a) Background

Research concerning the cost efficiency of radiosondes and alternative methods of developing a more cost-efficient form of data gathering is expanded in various scientific investigations and published works. Conclusions based on such works that are suited best for our project goals are summarized below.

Reuder et al (2009) analyzed an alternative method of gathering upper air data via the Small Unmanned Meteorological Observer also known as SUMO. Here the idea was to create a way to gather meteorological information in a cost-efficient manner in between radiosonde and tower/mast measurement stations. While SUMO is simply a Styrofoam model airplane, it has successfully reported measurements for temperature, humidity, wind speed, and wind direction up to 3500 m above the ground. In fact, through a direct comparison with a Vaisala RS92 radiosonde the SUMO proved to be an effective method of measurement. Both instruments were simultaneously released from the same environment, which in this instance was an Antarctic Iceland one. The findings showed a good agreement between the two with less than a 0.5 K variation in temperature throughout the sounding.

There were more variations between humidity as well as wind speed and direction because the SUMO ascended straight up vertical in a stationary position while the radiosonde drifted in the horizontal with height. Although the SUMO is a constantly developing method of measurement to overcome its limitations, it has proven to be an effective and efficient tool for atmospheric boundary layer analysis. Each SUMO airframe with propulsion and autopilot system, including the meteorological sensors, costs roughly 1200 Euro, which in today’s currency exchange corresponds to $1,600. The key advantages of SUMO over radiosondes are that it handles easily and maintains a cost-efficient performance.

Although there have been other methods similar to the radiosonde system such as attaching sensors to ozone profilers (Wolff 2008) or recent research on developing glidersondes that return the sensors to the launch point (Douglas 2010), a solution with support of low maintenance, minimal costs and enough provided funding has yet to be developed.

Durre et al (2005) recognizes the vital applications of the radiosonde network, which include climate monitoring, satellite data analysis, and weather prediction. This study analyzes how spatial sampling over remote areas such as oceans creates problems for all users by causing poor data. Here it is suggested that continuous radiosonde observations are not only observed on a national scale, but also on a global one. Obviously, the potential benefits need to outweigh the potential costs to make corrections to the radiosonde measurements. It has been known for years that the radiosonde network has room for improvement, especially the sensors (Jeannet 2005). But the network cannot solely rely on the United States for these detailed improvements because there needs to be specific criteria for the sensors on an international level.
On the international level there is also research related to radiosondes on improving forecasts. "THORPEX: a Global Atmospheric Research Programme" was founded in 2003 and is a ten year international research program focusing on improving the ability to forecast high-impact weather events. The program is working with academic institutions, operational forecast centers and the use of forecast products to achieve satisfying results (Rabier). This again shows that the network cannot solely rely on our country's network.

Studies such as "Multifunctional Mesoscale Observing Networks" conducted by Dabberdt el al (2005), acknowledge the limitations of radiosondes only being released twice a day across the United States. This project also concluded that quality relative humidity sensors deployed on radiosondes are essential to nowcasting in mesoscale event preparations. Consequently, if the radiosondes need to cut costs somewhere Dabberdt et al (2005) states that the NWS should not reduce the quality of the relative humidity data nor cut back on the number of radiosonde sites.

This being said, economic concerns may impact the NWS's decision and possibly lead to a cut back of radiosonde releases if there was a lack of funding. Relating the cost efficiency to the effectiveness of the current radiosonde network itself and comparing it to alternative methods requires knowledge in both fields. In terms of relating the radiosonde network to economics Morss et al (2004) states it best:

When applied to a particular arena, such as weather forecasting, economic efficiency requires that the most valuable forecasting activities are undertaken first, and that the resources expended on each activity are such that no reallocation of resources either among forecasting activities or between forecast production and activities in other sectorscould increase net societal benefit.

Morss et al (2004) that provides a step-by-step analysis of the cost efficiency of radiosondes, in the research entitled "A Systematic Economic Approach to the Evaluating Public Investment in Observations for Weather Forecasting." Based on this previous work, we intend to further the research for the current costs and compare it other methods of data gathering instruments to the process if possible. Morss et al combined the cost of the actual sounding itself along to the United States with its relationship to forecast skill. The research analyzed not only if soundings helped improve forecast skill, but also if they were worth the cost to the country and concluded that increasing the number of released soundings would be economically efficient by benefitting society.

From this, one can conclude that economic costs ultimately drive the completion of an adaptive radiosonde network strategy or improvements of the current system. It has yet to be determined if the project will further the ideas of an adaptive strategy or attempt to alter the current network. A combination of all background work in the area of interest will be taken into consideration during the upcoming research process.


We want to evaluate the current United States radiosonde strategy, including its budget to look into the cost-effectiveness of the instrument and the data it collects. When evaluating the United States we will also look into the Canadian radiosonde network problem this year to see how they are adapting as a guide. We then will research possible adaptive radiosonde observing approaches, two or three at maximum, to compare. Research of these approaches will contain data sets of run of the instrument and the expenses involved with using it. By looking at this research and the research of the current network we then will compare them for cost differences and accuracy differences. These two variables will be compared because if they instrument is not as accurate as the current strategy then it should not be an option as an adaptive strategy. In order for the adaptive strategy to be more applicable the accuracy must be the same or greater than the current strategy because accuracy will always out weigh the cost. Once these strategies are compared then conclusions will be drawn to find most accurate and cost-effective option available. If this option is different than the current radiosonde network we will then continue to develop the adaptive network in our findings.

Another of our objectives is to evaluate the special soundings that are launched during high impact weather days. We want to determine how effective these soundings are to forecasting by looking at data of when they are launched and collecting input on its aid to forecasters during this type of weather. Since a high impact weather day is a broad term we will identify its meaning before starting research. Then we will start by assessing the number of events that benefit each year from these soundings if any at all as well as if there were events that did not get a special sounding that could have benefited from one. Also we will consider if the adaptive networks researched could be valuable during these launches and if not found valuable for routine soundings if possibly only useful for special soundings. As these soundings are expensive, a goal is to determine how useful they are and if they are a viable supplements to weather forecasting. We will use rough estimates due to time constrictions to reach our project goals.


First we will conduct our research over the current radiosonde network in the United States. This will include researching related studies, articles and books in the National Weather Center library as well as online through scholarly sources for general information on the current radiosonde strategy. This general information will include, the current cost for a radiosonde instrument, employment that is created by this network and the amount of government funding and spending each year to keep the network running. Another way to collect more information, we plan to contact key staff to obtain specific information on the network's operation, performance and financing. This will lead us to contacting the people that are employed by the National Weather Service and government to release the balloons. From these individuals we will ask them about their opinion on the amount of money they make to launch the radiosondes and if qualified to make a comment we will ask them for their opinion on the cost-effectiveness of the instrument as well as its performance in relation to its accuracy. We then plan to look at the data that is received from the current network during different types of weather events for comparison later on to a possible adaptive network.

Next we will focus on the cost-effectiveness of an adaptive radiosonde observing strategy in the United States. We will also look into concentrating on only two or three alternative observing systems to compare as well. Examples of alternatives would be Amdar, profilers, GPS MET, Radiometric Profilers and more. Also, we plan to research to see if there are other options in other countries that could be applied here in the United States. Like mentioned before hopefully learning something from the Canadian problem this year. Once we identity the adaptive strategies and alternative observing systems we will start researching each one on similar topics, such as cost and accuracy. We will also collect comparable data matching the different weather events we collected for the current network. This is so once all the research for all the observing systems is collected we can compare to find which one is the most beneficial. To compare we will graph and use other presentation visuals to show the differences and similarities.

Last we will focus on the topic of special soundings. We will evaluate the use of these soundings as well as their cost-effectiveness. We will start with research on how often they are used, how much does each sounding additionally cost the National Weather Service, how many are launched on average each year and how much impact they have on forecasts. In order to accomplish this we plan on visiting the National Weather Service forecast office and Storm Prediction Center in Norman, Oklahoma and collect information on how useful the special soundings are to their forecasting and what type of weather events they use the special sounds most frequently, or for what events are they most beneficial. We would also ask them if there was an event they feel a special sounding could be helpful for forecasting that usually a special sounding does not get launched for. This information would help us understand if the money spent on special soundings is really a valuable investment. Also, to further this research on both routine and special soundings, we want to create a survey to send to all the National Weather Services across the country. This survey would ask questions to each station. It would touch as well on what extent it would touch as well on to what extent the data received by the network aids their forecasts and there would be a section asking questions about special soundings. The special sounding section would be an optional section of the survey and only to be answered if the use or launch of the sounding is applicable to the station. Lastly, we would gather information from insurance and damage surveillance companies in order to do a direct comparison of the amount of weather related events that caused damage or casualties . We would analyze this study to see if the special sounds are cost-effective towards the safety of the public. With the different ways to collect information about the special soundings we would then see if they are influential to the budget created each year for the current radiosonde network.

Stated above are the different approaches to finding the most beneficial upper air observing system will help to identity the scientific issue presented here.


This scientific issue of the need of an adaptive radiosonde network has an impact on all parts of society. Haimberger et al (2006) mentions that since the 1940s that the radiosonde network has been an important part of the global atmospheric observing system and states that they are a unique source for information about the upper air climate.

Since we have the advantage of being able to reach farther back into archived data and have higher resolution radiosonde data has a higher quality than satellite data. This higher quality of data leads to better forecasting which then leads to informing the public faster and accurate. Giving forecasters the opportunity to have this type of data, especially during more profound weather events, helps them to save more lives.

Although the government spends a large amount of money, tax payers money, on radiosondes they are known to impact long-range forecasts (5-6 days). Andersson et al (2007) showed that the combination of the use of the humidity sensor associated with the radiosonde and surface conditions help to make these long-range forecasts more accurate. Again this helps the safety of the public because it gives them more preparation time for the high impact weather that could be in the near future.

The data that needs to be received is also easily controlled so that what is needed and not needed at the time is received at a routinely manner. This idea would make an impact because it would help to control spending to a minimal amount. The current radiosonde network is not extremely flexible according to Douglas et al (2009) so this is something that could be adapted to help cost-effectiveness.

With budget cuts in our government happening frequently it would be a great disadvantage if the radiosonde network were to loose funds or to be cut out completely. By finding an adaptive network that would be more cost-effective we could lower the cost to ease the any opportunity for this to happen. This network is helpful to forecasters, which greatly effects the public. This is why our research is of great importance and will create a broader impact on the society.