GRANDE Final Report: Chapter 1:
Assessing the impact of natural disasters on geoscience education and research

Introduction

This chapter provides an overview of the impacts of natural disaster events on geoscience academic departments in the United States from 2000 through 2019, and explores how natural disaster events influenced department operations, infrastructure, pedagogy, research, student engagement, and faculty composition. We address critical questions regarding the frequency and intensity of natural disasters that impacted geoscience academic departments, the threshold at which they significantly disrupted department functions, and whether departments adapted and evolved or experienced capacity reductions as a result.

We used a multi-faceted approach to understand the breadth of impacts across geoscience academic departments and the responses to these impacts. We integrated hazard event data from FEMA, NOAA, and the USGS, mapped these events to the locations of US geoscience academic departments, and assessed the frequency of hazard impacts on geoscience departments from 2000 through 2019. We also evaluated research and curricula production related to natural hazards to examine how the geoscience discipline has leveraged natural hazard events to enhance educational and research opportunities. We also assessed changes in departmental faculty composition and the changes to faculty size relative to impactful events on programs. Additionally, we evaluated trends in federal funding for hazard-related research using National Science Foundation award and solicitation data. This multi-faceted approach provides a baseline of understanding regarding how geoscience departments have leveraged natural hazard events to further educational opportunities and research activities.

Mapping geoscience academic departments to federally declared disasters (since 2000)

To establish the necessary data baselines to identify geoscience academic departments that were impacted by natural hazard events from 2000 through 2019, we integrated disparate objective sources of information (i.e., data from FEMA, NOAA and the USGS) that comprise a wide public safety warning system about natural hazard events. The data from FEMA’s Integrated Public Alert and Warning System (IPAWS), the OpenFEMA Disaster Declarations dataset, the National Weather Service’s weather warnings archive, and the USGS ShakeMap archive were integrated into a single events database. The data covered the period from January 1, 2000 to December 31, 2019. The aim of this task was to create an inventory of academic departments that had experienced natural hazards that would be invited to provide feedback through surveys about their experiences.

The density of archived data from the federal sources was uneven between 2000 and 2011 but was complete and highly dense from 2011 through 2019. Much of the unevenness was the eventual integration of disparate public safety warning systems into the IPAWS structure. This integrated database, which also included spatial extents, metrics of severity and risk, as well as temporal data, was established as the baseline data for disruptive events.

For this project, we defined the scope of qualified hazard events as: “Any event that involves a non-anthropogenic force that impedes or requires changes to regular activities, such as teaching, research, or basic operations. Also encompasses events of sufficient note or demonstrable properties that they promote specific changes to research, pedagogy, or planned operations.” Any events that were set as immediate and/or urgent notices in the baseline event data were considered as qualifying events, and for seismic events, we included any events which had a confirmed report of observable ground shaking in the same postal code as the geoscience department.

One challenge was in the outcomes of specific event classes relative to potential impacts. For instance, we had to include severe thunderstorm warnings as potentially impactful events because events such as the 2012 derecho that impacted the Midwest and Mid-Atlantic were only recognized in public safety alert systems as severe thunderstorm warnings.

All geoscience departments listed in the AGI Directory of Geoscience Departments at any point from 2000 through 2019 were geocoded. Based on this data set, each department’s event history was generated by mapping the location of the department onto the spatiotemporal data from the federal hazard data baseline. This dataset was then filtered based on indicators of severity and actions (issuing/cancelling of warnings).

Our expectation was that there would be a selection of departments that experienced potentially disruptive events; however, as determined through this process, all geoscience academic departments experienced potentially disruptive events over the project time frame. The data can be viewed in the Data section of this website.

The departments with the fewest number of alert notifications over this period were located in Guam, Idaho, and California, and are listed below. Hazard alerts for the departments in Guam were predominantly focused on severe weather, such as hurricanes and thunderstorms, as well as flooding. Hazard alerts for the department in Idaho were primarily related to severe weather, such as thunderstorms and winter storms, fires and floods. For departments in California, hazard alerts were primarily related to earthquakes, floods and winter storms.

  • College of Natural Sciences, University of Guam (291 alerts)
  • Department of Earth Sciences, Lewis-Clark State College, Idaho (322 alerts)
  • Department of Earth & Climate Sciences, San Francisco State University (385 alerts)
  • Department of Earth Sciences, City College of San Francisco (417 alerts)
  • Department of Physical & Agricultural Sciences, College of Western Idaho (493 alerts)

Departments with the most alert notifications over this period were located predominantly in Wyoming, with one department in Illinois, and are listed below. Hazard alerts for the departments in Wyoming were predominantly related to severe weather, especially severe thunderstorms, windstorms, and winter storms. Hazard alerts for the department in Illinois were overwhelmingly related to flooding events.

  • Department of Earth, Energy & Environment, Central Wyoming College(12,841 alerts)
  • Department of Geology, Augustana College, Illinois (8,904 alerts)
  • Department of Earth and Environmental Sciences, Casper College, Wyoming (8,236 alerts)
  • Department of Geology and Geophysics, University of Wyoming (6,622 alerts)
  • Western Wyoming Community College (6,140 alerts)

Given the overwhelming number of alerts for potential hazard impacts, we also analyzed the data to determine the proportion of departments located in areas covered by a disaster declaration over the 2000-2019 period. This analysis was done to assess whether there were geoscience departments that were particularly impacted over this period. The results indicated substantial impacts across departments from a variety of hazard types, especially related to severe weather, such as winter storms, thunderstorms, tornadoes, as well as flooding, hurricanes, and fires. Note that in the case of multiple hazards (aka multi-hazards in the chart below), severe weather often occurred with flooding and debris slides in the disaster declarations.

Academic departments located in disaster declaration areas (2000-2019)

Note that the hurricane category in the chart above is dominated by the wide geographical area covered by disaster declarations from Hurricane Katrina which was related to the relocation of refugees from the impact zone to communities across the country.

Academic departments located in disaster declaration areas (2000-2019) by multi-hazard type

Given the wide prevalence of impacts from natural disasters across geoscience academic departments, our final inventory of departments included all geoscience academic departments. These results contrasted with our original hypothesis that there would be a few departments that had been particularly impacted by natural hazards over this period.

AGI publishes the GeoRef bibliographic database that can be used to identify peer-reviewed literature and can be parsed at both an individual and institutional level. We analyzed the GeoRef database to identify peer reviewed publications published by US-based first authors during 2000-2019 to assess the coverage of specific topics which would be related to naturally occurring disruptive events such as hurricanes, landslides, floods, etc. to establish a baseline for evaluation of core changes in research portfolios in response to identifiable disruptive events.

Our results indicated no statistically significant signal in changes in research intensity as seen in the published literature in response to natural hazard events. This outcome suggests that there was no significant impact on research program topics and exercises that led to publications.

Traditional sources of new curricular materials and ideas, such as the Journal of Geoscience Education (JGE) and the Science Education Resource Center at Carleton College (SERC) website were examined to identify trends in production of these products relative to impactful events since 2000.

We downloaded the titles and abstracts from 1,392 articles in the Journal of Geoscience Education (JGE). We used the phi4 Large Language Model to assess if the JGE articles were related to natural hazards. Analysis of the JGE literature indicated that hazard-related articles comprised 8% of publications from 2000-2019, and 4% between 2020-2024. Half of the resources were related to multiple hazards, predominantly volcanoes and earthquakes, followed by floods and severe weather. Drought and earthquakes were the most common hazards in non-multi-hazard resources.

We also downloaded the titles and descriptions of 50,253 resources from the Science Education Resource Center at Carleton College’s (SERC) catalog (https://serc.carleton.edu/serc/search.html). We cleaned the list of downloaded resources to remove duplicate content, content with broken links, workshop and webinar surveys and evaluations, discussion posts, posts related to logistics for travel and meetings, and news items. After cleaning, our dataset used for analysis comprised 9,495 resources. We used the phi4 Large Language Model to evaluate if these resources were related to natural hazards. Only 5% (482 resources) of the cleaned set of resources were hazard related. One-third of hazard-related resources pertained to earthquakes, 16% to multiple hazards, 13% to volcanoes, and 10% to floods. Resources pertaining to multiple hazards were primarily related to earthquakes, volcanoes, and tsunamis, followed by floods, hurricanes, severe weather, and slides. When the cleaned set of data was analyzed by year, 3% of resources between 2000-2019 were hazard-related and 5% of resources between 2020-2024 were hazard related.

Identifying changes in faculty size and focus across programs impacted by disasters

Through the annual updates to the AGI Directory of Geoscience Departments, which includes a census of faculty and their specialties, we analyzed temporal changes to faculty size and focus to help identify program restructuring after events, and or development of core expertise for hazards research across academic departments. Our analysis of faculty specialties indicates there were no material changes to faculty size relative to impactful events on programs. Generally, geoscience departments are small, and do not have the capacity for the critical mass required to develop a program or portfolio in natural hazards related research.

Using the nature, location, and timing of impactful events, the NSF research funding database was analyzed for the 2000-2019 period to identify any changes in the number of awards and funding levels of research related hazards, both for specific events and hazard types in general.

We downloaded 239,260 NSF awards and 3,023 funding opportunities that were active between 2000 and 2019 to assess trends in funding for hazards research. We used the mixtral-8x-7b Large Language Model (LLM) to determine if the awards and solicitations were related to natural hazards and to identify if they were focused on specific named natural disasters. We also used the LLM to classify the awards by research focus. All AI-generated responses were audited by members of the project team. NSF award trends showed that 3.9% of all NSF awards between 2000-2019 funded natural hazards-related research while 1% of funding opportunities were specifically for hazards-related research. The Directorate for Geosciences (GEO) awarded the largest share of awards (5,239) over this period which comprised 15.2% of all GEO awards over the period. The Directorate for Engineering (ENG) awarded the next largest share of hazard-related awards (1,521) which comprised 4.6% of all ENG awards over the period.

Our analysis of total NSF funding indicated that just over $4 billion was awarded to natural hazard research, which was 3.2% of all NSF research funding over 2000-2019. Just over half of the funding ($2.3B) was from GEO and $797M was invested by ENG. For perspective, the $2.3B invested by GEO comprised 15.2% of the directorate’s 2000-2019 total funding, while the $797M invested by ENG comprised 4.6% of its total funding over the same period.

Since awards are multi-year projects, we also assessed the total number of active awards over the 2000-2019 period. The GEO directorate drove the increase in natural hazard related research over the 2000-2019 period, with 95 active awards in 2000 to 1,522 active awards in 2019. Proportionally, hazard research funded by the GEO directorate increased from 47% of all hazard-related NSF awards in 2000 to 58% in 2019. Within the GEO directorate, the percentage of awards over this period related to hazard research increased from 9% of all GEO awards in 2000 to 21% in 2019.

We analyzed the types of awards funded (i.e., research, facility and instrumentation, meeting/workshops, and travel) to better understand funding levels for research vs. other types of activities. Research awards were the most funded activity followed by funding for facilities and instrumentation. Facility and instrumentation awards comprised nearly half of ENG funding and over a third of GEO funding.

Hazard-related awards predominantly focused on multi-hazard research (21%), earthquakes (20%), volcanoes (15%), and hurricanes (10%). Hazards commonly mentioned together in multi-hazard awards included earthquakes and tsunamis, earthquakes and volcanoes, hurricanes and floods, drought and flood, earthquakes and landslides, earthquakes, tsunamis, and volcanoes, hurricanes and tornadoes, and floods and landslides. The proportion of awards granted for multi-hazard research increased from 10% of all active hazard related awards in 2000 to 28% in 2019.

Funding for natural hazard research varied by directorate in terms of types of hazards. For example, over half of the funding from the Directorate for Biological Sciences (BIO) for natural hazard research focused on drought, while half of ENG’s natural hazard funding focused on earthquakes, and nearly 2/3 of the funding from the Directorate for Education and Human Resources for natural hazard research focused on multi-hazard awards. For GEO, 42% of natural hazard funding was awarded to multi-hazard awards, 19% for earthquake awards, 11% for volcanic activity awards, and 10% for weather hazard research.

We next examined types of research being funded by NSF directorates to understand the proportion of research invested in applied and basic research as well as research focused on response, recovery, mitigation and preparedness to natural hazard events. Across directorates, applied research and research related to natural hazard preparedness were the most common types of research. Of note is that the hazard research funded by the Directorate for Education and Human Resources and the Directorate for STEM Education were overwhelmingly for hazard preparedness.

We examined the awards to see if there were notable events specified in the award title and/or abstract. Hurricanes, earthquakes, and volcanoes were the most commonly named events in the natural hazard-related awards. Top named events in the 2000-2019 awards included Hurricanes Katrina, Harvey, Irma, Maria and Sandy, as well as the Tohoku earthquake in Japan and the 2004 Indian Ocean earthquake and tsunami.

Awards with named (specific) events have changed in topical focus over time. Awards that focused on pre-1950 events most commonly mentioned historic volcanic events, while those awards mentioning events from 1950-1999 were predominantly focused on the earthquakes in Turkey, Taiwan, and Northridge, California. Awards with named events between 2000-2004 were primarily focused on multi-hazard events, such as the Indian Ocean earthquake and tsunami, and the 2002 Denali earthquake. Awards with named events in 2005-2009 primarily focused on Hurricanes Katrina and Rita, and those with named events in 2010-2014 were focused on the earthquakes in Japan, New Zealand, and Chile as well as Hurricane Sandy. Those with named events in 2015-2019 were focused mostly on Hurricanes Harvey, Irma, Maria, Florence, Michael, and Matthew, as well as the Nepal earthquake.

We next reviewed institutions of the primary investigators (PIs) on awards for natural hazard-related research to examine how funding has been awarded across the academic enterprise. We focused on the institutions of PIs that received 100 or more natural hazard-related awards between 2000 – 2019. The following 18 institutions in this cohort were awarded 29% of all natural-hazard related NSF awards during the 2000-2019 period, which totaled $1.09B.

  • Columbia University (n=243)
  • University of Colorado, Boulder (n=220)
  • University of Washington (n=212)
  • Oregon State University (n=176)
  • University of California, Berkeley (n=175)
  • University of Texas, Austin (n=169)
  • University of Hawaii (n=151)
  • Pennsylvania State University (n=139)
  • Colorado State University (n=135)
  • University of Wisconsin, Madison (n=132)
  • University of Southern California (n=131)
  • University of California, San Diego Scripps (n=126)
  • Woods Hole Oceanographic Institute (n=123)
  • University of Illinois, Urbana-Champaign (n=115)
  • Stanford University (n=109)
  • University of Oklahoma, Norman (n=109)
  • Cornell University (n=105)
  • University of Arizona (n=103)

We next examined NSF funding opportunities to understand the percentage of opportunities focused on natural hazards research. We analyzed Dear Colleague Letters, Program Summaries, and Program Solicitations from the funding section of the NSF website. Of the 3,023 opportunities listed between 2000 and 2019, only 1% (42 opportunities) were related to natural hazards. Of those opportunities, 9 were in response to specific natural hazard events such as Hurricane Katrina, the Haiti earthquake of 2010, the 2011 earthquakes in Japan and New Zealand, Hurricane Harvey, Hurricane Irma, and the 2018 Hurricane season. Most opportunities were focused on multiple hazards, followed by those focused on earthquake and volcanic activity. By directorate, most of the funding opportunities were from GEO or ENG, while 12 of the 42 opportunities were funded by multiple directorates.

The full suite of data and charts for this analysis of 2000-2019 awards and solicitations is available in the Data section of this website.

Next, we extended our analysis to the 2020-2024 awards and funding opportunities to assess if there was a continued increase in the proportion of awards for hazard-related research funding. We downloaded 52,355 NSF awards and 279 funding opportunities that were active between 2020 and 2024. We used the llama 3.1-8b Large Language Model (LLM) to determine if the awards were related to natural hazards and used the phi4 Large Language Model to analyze solicitations.

Of the awards granted over this period, only 1.6% (814 awards) were related to hazard research. One-fifth of these hazard-related awards were for multi-hazard research, 16% were for weather hazards and 16% were for volcano hazards. In addition, research related to earthquakes comprised 11% of awards, and research related to floods comprised 10% of awards. For multi-hazard awards, the hazards most mentioned together were earthquakes, volcanoes, and tsunamis followed by drought and temperature extremes.

We analyzed Dear Colleague Letters, Program Summaries, and Program Solicitations from the funding section of the NSF website for the period 2020 to 2024. Of the 279 opportunities listed 5% (14 opportunities) were related to natural hazards, all of which mentioned either multiple specific hazards, or natural hazards more broadly and were primarily focused on research related to mitigation and adaptation to climate change impacts and related hazards.

Conclusions

Natural hazards are ubiquitous across the United States and affected all geoscience academic departments over the 2000-2019 period as mapped to federal hazard event datasets. As a result of this analysis, there was no discernable subset of departments that were particularly prone to impacts from natural hazard events. These results altered our approach in the subsequent activities of the project that pertain to surveying impacted departments. To gain a deep understanding of the impacts to and recovery from these natural hazard events, our final inventory of departments included all geoscience academic departments.

Even though all geoscience academic departments experienced multiple hazard events over the 2000-2019 period, there was little response in terms of research production, faculty growth and specialization, and curricular resource production over the same period. Based on our analysis of geoscience research output related to natural hazards in scholarly literature, there was no statistically significant changes in research intensity or shifts in program focus directly attributable to natural hazard events. Furthermore, faculty size and specialization patterns also showed no evidence of significant faculty loss or program reorientation related to natural hazard impacts. Production of curriculum resources and education literature related to natural hazards showed very limited engagement since 2000. Between 2000 and 2019, hazard-related articles comprised 8% of the Journal of Geoscience Education (JGE) publications, dropping to 4% between 2020 and 2024. Hazard-related curriculum resources from the SERC catalog comprised 3% of the catalog’s resources between 2000-2019 and increased to 5% over the 2020-2024 period, a time when there was a substantial expansion of total resources added to the catalog.

Since research funding can drive engagement with specific topics, we also examined funding patterns from the National Science Foundation (NSF) since 2000. The analysis indicated a low priority for hazard-focused research. From 2000 to 2019, just 3.9% of NSF awards funded hazard-related projects, and only 1% of funding opportunities specifically targeted hazards. The Geosciences Directorate (GEO) awarded the most hazard-related grants, comprising 15.2% of its total awards during this period.

Between 2020-2024, 1.6% of NSF awards and 5% of NSF solicitations were related to hazards research. These were primarily in the context of climate change impacts and subsequent related hazards, instead of response to specific natural hazard events. In addition, there were no discernible impacts from the Catalyzing Human-Centered Solutions through Research and Innovation in Science, the Environment and Society (CRISES) and Confronting Hazards, Impacts and Risks for a Resilient Planet (CHIRP) funding mechanisms, but this may be because they were launched in 2023 and 2024, respectively.

Overall, our analysis indicates that natural hazard events did not significantly alter the capacity, focus, or trajectory of geoscience departments over the 2000-2019 period. It may be that natural hazard events may not have been severe enough and the impacts may not have been lengthy enough to cause discernable disruptions to departments’ existing operational and research patterns. Overall, there was no sustained or increased investment in hazards-related research, scholarly research, or curriculum development over this period, indicating a low priority for engagement with natural hazards across the geoscience discipline.

Table of Contents