convective rainfall, and a record six tornadoes were observed and surveyed across the state. 

Simulated Warning Decision Training

I was very fortunate to perform simulated warning decision training during my tenure at the National Weather Service in Wichita. What is simulated training? The way it works is that previous weather events are archived in the computer system, and are automated to play again in realtime. Data such as all of the radar imagery, satellite data, surface observations, and other realtime mesoscale analyses are provided for the forecaster to assess the situation before the simulation starts. 

Many simulations were conducted, including the significant tornado outbreak that occurred on April 14th, 2012. This was an intense event, as numerous discrete supercells moved across the central Plains producing long-tracked tornadoes, including one that was rated as an EF-4 in central Kansas. 

Image of the beginning of the simulated training for the April 14th, 2012, tornado outbreak across central Kansas.

This was a particularly high stress situation, and took two work days to complete the entire simulation. Warning issuance is very strategic in many ways. Severe thunderstorm warning criteria is winds ≥58mph and hail ≥1.0″ in diameter. Tornado warning criteria revolves around the Doppler radar detecting strong enough rotation within the storm to warrant a warning, dual-polarimetric detecting debris, or confirmation from storm spotters and/or law enforcement. On a day like April 14th, the false alarm ratio was low as local storm reports and Doppler radar confirmed destructive tornadoes. 

One thing that I found particularly interesting was monitoring the supercells as they moved across northern Oklahoma and western Kansas before they entered the county warning area (CWA). The reason for this is that contemplating when to issue a warning became critical, because in the real world situation you want to issue a warning to give enough lead time without causing warning fatigue.

In the latter part of the simulation during the evening hours, an intense supercell moved across north-central Oklahoma and into the southern part of the county warning area. Ahead of it, I issued a severe thunderstorm warning for Harper County. However, the supercell already had a history of producing tornadoes.

My reasoning for issuing a severe thunderstorm warning before the core of the supercell entered into the county warning area was to give an appropriate lead time for the approaching storm. In the warning text; however, I had enhanced wording about the storm as well as statement that tornadoes could be possible. Eventually, as the hook echo approached the Kansas/Oklahoma border, a tornado warning was then issued. This was a highly stressful case, especially with having confirmed large and violent tornadoes reported by the public. It was a very rewarding experience to be able to do a simulation from such a significant event, however.


A multitude of other different case events were performed in the warning simulator, including the May 8th, 2009, derecho event that occurred across central Kansas. Unlike the April 14th, 2012, tornado outbreak, this was a linear event compared to discrete supercells. This made for an entire different tactic for hazardous warning issuance. 

The spatiotemporal distribution of warnings for this event was much different from other events that I performed on the simulator. Given that this was a high impact event that affected the more populated areas of Kansas, the decisions to keep warning times extended was on the table. As the derecho approached the northwestern tiers of the CWA, severe thunderstorm warnings were issued for periods of 45-60 minutes. This was done in order to reduce warning fatigue, and to give ample lead time to those in the path of the oncoming derecho. 

Image of the May 8th, 2009, derecho event simulation. This was at the peak of the event as the derecho moved through south-central Kansas.

Swaths of damaging straight-line winds in excess of 80mph were observed across much of the county warning area, including a wind gust surveyed at 120mph by the National Weather Service that occurred in New Albany, Kansas, that destroyed mobile homes. 

In the image above, you will notice that the severe thunderstorm warning polygon is rather large. I warned the entire southern tier of the county warning area with one severe thunderstorm warning polygon. This was done for a few reasons: 1) the timescale of which the warning was issued was 60 minutes; 2) the incoming derecho covered a large portion of that part of the county warning area.

Unlike tornado warnings where warning polygons can be confined and targeted to a particular area, the potential for severe winds to occur along the leading edge of the system was high. As such, the entire area ahead of it was warned as Doppler radar detected winds well above severe limits (public reports also contributed to the decision making in the wording within the warning text). Also, the polygon encompassed the southern half the county warning area in order to reduce any repetitiveness with the warning dissemination.

My time at the National Weather Service Wichita field office proved invaluable. Gaining hands on experience in operational forecasting, to warning simulated training, to understanding how the National Weather Service conducts their daily forecast and scientific operations is an invaluable experience. Rounding out your experiences in undergraduate work, as mentioned, is very important no matter which direction you go in any field. Learning from some of the best in the field is a humbling experience, and to be able to apply what you learn in the classroom to the real world is essential to advancing in your career path.