The Effect of Office Culture and Human Bias in the
Tornado Warning Decision Process in the NWS Central Region (100)
Theodore Funk, National
Weather Service, Louisville, KY
Abstract:
In late 2016, the National
Weather Service (NWS) Central Region (CR) Tornado Warning Improvement Project
(TWIP) team sent out a comprehensive volunteer survey to all meteorologists at
CR Weather Forecast Offices (WFOs). The survey was designed to provide a
detailed understanding of the current state of the CR tornado warning decision
process, and included questions and request for feedback on office culture,
radar data usage, human factors/bias, risk communication, and training needs. A
total of 326 people responded or about 60-65% of CR meteorologists, including
247 forecasters and 79 managers.
Some findings were striking,
strongly suggesting an inconsistent warning decision process among meteorologists.
For example, varying levels of knowledge and experience make valid
tornadogenesis conceptual model training a clear need, especially for
quasi-linear convective systems (QLCSs). Effective data interrogation
strategies also were desired. Many roadblocks to a reliable decision process
center on local office culture, human collaboration, and personal bias/filters.
For example, management-forecaster interaction, inconsistency from
person-to-person within and between offices, and varying levels of communication
all contribute. Fear of missing or overwarning a tornado, and the effect of
storm history can play a significant role, differing between forecasters.
Office noise and mental/physical fatigue can have substantial and varying
impact as well. Confidence changes based on storm mode, and even storm location
and time of day/night weighs on decisions.
Understanding how the human,
not just the science, affects warning decisions and the protection of life and
property is absolutely crucial. This presentation will highlight factors which
inherently contribute to an inconsistent tornado warning decision process
across CR (and potentially other NWS regions), and discuss key TWIP
recommendations to mitigate human and cultural bias in order to facilitate a
more objective and science-based warning approach.