Stella Choi, a Ph.D. student in atmospheric sciences at the University of Illinois at Urbana-Champaign, received a Tinker Field Research Grant in 2014 to study the effects of severe weather in Argentina and to explore possible methods for improving notification and warning systems.
Record-breaking flash floods caused by heavy rainfall across the Buenos Aires province in Argentina made international headlines on 2-3 April 2013. The flooding was deemed an unprecedented catastrophe by local government officials, and resulted in overwhelming damages in the provincial capital, La Plata. Reported floodwater levels reached about 2 meters in depth, and the intense rainfall covered over 60 kilometers.
Studies have shown that subtropical South America experiences some of the most intense deep convection in the world (Zipser et al. 2006). Between 1980-2010, flooding was the most destructive natural disaster in Argentina, affecting 343 million people and costing over $222 billion USD . As such, my research focuses on improving severe weather prediction in subtropical South America using high-resolution numerical models to mitigate the societal impacts of extreme precipitation. With the Tinker Fellowship, I conducted on-site research in Buenos Aires, Argentina to listen to the personal narratives from locals who have experienced the high impact weather in the region.
I proposed to investigate the social impacts and predictability of severe weather events in Argentina to conduct preliminary research for ALERT.Ar, a project led by Dr. Paola Salio at the University of Buenos Aires that aims to improve early warning forecasts for severe weather in Argentina. Currently, forecasts in Argentina are made using large-scale features, which tend to overestimate the impacted areas, which can lead to many problems including public distrust of forecasters and decision makers. ALERT.Ar will construct a framework for training forecasters to effectively deliver hazardous weather outlooks in order to protect lives and mitigate societal and economic losses.
I drafted a survey to release to stakeholders who were involved in ALERT.Ar, including those in the agricultural sector, aviation companies, and members of the general public. I hoped to gain information on the general public’s decision making process under severe weather emergencies, which can be used to mold and refine the framework for delivering effective forecasts. I was prompted by reports I read regarding severe weather alerts in the US. Learning how the Argentine public across sectors (urban, suburban, rural) receives and interprets forecasts is a crucial area of study to ensure successful hazardous weather forecasts. My original survey addressed questions including:
- Do forecasters know their audience, and the public’s concerns?
- How can forecasts be communicated to maximize understanding, by using graphics for example?
- How does the public comprehend uncertainty (so as to minimize the chance of watch fatigue in the case of false alarms)?
- Does the public respond differently depending on who delivers the warning, or the language that is used?
A successful forecast not only involves accurate science, but also depends on an appropriate response by the public, which protects lives and minimizes societal impacts. I thought this information would be useful not only for forecasters, but also emergency managers, engineers, and urban planners who can design resilient cities to combat hydrologic disasters, and mitigate impacts in vulnerable regions lacking infrastructure and preparation, e.g. subtropical South America.
After discussing this plan with Dr. Salio and sociologists at the University of Buenos Aires, our team refined the project and began the investigation at the Argentine national weather service in Buenos Aires, the Servicio Meteorológico Nacional (SMN). Upon my arrival in Buenos Aires, I met with principal forecasters and the head of the SMN. We thought releasing our survey to all of the stakeholders would not guarantee the information that we were looking for. As such, the research project was refined and I created a survey for the forecasting department at the SMN in Buenos Aires.
After a few weeks of shadowing forecasters and monitoring the daily operations at the SMN in Buenos Aires, I became aware of the overarching issues at the organization. These issues ranged from practical/functional problems (e.g. computer malfunctions, to outdated programs) to the tedious procedures a forecaster has to complete to process a “nowcast” (a short term, very precise forecast that is valid for 3-hours). However, the main problem almost everyone told me about was the lack of employees at the organization, locally and nationally. The main SMN branch is located in downtown Buenos Aires, and this single office is responsible for forecasting for the entire country of Argentina. There are a few other locations at airports throughout the country, but these offices only focus on aviation. Not only is the lack of employees at the organization an issue, forecasters have said they have too many tasks they must complete during their shifts. In light of this new insight, my new survey had three components:
- Profile of the forecaster: to provide a background of the participant, and employees at the weather service
- The weather service: to understand the operations and issues that exist within the organization
- Perception of danger and public response: to understand the behavior of the general public during weather emergencies, and to determine if forecasters know their audience
The survey is still open, as it was released in late September, and I remain in the process of translating and analyzing the responses. However, preliminary results are showing key problems within the organization that all forecasters seem agree on.
The three overarching issues at the Buenos Aires office is the lack of employees, outdated infrastructure and services, and communication interruptions. There are only one or two forecasters during each shift, even during weather emergencies, and nearly all of the weather service employees have second jobs.
There are too many tedious tasks the forecasters must complete, which is particularly problematic when a severe weather watch is in effect and more time must be spent watching the development of weather conditions. Crafting a forecast involves both computer modeling and hand analyses (photos below). One forecaster mentions, “…it is so challenging because of the large area for which we need to forecast. We are always going to miss something.”
Unreliable computer programs have also caused complications for employees. At the turn of a shift, the next forecaster will begin his or her work by reviewing the current state of the atmosphere. If, for example, he or she notices even a single error in a temperature plotall data points must be reentered (this is done manually, point by point on the computer for about 62 locations).
Beyond the issues within the weather service organization, severe weather communication must also be improved. Before my discussions with forecasters, I hypothesized that the tone and word choice used in severe weather alerts must be simple and have a sense of urgency for the general public to understand and react accordingly. However, some of the issues at the SMN are even more fundamental—the acronyms used in severe weather alerts are written in English and receivers are not familiar with meteorological products to be able to translate the acronyms.
There is also often a lag in the distribution of severe weather alerts. Once the meteorologist completes the hand analysis on a map, the map is given to the auxiliary forecasters and assistants who type the forecasts into a computer program which updates the website. Thus there is a delay between making the forecasts and releasing alerts to the public, which is not ideal for nowcasts, as they are short-term forecasts valid for 3-hours. The nowcasters are not even aware of how long it takes for these organizations to release the alerts to the public.
Moreover, these alerts are only available in text format, and there are no graphics to serve as aids for the public to understand the weather emergency. As such, people will often call the weather service office to ask questions.
There is no official license to be considered a meteorologist in Argentina. The position requires no formal training, and most television stations have their own forecasting departments, which are completely independent of the SMN. As this is a popular medium for the general public to receive forecasts and severe weather updates, there is often blame that is unfairly directed to the weather service when the meteorologists on television programs provide incorrect forecasts or false alarms. One forecaster said, “Es un trabajo muy difícil el nuestro, muy ingrato. Cuando hacemos bien nuestro trabajo, nadie nos dice nada. Pero si por desgracia llueve de madrugada y se inunda la ciudad de Buenos Aires, te dicen de todo en todos lados...” (Translation: “It is a very difficult and thankless job we have. When we do our jobs right, no one says anything. But if there is early morning rain and the city of Buenos Aires is flooded, they tell you all about it.”)
As my survey remains open, my research is incomplete. A more thorough analysis will be presented at the Tinker Field Research workshop. I hope to continue the social impact studies coupled with my technical atmospheric science research. Speaking with locals in Argentina gave me deeper insight into the societal impacts of the extreme subtropical precipitation over the region.
Zipser, E. J., D. J. Cecil, C. Liu, S. W. Nesbitt, and D. P. Yorty, 2006: Where are the most intense thunderstorms on Earth?, Bull. Am. Meteorol. Soc., 87, 1057-1071.