1. Introduction
One of the major effects of climate change is the increased occurrence of extreme weather events (United Nations Office for Disaster Risk Reduction 2020; Pörtner et al. 2023; van der Wiel and Bintanja 2021). Climate change alters the spatial patterns and temporal variability of temperature and precipitation (Samset et al. 2019; Hansen et al. 2012) resulting in changes in global climatic patterns (Wang et al. 2021; Trenberth 2011). The subsequent shifts in the frequency, severity, and temporal variability of extreme weather events, such as heat waves, droughts, and floods are a direct expression of these changes in climatic patterns (Ebi et al. 2021; van der Wiel and Bintanja 2021). These events can have devastating effects on communities, causing destruction to property, major economic losses (United Nations Office for Disaster Risk Reduction 2020; Pörtner et al. 2023; Clarke et al. 2022), and negative impacts on public health (McMichael et al. 2003; Luber and Lemery 2015; Pörtner et al. 2023; Clarke et al. 2022). Despite global efforts for reducing current and future carbon emissions, the accumulation of greenhouse gas emissions since the 1950s has caused long-term changes to the climate (IPCC 2013). To mitigate the impacts of these changes (i.e., climatic extremes), many governmental organizations, scholars, and community leaders argue for efforts toward increasing climate resilience (Hallegatte et al. 2020; Denton et al. 2014; Gonzalez et al. 2017). Recognizing the urgency of addressing the impacts of climatic extremes, there is a growing need for innovative tools that can comprehensively assess climate resilience at various scales and facilitate effective adaptation strategies. In response to this challenge, we present the climate extremes resilience index (CERI), a new approach that combines measures of community disaster resilience and climatic extremes to identify areas at high risk and assess potential climatic impacts.
Why another index?
The development of a climate-resilience index is not a novel concept. Several indices have been previously developed to assess climate resilience (i.e., Joerin et al. 2014; Feldmeyer et al. 2020; Kusumastuti et al. 2014; Marzi et al. 2019); however, these prior indices are not designed to address climatic events across explicit or current temporal scales. Rather, they examine hazards through static measures such as the frequency of extreme events or the severity of natural events over a set period. Over the past two decades, the number of climate-related disasters has nearly doubled from 2656 events from 1980 to 1999 to 6681 events from 2000 to 2019 (United Nations Office for Disaster Risk Reduction 2020). This rapid rise in extreme weather events suggests the need for more event-driven, or temporally dynamic climate-resilience indices that can be scaled for specific events.
In recent years, event-driven indices such as the global climate risk index (Eckstein et al. 2021) and the extremes vulnerability index (EVI; Pauline et al. 2021), as well as hazard-specific indices such as the FEMA national risk index (Burns et al. 2018) and the flood resilience index (Leandro et al. 2020) have become increasingly popular. However, these indices are not without limitations. For example, the EVI enables the examination and comparison of climate extremes and vulnerability within each of the United States’s 344 climatic divisions on a monthly basis; however, this framework was not designed to examine extreme events occurring at finer temporal resolutions. Short-term events like extreme heat, cold, or precipitation may not be reflected over a monthlong time step and can be missed by coarse temporal aggregations. Thus, it may be beneficial to use daily aggregations over the longer monthly or even weekly time steps. Such challenges and biases associated with temporal selections are known as the modifiable temporal unit problem (Cheng and Adepeju 2014; Çöltekin et al. 2011), which can be addressed by a temporally agile framework such as the one proposed here.
For spatial data aggregation, many climate indices often ignore political boundaries between states and counties. Rather, climate data are often aggregated across large scales such as climatic divisions (Pauline et al. 2021), climatic regions (Gleason et al. 2008), or even continental scales (Karl et al. 1996). Although it is reasonable to aggregate climate data by climatic regions in a purely climatological setting, when applying these indices to analyze impact on populations, special considerations must be made in the societal context and the actionable levels for governmental policy and resource allocation. Resilience/vulnerability indices are often created at political boundary scales to increase policy relevance, boost stakeholder engagement, and to allow for comparative analysis between administrative zones. For example, the comprehensive disaster resilience index (CDRI) was created at the municipality scale in Italy (Marzi et al. 2019) while the baseline resilience indicators for communities (BRIC; Cutter et al. 2008) and social vulnerability index (SoVI; Cutter et al. 2003) are available at the county scale in the United States. Indices developed for use both inside and outside of the scientific community (e.g., NGOs, policy makers, and community leaders) should consider using spatial scales most beneficial for their intended users. Public authorities are typically only able to act within their legal capabilities and legislative zones. Therefore, indices developed to aid in legislation must be adapted so that authorities can design, execute, and evaluate adaptation measures using information relevant to their spatial reach (Feldmeyer et al. 2020).
Recently, some steps have been taken to address the concern of scale. The FEMA national risk index is unique in that it allows for the examination of both community resilience (at the county level) and social vulnerability (at the census tract level), as well as a suite of natural hazards (at the census tract level). Further, it estimates risk based on resilience, vulnerability, and the expected annual loss from these events. Despite its broad inclusivity of hazards and multiple measures of community health, this index was developed only for planning purposes and broad, nationwide comparisons (Zuzak et al. 2023). Additionally, the index does not allow for the analysis of specific periods in time or events; rather, it captures only the impacts of hazards as a “snapshot in time” (i.e., cumulative data available up to 2019). Although the FEMA national risk index is robust in both its analysis of hazards and attention to scale, its inability to examine the impacts of specific events limits its use to baseline measurements of risk and static applications. There has yet to be an index that 1) examines the potential impacts of hazards at a scale that is adaptive to administrative zones and 2) is temporally scalable to examine specific events.
The CERI extends upon the extreme vulnerability index (Pauline et al. 2021) by incorporating additional elements that enhance its applicability to community-based assessments. While both indices examine human resilience/vulnerability and extremes, the CERI takes a distinct approach by shifting its focus toward community and public engagement. This is achieved by incorporating a more comprehensive resilience index, implementing a county-scale aggregation for improved usability, refining the measurement of soil moisture extremes to reflect their impact on communities, developing separate indices for state and national levels of resilience, and creating a web-based tool to enhance accessibility. The CERI overcomes the weaknesses of previous climate resilience/vulnerability indices by adopting a spatially and temporally explicit framework. Using a dynamic calculation method and daily climate data, inter- and intrastate county comparisons can be made for events spanning any duration (one-day as well as multiday events). The CERI combines elements of a recently revised version of the U.S. climate extremes index (Gleason et al. 2008; Pauline et al. 2021) with the baseline resilience indicators for communities (Cutter et al. 2014) into one index. In this paper, we demonstrate the utility of CERI by applying it to a case study for 25–30 June 2021 to show its efficacy in identifying at-risk counties in the conterminous United States during the 2021 western North America heat wave.
2. Methods
a. Vulnerability-resilience framework
CERI employs portions of the disaster resilience of place (DROP) model (Cutter et al. 2008) as its framework for examining the potential impacts of extreme climatic events. Although the DROP model was created specifically to examine the effects of rapid-onset natural hazards such as earthquakes and hurricanes, it is applied here to examine the effects of climate change by focusing on events associated with climatic extremes. When extreme droughts, heat/cold waves, or extreme precipitation occur, there is an immediate effect on a community like that of a rapid-onset disaster: damage to homes and infrastructure, a sudden disruption in day-to-day activities, an increase in stress in members of a community, and the loss of life.
The DROP model defines the total hazard or disaster impact as a combination of the antecedent conditions, event characteristics, and coping responses (Cutter et al. 2008). Although the coping responses of a community can be interpreted as a function of their antecedent conditions, without directly measuring these responses, it is impossible to truly identify the impact of the hazard event. Therefore, we examine the antecedent conditions and the event characteristics prior to the addition of coping responses to identify regions most likely to suffer heavy impacts. To assess these antecedent conditions, we use the BRIC index (Cutter et al. 2010). The BRIC was developed using the DROP model as its conceptual basis and follows the DROP model’s premise that antecedent conditions include both inherent vulnerability and inherent resilience (Cutter et al. 2008). We define event characteristics as the number of extreme events occurring within each community (county). These events or “extremes” are identified by using modified version of the U.S. climate extremes index (CEI). With these modifications, we use the following framework based on the DROP model to identify areas of significant impacts (Fig. 1).
Modified DROP model for identifying short-term impacts of extreme climatic events.
Citation: Weather, Climate, and Society 16, 1; 10.1175/WCAS-D-23-0008.1
b. BRIC index
The BRIC considers six broad categories of community resilience (social, economic, community capital, institutional, infrastructural, and environmental) using 49 individual indicators at the county level (Cutter et al. 2010). Each of the 49 indicators within the BRIC are scaled and sorted into their respective category to create an overall categorical score; these categorical scores are then summed to calculate the overall resilience score for each county. This comprehensive approach to measuring community resilience enables a more thorough evaluation of risks and impacts. Climatic extremes have a wide range of effects on communities, impacting not only physical infrastructure but also social well-being, economic stability, environmental sustainability, and overall quality of life. Integrating a comprehensive community resilience index like the BRIC into the CERI recognizes the interconnectedness of community characteristics, resources, and institutions, which significantly influence the ability to cope with and recover from extreme events. The 2015 BRIC dataset, available from the Hazards and Vulnerability Research Institute (HVRI), is included in the calculation of the CERI. Note, however, that community resilience changes throughout time (Cutter and Derakhshan 2020) and that the scores for 2015 may not be truly reflective of community resilience prior to or beyond that year.
Across the conterminous United States, the highest levels of resilience are observed in the Midwest and Northeast (Cutter et al. 2014; Cutter and Derakhshan 2020). However, this does not mean that all counties in these regions are equally resilient to climate change. Resilience can be interpreted as a relative concept, that is, one place may appear much more resilient than another on a local scale, but they may be comparable to each other on a national scale. Local differences in population demographics, governmental programs, and resources can affect resilience, with some counties within each state demonstrating greater resilience than others. As a result of these local variations, county resilience values can differ markedly within regions (Cutter et al. 2014; Cutter and Derakhshan 2020). Therefore, it is important to examine resilience as a relative factor and at multiple scales to understand how the effects of extreme events can vary between communities. To achieve this goal, resilience is assessed at both the state and national levels to produce two versions of the CERI.
(top) Nationwide and (bottom) statewide BRIC scores in CONUS for 2015. Scores range from 0 to 1, with smaller values indicating higher levels of resilience and larger values indicating lower levels of resilience.
Citation: Weather, Climate, and Society 16, 1; 10.1175/WCAS-D-23-0008.1
c. U.S. CEI
The U.S CEI was first developed by Karl et al. (1996) with the goal of summarizing extremes in the climate for U.S. citizens and policy makers. The operational CEI (Gleason et al. 2008) is the annual arithmetic average of the following five indicators of the percentage of the conterminous U.S. area:
-
the sum of (i) percentage of the United States with maximum temperatures much below normal and (ii) percentage of the United States with maximum temperatures much above normal,
-
the sum of (i) percentage of the United States with minimum temperatures much below normal and (ii) percentage of the United States with minimum temperatures much above normal,
-
the sum of (i) percentage of the United States in severe drought based on the Palmer drought severity index (PDSI) and (ii) percentage of the United States with severe moisture surplus based on the PDSI,
-
2 times the value of the percentage of the United States with a much-greater-than-normal proportion of precipitation derived from extreme 1-day precipitation events, and
-
the sum of (i) percentage of the United States with a much greater-than-normal number of days with precipitation and (ii) percentage of the United States with a much greater-than-normal number of days without precipitation.
In these cases, above or below normal is defined as the upper or lower tenth percentile of the local period of record (NCEI 2022a).
The CEI is composed of five indicators that describe extremes in temperature, one-day precipitation events, days with or without precipitation, and drought on an annual or seasonal basis. However, since its original inception in 1996, the definition of extremes has been modified by others to both improve accuracy and fit research needs (Gleason et al. 2008; Pauline et al. 2021). Within the current operational CEI, extremes in temperature and soil moisture/surplus are identified as the top and bottom 10th percentile of each distribution however, this calculation is heavily based on the assumption that 20% of the country on average should be experiencing extremes (Pauline et al. 2021). Temperature simulations by Pauline et al. (2021), suggest the identification of extreme temperatures based on percentiles, rather than mean values, may not produce consistent results. Often, observations considered to be extreme by the CEI were less than two standard deviations away from the mean. Rather than using percentiles, Pauline et al. (2021) developed a robust method using Z scores for identifying extremes. This method was shown to identify extremes accurately at fine scales, as well as describe a range of conditions from normal to abnormal to extreme—an element that the original CEI lacks.
Adapting the Z-score method of extreme identification for components 1, 2, and 4 of the CEI, the CERI uses the following definitions to identify climatic extremes:
-
an average maximum temperature during the period of interest that is more than 3 standard deviations from the mean,
-
an average minimum temperature during the period of interest that is more than 3 standard deviations from the mean,
-
the presence of severe drought or moisture surplus during the period of interest as indicated by Palmer’s Z index, and
-
total precipitation occurring during the period of interest that is greater than 3 standard deviations from the mean.
Following Pauline et al. (2021), components with a Z score of 3 or greater were flagged as extreme. The total number of extreme components present is then summed and implemented in the final CERI calculation. As the fifth indicator of the original CEI (days without precipitation) cannot be modified to be determined using Z scores, we elect to omit it as an indicator. Additionally, significant modifications were made to component 4 to apply the Z-score method. Rather than focusing on extremes in 1-day precipitation for each day during the focus event, our approach involves analyzing the cumulative rainfall during the entire event period. We then compare this total rainfall with the expected or normal amount of precipitation that would typically occur during the same period using the Z-score method.
d. Revisions to CEI component 3: Drought
The CEI employs the PDSI as a measure for drought and moisture surplus conditions; however, this measure focuses on long-term drought patterns rather than short-term conditions making it difficult to identify rapidly emerging conditions (Palmer 1965; Karl 1986). The PDSI was designed to be strongly autocorrelated (Szép et al. 2005; Karl 1986) to identify long-term drought conditions, therefore the usability for time scales shorter than 12 months is severely limited. Further, the PDSI also carries a strong sensitivity to the calibration period and slow reaction time (Karl 1986). The purpose of the CERI is to examine the impact of climatic extremes on communities therefore, it is important to use a climatic gauge that can detect emerging and rapid onset events. Because of these drawbacks associated with the PDSI, the Palmer’s Z index is used in favor of the PDSI to identify drought and moisture surplus conditions.
For the purposes of developing a climate-resilience index, it is advantageous to use the Palmer’s Z index over the PDSI as a measure of soil moisture conditions for several reasons. First, is the ability for the Palmer’s Z index to identify rapidly emerging conditions (Karl 1986; Keyantash and Dracup 2002). In one case study, droughts that had begun to spread throughout the eastern United States in April and October of 1963 were not accurately reflected in the PDSI but were reflected with the Palmer’s Z index (Karl 1986). Second, is the rapid impact that short-term droughts can have on agriculture, wildfires, and human mortality. Because of the Palmer Z index’s focus on monthly variability, it is often regarded as being more useful for monitoring short-term and emerging drought/moisture surplus conditions in comparison with the PDSI, which is better suited for assessing long-term drought trends (Palmer 1986; Keyantash and Dracup 2002; Brázdil et al. 2015). When considering the impact of drought on agriculture, even a short-term deviation from normal moisture conditions can lead to negative impacts on crop yield and quality (Soulé 1992; Kaur and Behl 2010; Gilbert 2021). Considering the human health impacts, short-term droughts1 have been observed to carry a stronger association with mortality over long-term droughts (Alam et al. 2022). When coupled with high temperatures and poor air quality, short-term droughts2 have also been linked to enhanced daily respiratory mortality (Salvador et al. 2020). These observations indicate that changes in soil moisture conditions do not necessarily have to be prolonged to cause devastating impacts on communities. Due to the rapid appearance (and potential disappearance) of these events, the occurrence of these events may not be detected when using the PDSI and are, rather, only detected using a shorter assessment period like that employed by the Palmer Z index. Finally, the simpler calculation and reduced temporal scale of the Palmer’s Z index makes it relatively easier to use than the PDSI. This allows for broader applications in data scarce regions thus promoting the application of these methods beyond CONUS.
Following the Z-score method, components such as extreme temperature and extreme precipitation are well represented with Z scores; however, drought extremes were observed to have a limited representation following this method. Pauline et al. (2021) tested the Z-score method on the month of December 2015 and observed unusual conditions but no extremes in drought and moisture surplus conditions. In these identifications, unusual conditions are defined as PDSI values between 2.01 and three standard deviations and extreme conditions are defined as values greater than three standard deviations from the mean. However, following the NCEI interpretation of drought and soil moisture conditions (Table 1), both the non-Z-scored PDSI and Palmer’s Z index values classified soil moisture conditions to be extreme (wet) across the Midwest and parts of the Southeast (NCEI 2016). The definition of drought and moisture surplus varies between indices; in fact, over 50 indicators and indices are listed in the World Meteorological Organization’s “handbook of drought indicators and indices” (Svoboda and Fuchs 2016). However, it is suspected that using standard deviations of drought indices leads to an underestimation of extreme conditions. An examination of drought and soil moisture extremes reveals that many locations in the United States have never experienced extreme drought and/or moisture surplus conditions over the 1981–2022 period following the PDSI Z-score method (Figs. 3a,b). In fact, many regions commonly known for severe and lengthy droughts, such as California and New Mexico, had limited extreme droughts identified over this 41-yr period (Fig. 3a). Following the Z-score method on the Palmer’s Z index, extremes in moisture surplus are well identified throughout the United States (Fig. 3d), though the occurrences of drought extremes carry a similar underrepresentation (Fig. 3c). Although more areas of the United States are shown to have at least one occurrence of drought/moisture surplus extremes using the Z scores of the Palmer’s Z index (relative to the PDSI), the high threshold for defining extremes likely limits the identification of extreme drought conditions, especially conditions that may impact communities.
NCEI PDSI and Palmer’s Z-index interpretation.
Occurrence of extreme PDSI and Palmer’s Z-index drought and moisture surplus events over the 1981–2022 period: (a),(b) PDSI and (c),(d) Palmer’s Z (left) drought and (right) moisture surplus occurrence. Areas are marked in orange or blue where the presence of extreme conditions has occurred at least once over the 1981–2022 period. The presence of extreme drought and moisture surplus occurrence is identified by PDSI/Palmer’s Z values at least 3.01 standard deviations from the mean.
Citation: Weather, Climate, and Society 16, 1; 10.1175/WCAS-D-23-0008.1
Often, drought events considered “severe” but not “extreme” can have devastating impacts on human health. In Bangladesh, Alam et al. (2022) observed significant associations not only between human mortality and extreme droughts [standardized precipitation index/standardized precipitation–evapotranspiration index (SPI/SPEI) values ≤−2.06], but also both severe droughts (SPI/SPEI values from −2.05 through −1.28). Over the United States, even severe droughts (SPEI values < −1.28) are associated with enhanced air pollution and ground level ozone (Wang et al. 2017), which can lead to increased cases of respiratory distress. Because of the potential underestimation of extreme conditions when using standard deviations of drought indices and significance of even severe events, we opt for a more inclusive method for identifying extremes in soil moisture. While the Palmer’s Z-index Z scores can identify extremes in moisture surplus, employing different thresholds or approaches for each condition may inadvertently favor one type of extreme over the other, creating bias and neglecting certain aspects of their potential impacts on resilience. Therefore, the non-Z-scored values of the Palmer’s Z index are used in place of the Z scores to identify drought/moisture surplus extremes. For the CERI, extreme values defined by the National Centers for Environmental Information (NCEI 2022a). Following the NCEI interpretation, Palmer’s Z-index values greater than 3.5 indicate an extreme moisture surplus, while values less than −2.75 indicate an extreme drought (Table 1).
e. Revisions to scale
The original CEI examined extremes at a continental scale (Karl et al. 1996); however, since then, the application scale has been modified to fit the evolving needs of researchers and the public. Previously, Gleason et al. (2008) examined these extremes at a regional scale while Pauline et al. (2021) examined extremes at the climatic divisional scale. Though these previous works were able to successfully downscale the CEI, these scales are not useful for local policy development and for community use. To increase the usability of such climate-vulnerability indices, indices must be developed at scales geared toward their intended audiences. Because of this, the CEI is calculated on a county basis to enhance its use.
f. Revisions to CEI data sources
Both the original and revised versions of the CEI (Karl et al. 1996; Gleason et al. 2008) were calculated from data from the U.S. Historical Climatology Network (HCN), a network of weather stations spanning most of the United States. However, within the United States there is an uneven distribution in station coverage with higher concentrations of weather stations found in the eastern portion of the United States (U.S. Department of Energy 1996). For an accurate climate assessment to be conducted at the county level, it is necessary to have data representative for the area. Therefore, gridded daily temperature and precipitation data from the Parameter–Elevation Regressions on Independent Slopes Model (PRISM; PRISM Climate Group 2022) and gridded daily Palmer’s Z-index values from gridded surface meteorological dataset (gridMET) drought: CONUS indices (Abatzoglou 2012) are utilized in the calculation of the CEI. These gridded data sources provide an advantage over the previously used HCN stations, which may not always contain spatially or temporally continuous information.
To scale these datasets to the county level, the average temperature values over the period of interest are calculated for each county in the United States. These values are then compared with the 1991–2020 average for the same date range of interest. Similarly, daily precipitation is averaged over each county; however, total precipitation (sum) values are used to identify extremes over the period of interest. These precipitation values are then compared with the average total precipitation occurring over the 30-yr period for the same date range. The calculation of drought and moisture surplus using Palmer’s Z-index data from gridMET is based on climatological averages from 1979 to 2018 (Abatzoglou 2012). The average Palmer’s Z-index value is calculated for each county and these values are averaged over the period of interest. These averaged values are then interpreted using the NCEI interpretation scale shown in Table 1.
g. CERI calculation
To equalize the weighting of the antecedent conditions and hazard events in the CERI, the total number of extremes are summed and multiplied by 0.25. This produces a range of extreme values from 0 to 1, where 0 indicates no extremes, and 1 indicates all four extremes present. This “extreme” value is then summed with the resilience score (which also ranges from 0 to 1) to produce a combined climate and resilience score ranging from 0 to 2. Because of the unusual nature of a 0–2 scale, this combined score is then rescaled from 0 to 100 to produce the final CERI score. These methods are outlined in Fig. 4. Following the scoring interpretation outlined in Table 2, counties with lower CERI scores are considered less likely to be significantly impacted than counties with higher resilience scores. In the following sections, the response of the CERI is tested on the most severe days of the 2021 North American heat wave.
Calculation method for calculating the climate extremes resilience index.
Citation: Weather, Climate, and Society 16, 1; 10.1175/WCAS-D-23-0008.1
CERI score interpretation.
h. The 2021 western North American heat wave
To test the suitability of the CERI, the index is tested during the peak dates of the 2021 western North American heat wave (25–30 June 2021). The 2021 western North American heat wave was a major heat wave caused by a combination of high pressure and dry conditions (Thompson et al. 2022). Considered a once in a millennium–type event, the conditions leading to this event were made 150 times as likely to occur because of climate change (Schiermeier 2021). During the heat wave, northwestern regions of North America recorded record-breaking temperatures. In the state of Washington, a state record temperature of 120°F (48.8°C) was set in Hanford (NWS Pendleton 2021). In Oregon, temperatures peaked at 119°F (48.3°C) in Pelton Dam tying the state record (NWS Pendleton 2021). These extreme temperatures coupled with overall drought conditions from the broader 2021 western drought and heat wave led to the destruction of crops and failure of infrastructure in the United States causing an estimated $9.1 billion (U.S. dollars) in damages (NCEI 2022b). Losses from this event, however, extended far beyond monetary damages. An analysis by the New York Times determined that this event led to the excess deaths of approximately 600 people (Popovich and Choi-schagrin 2021). While the event extended across national borders affecting millions of people across the western United States and Canada, evidence has suggested a disproportionate impact across populations. In Multnomah County, Oregon, alone, 72 deaths were recorded, attributed to a lack of access to adequate air conditioning (Multnomah County 2022). Groups disproportionally affected by the event included: males (67% of victims), individuals at an advanced age, individuals living alone, individuals living in multifamily dwellings (e.g., apartments), and those experiencing homelessness or unstable housing (Multnomah County 2022). Though extreme events of this degree are generally rare, it is expected that the frequency of extreme events will increase into the future due to climate change (Pörtner et al. 2023).
3. Results
a. Recalculating the CEI components
As expected, across the Northwest, high temperatures dominated the region during the peak dates of the heat wave. Abnormally high maximum temperatures (Z scores from 1.01 through 2.00) could be seen in Montana, Idaho, Nevada, Oregon, and California. Similarly, severe high maximum temperatures (Z scores from 2.01 through 3.00) were found in the same states with the exception of Nevada. Extreme high temperatures (Z scores ≥ 3.01), however, were concentrated in Washington, northern Oregon, and northwest Idaho and Montana (Fig. 5, top-left panel). Interestingly, in the Southwest, the remnants of a prior heat wave can also be seen. Earlier in the month between 15 and 20 June, a large heat dome settled across the southwestern United States (Di Liberto 2021). Though it led to temperatures exceeding 100°F (37.7°C) over the course of multiple days (Di Liberto 2021), the dome quickly settled with some portions of the event escaping northward to join with the evolving heat wave in the Northwest.
CEI component calculations for 25–30 Jun 2021: (top left) component 1—extremes in maximum temperatures, (top right) component 2—extremes in minimum temperature, (bottom left) component 3—extremes in soil moisture conditions (identified using the Palmer’s Z index), and (bottom right) component 4—extremes in precipitation. For temperature and precipitation (the top-left, top-right, and bottom-right panels), counties in gray indicate the presence of normal conditions (Z scores from −1.00 through 1.00), counties in light blue indicate abnormal conditions (Z scores Z scores from −1.01 through −2.00 or from 1.01 through 2.00), dark blue indicate severe (Z scores from −2.01 through −3.00 or from 2.01 through 3.00), and gold indicate extreme conditions (Z scores ≤ −3.01 or ≥ 3.01). For soil moisture conditions in the bottom-left panel, severity of extremes is shown following the same ranking but using NCEI numerical scaling for severity classification.
Citation: Weather, Climate, and Society 16, 1; 10.1175/WCAS-D-23-0008.1
High minimum temperatures were also observed in the northwest region. Surprisingly, the spatial extent of abnormally high minimum temperatures extends farther than that of abnormally high maximum temperatures, stretching deeper into California, Nevada, Idaho, and Montana, and even stretching into Utah (Fig. 5, top-right panel). Similarly, the extent of severe and extreme high minimum temperatures also stretched farther out from the Northwest. In Oregon, severe and extreme high minimum temperatures dominated the state. In Idaho, every county (apart from Bear Lake County in the far southeast corner of the state) experienced at least abnormally high minimum temperatures. Washington experienced the brunt of both extremes with every county in the state encountering both extreme high minimum and maximum temperatures.
Much of the United States experienced at least some abnormal soil moisture conditions (Palmer’s Z-index value of greater than 3.50 or less than −2.75). In the northern United States, the Palmer’s Z index identified mostly drought conditions while in the south, mostly moisture surplus conditions were identified (Fig. 5, bottom-left panel). Extreme drought conditions were observed across Washington, Oregon, Montana, Idaho, Wyoming, North Dakota, South Dakota, Minnesota, and Iowa. While this range seems rather expansive, during this period, agriculture in these regions was severely affected by the drought conditions. Approximately 93% of the spring wheat–producing and 72% of barley-producing regions (the majority of these regions are located in these states) experienced droughts causing both crops to be “in the worst condition ever recorded up to this point in the season since at least 2001” (NCEI 2021). Further, droughts in corn producing areas were reported to threaten corn yields as they approached a critical stage in their growing cycle (NCEI 2021).
The Northwest was not affected by any extreme precipitation conditions during this period. However, elsewhere, extreme conditions were noted across the center of the continent stretching from New Mexico to Michigan (Fig. 5, bottom-right panel) An early monsoon season brought unusually high rainfall to New Mexico and Texas, producing over 400% of normal precipitation (PRISM Climate Group 2022). This created several flash floods in southeastern New Mexico and western Texas. Simultaneously, across the Great Plains and Midwest, severe storms brought significant precipitation producing 32 tornados ranging from EF0 to EF3 severity (NCEI 2022c). The extreme conditions created by these events can be seen in the bottom panels of Fig. 5.
b. Comparison of PDSI versus Palmer’s Z index
Relative to the PDSI, more drought and moisture surplus conditions were identified from the Palmer’s Z index (Fig. 6). In New Mexico and west Texas, a large difference between the PDSI and Palmer’s Z index is observed in extreme moisture surplus identification. During this period, monsoonal rains across New Mexico led to widespread flash flooding and flooded roadways in the southeast portion of the state (Davies 2021). Despite these conditions, the PDSI denoted the condition of the soil as midrange/moderate. In the Southeast, heavy rains from Tropical Storm Claudette earlier in the month (19–20 June 2021) caused severe flooding across northern Alabama and Mississippi (NWS Birmingham 2021) and southeast Louisiana (NWS Lake Charles 2021). Although the PDSI was able to identify some counties experiencing this extreme moisture surplus, the number of counties affected by Claudette was better reflected with the Palmer’s Z index (Fig. 6).
Differences in extreme conditions identified by the PDSI and the Palmer’s Z comparison for 25–30 Jun 2021.
Citation: Weather, Climate, and Society 16, 1; 10.1175/WCAS-D-23-0008.1
In the north, drought conditions that had begun to spread across Montana (NWS Great Falls Montana 2021), Minnesota (Minnesota Department of Natural Resources 2022), and South Dakota were only flagged as moderate by the PDSI but were classified as extreme by the Palmer’s Z index. Throughout 2021, much of California and Nevada experienced significant drought conditions (NCEI 2022d) and the PDSI captured these static conditions. However, the rapid onset of extreme soil moisture conditions in the month of June were only detected with the Palmer’s Z index.
c. Final CEI calculation
To calculate the final CEI, the total number of extreme components are summed for each county; these totals are displayed in the top panel of Fig. 7. The final CEI map reveals that during the study period, much of the United States experienced at least some extreme conditions in temperature, drought, or precipitation (Fig. 7, top panel). As expected, counties with the greatest number of extremes components were concentrated in the Northwest where extreme temperature conditions dominated the region. However, across the central United States and parts of the South, the CEI also identified extreme precipitation and moisture surplus conditions and in the north emerging extreme drought conditions were identified.
Total extremes identified during 25–30 Jun 2021 at (top) the county scale and (bottom) the climate-division scale. Soil moisture extremes in both panels are identified using the Palmer’s Z index.
Citation: Weather, Climate, and Society 16, 1; 10.1175/WCAS-D-23-0008.1
When comparing the county scale used within this study and the climate divisional scale used by Pauline et al. (2021), strong similarities are observed between the two maps (Fig. 7). However, in many cases, finer-scale conditions found within individual counties are not represented in the climate divisional scale, an example of the modifiable areal unit problem (MAUP; Openshaw 1981). This is an important consideration when examining large climate divisions that may encompass many counties. For example, the larger climate divisions found across the western United States make it difficult to identify extreme events occurring at county scales. Rather, these occurrences are “averaged out” and lost within the larger climate divisions. While this is not to suggest that using climate divisional aggregations is inherently wrong, for community applications, county-level examinations of climate extremes are likely more appropriate for capturing localized impacts and addressing specific vulnerabilities and needs.
d. CERI results
During the 2021 western North American heat wave, the most at-risk counties in the Pacific Northwest3 were found in eastern Washington (WA), northern Oregon (OR), northern Idaho (ID), and western Montana (MT), with extremes in minimum temperature, maximum temperature, and soil moisture occurring in each state (Fig. 8a). On a state level, assessment of Washington, Adams County was observed to be the most at-risk with a CERI state score of 86.68, while in Oregon, Morrow County was found to be the most at-risk with a CERI state score of 87.50 (Fig. 8b). In Idaho, Bonner County was the most at-risk county with a CERI state score of 68.79 and in Montana, Lincoln County was most at risk (CERI state score: 84.65). Each of these counties had some of the lowest resilience scores of their respective states (Table 3), which, when compounded with a large number of extremes (3 in each county) led to larger scores, relative to their state. These larger scores indicate the presence of a heighted risk of impact from extremes during the event.
(top) CERI state-level and (bottom) national-level results. (right) Plots that highlight the data (left) over the northwestern United States. The CERI was calculated over 25–30 Jun 2021, during the peak dates of the 2021 western North American heat wave. Resilience and climatic extreme data used to calculate the index were measured at the county level. The most at-risk counties of each of the four most affected states (Washington, Oregon, Idaho, and Montana; top-right panel) from this event were Adams County, Washington (B1); Morrow County, Oregon (B2); Bonner County, Idaho (B3); and Lincoln County, Montana (B4). On a national comparative scale (bottom-right panel), Bonner County, Idaho (D1); Adams County, Washington (D2); Ferry County, Washington (D3); Morrow County, Oregon (D4); and Benewah County, Idaho (D5), were identified as the most at risk.
Citation: Weather, Climate, and Society 16, 1; 10.1175/WCAS-D-23-0008.1
Most at-risk counties in each of the Pacific Northwest states affected by the 2021 western North American heat wave. Note that larger CERI state scores and resilience scores indicate a higher level of risk and lower level of resilience, respectively.
On a national scale, the top five most at-risk counties during this event were Bonner County, ID; Adams County, WA; Ferry County, WA; Morrow County, OR; and Benewah County, ID (Fig. 8d). Relative to the United States, states in the Pacific Northwest are slightly above average in terms of community resilience (national average resilience score of 0.47; Fig. 8c). However, all five of these counties had three extremes identified (in minimum and maximum temperature and soil moisture) and had low baseline levels of community resilience prior to the event (Table 4). Although it can be argued that Oregon and Washington were the most devastated states from this event, Bonner County, Idaho, was identified as the most at-risk county nationally (CERI national score: 72.47). Despite this ranking, when its resilience is measured relative to its state (rather than to the nation) and compared nationally, Bonner County dropped from being the most at-risk county to being the 12th most at-risk county (CERI state score: 68.79). These differences in ranking demonstrate the importance of understanding the relative nature of resilience when drawing comparisons. For example, counties with a low baseline level of resilience relative to their state may have a high level of resilience on a larger (national) scale. Similarly, counties with a high baseline level of resilience relative to their state may have a low level of resilience on the national scale. To account for the relative nature of resilience, it is recommended that the CERI national be used for interstate comparisons and CERI state be used for intrastate comparisons and within local governments and organizations.
The top five most at-risk counties during the 2021 western North American heat wave. Note that larger CERI national scores and resilience scores indicate a higher level of risk and lower level of resilience, respectively.
4. Discussion and conclusions
The creation of the CERI marks a significant step forward in assessing the impacts of climatic extremes on community resilience at local scales. For the 2021 western North American heat wave, CERI demonstrated its utility in identifying the most at-risk counties in the United States, which were concentrated in eastern Washington, northern Oregon, and Idaho. Although the heat wave served as the primary event for testing the CERI, note that the index is designed to capture a wide range of climatic extremes beyond heat waves. These include extremes in temperature, precipitation, and soil moisture, all of which were detected in various regions across the United States during the 25–30 June 2021 period. Despite Oregon and Washington suffering the most damages from the extreme heat event in terms of casualties and physical damages, Bonner County, Idaho, was identified as the most at-risk county during this event due to its lower baseline level of resilience and high number of extremes (three extremes present). Similarly, the most at-risk counties for each state produced high CERI scores due to the combination of their lower baseline resilience scores and high number of extremes. While these county scores do not necessarily indicate the highest number of casualties or the most significant damages from the event, they do suggest that the impact of the extremes within these counties during the event may exceed that of neighboring counties within the same state. As a result, these counties may require additional assistance and support to fully recover to their pre-event conditions.
Previously, Pauline et al. (2021) was successful in enhancing the identification of temperature and precipitation extremes by using Z scores in their identification. However, we observe that these Z scores likely produce an imbalance in identifying soil moisture extremes (Fig. 3). In fact, soil moisture extremes identified by the Z score method were much rarer than extremes in temperature and precipitation. Following a normal distribution, approximately 99.7% of data falls within three standard deviations of the mean, yet according to the interpretations established by Palmer (1965), approximately 8% of PDSI/Palmer-Z values are regarded as extreme (Table 1). Though this is a much larger proportion of events than the 0.3% allotted by the Z-score method, the likelihood of encountering a drought or wet condition extreme enough to be flagged by the Z-score method is very low (Fig. 3). It is speculated the underperformance of the Z-score method in identifying soil moisture extremes is due to the high variability of soil moisture conditions. Unlike temperature patterns, soil moisture conditions do not follow a consistent seasonal pattern and can stretch to extreme ranges at various times within each season. As a result, the standard deviations of soil moisture for each season are relatively large, leading to a wide range of potential values and limiting the ability to identify extremes accurately. Consequently, only the most extreme events would be identified as such. Therefore, the underperformance of using drought Z scores may be tied to this reduced, or lack of, seasonality. Previous works examining the impacts of droughts found that even severe events, not necessarily extreme, can have devastating impacts on human health and mortality (Alam et al. 2022; Wang et al. 2017). While the climatic portion of the CERI focuses on examining extremes, the definition of an extreme should be relevant to assessing the impacts to communities. Therefore, the definition of soil moisture extremes is broadened to include more events. By doing so, we introduce more balance between the climatic variables and open the definition of extremes to be more inclusive to events that may impact communities.
While the PDSI remains one of the most widely used soil moisture indices worldwide, the Palmer’s Z index possesses several notable advantages. Previous studies have demonstrated that the Palmer’s Z index offers a more robust approach to capturing short-term and emerging deviations from typical moisture conditions (Karl 1986; Brázdil et al. 2015). The Palmer Z index’s shorter “memory” of past conditions enables timely identification of emerging drought or moisture surplus conditions that may otherwise go unnoticed when using longer temporal assessment. Additionally, its ability to detect short-term conditions allows it to capture conditions that would have been otherwise lost at longer temporal assessment scales. Incorporating the Palmer’s Z index allows the CERI to be more adept at identifying and assessing diverse conditions that may affect community resilience, thereby enhancing its effectiveness as a comprehensive tool for climate risk analysis.
Although many indices have been developed to examine the impacts of climatic hazards and extremes (FEMA 2021; Eckstein et al. 2021; Feldmeyer et al. 2020; Inostroza et al. 2016; Joerin et al. 2014; Kusumastuti et al. 2014; Leandro et al. 2020; Marzi et al. 2019; Pauline et al. 2021; Summers et al. 2020), accessibility to these indices is often limited due to their heavy data and software requirements. By revising data sources to open access, callable sources found within the Google Earth Engine framework, the CERI can be used by the public to actively engage in understanding the risks and vulnerabilities their communities face, enabling them to take proactive measures, advocate for necessary changes, and allocate resources efficiently. This empowerment of individuals and communities strengthens their capacity to build resilience, enhance preparedness, and contribute to effective climate adaptation and mitigation efforts at the local level. The CERI offers two distinct indices: the CERI state, suitable for local level examinations and statewide assessments, and the CERI national, designed for county-scale comparisons across multiple states or regions and is made operational by a Google Earth Engine app (https://ceri.cartoscience.com/).
Despite these advancements made toward climate-impacts research, several limitations exist for the tool. Climate data sources are limited to the conterminous United States therefore, this tool cannot be used to assess climatic impacts on Alaska, Hawaii, or Puerto Rico. Additionally, due to the temporal availability of gridMET and PRISM data, there exists a temporal limitation for when and how long CERI can be calculated using the web app. The most recent temperature and precipitation data available from the PRISM Climate Group and gridMET is available five and eight days before the current date, respectively. Therefore, there is an eight-day delay between the current day and the earliest the CERI can be calculated. Further, due to delays in data updates, the web tool is currently only functional for periods at least five days in length. Although the web app has its limitations, the greatest challenge lies in the difficulty of verifying the accuracy and reliability of climate-resilience indices. Like many indices that have come before, the CERI is at best, only an estimate of how climatic extremes may interact with community resilience. Validating climate-resilience indices is a difficult task, however, their validation is essential to build confidence among community planners and other stakeholders in its utility. Validation testing of this index has not yet been systematically conducted and future work should explore the validation of CERI to firmly establish a reasonable and effective tool for assessing climate resilience.
While the CERI is designed to examine short-term impacts, it is important to note that climatic events can trigger a cascade of secondary and indirect impacts that are difficult to measure. Assessing these long-term effects and their ramifications on communities requires a comprehensive understanding of complex systems and interdependencies. Community systems are dynamic, complex, and regionally unique, making them extremely difficult to completely capture with simple indices like the one proposed in this paper. Special care must also be taken when using these indices to avoid committing ecological fallacy. Climate resilience indices typically operate at aggregated levels, such as regional or community scales, and provide an overview of the resilience of a particular area. However, individual communities or households within that area may exhibit different characteristics, vulnerabilities, or capacities that are not captured by the aggregated index. Making assumptions or generalizations about individual-level resilience based solely on the index’s findings can be misleading and may overlook important variations and nuances within the population. Therefore, it is crucial to interpret and use climate resilience indices cautiously, recognizing their limitations and considering additional contextual information to avoid the ecological fallacy and ensure more accurate assessments of resilience at the individual and community levels. While the use of finer spatial aggregations (county scale) and multiscale (CERI state vs CERI national) spatial examinations aims to limit some of this risk, for best practices, it is important to understand the limitations associated with climate-resilience indices.
For the CERI framework, there is much room for improvement. As suggested by Pauline et al. (2021), we too would also like to include the fifth indicator of the CEI (greater than normal days with/without precipitation) in the final index calculation. As temperature, precipitation, and soil moisture data are often the most readily accessible forms of weather-related data, CERI is limited to assessing extremes related to these variables. To enhance CERI’s comprehensiveness, it would be beneficial to integrate additional datasets, such as wind patterns, atmospheric pressure, and air quality. This expansion would allow CERI to cover a broader spectrum of climatic extremes, including conditions associated with hurricanes, tornadoes, and wildfires. Additionally, because the BRIC includes individual broad measures of community resilience (social, community capacity, infrastructure, etc.), it would be worth investigating the interactions of these measures with climate extremes individually. Individual assessments can provide deeper insights into how different aspects of community resilience contribute to mitigating or exacerbating the impacts of climate extremes. For example, understanding how social cohesion and community networks influence community responses to extreme weather events can inform targeted interventions and support systems. Exploring the role of robust infrastructure in reducing vulnerability to climate-related hazards can guide investments in resilient infrastructure projects. Similarly, special considerations should be made when applying CERI and related climate resilience indices to indigenous and tribal nations. These nations may have their own governance structures, legal systems, and decision-making processes that both reflect their sovereignty and self-determination and impact their recovery from disasters. By including these specialized adaptations and assessments, we can enhance our understanding of the nuanced dynamics between community resilience and climate extremes, leading to more effective policies, strategies, and interventions for building climate-resilient communities.
The CERI is a fully operational index that combines measures of community disaster resilience and climatic extremes to identify areas of risk and to assess potential climatic impacts. While this work does heavily extend on elements of the EVI produced by Pauline et al. (2021), we believe that our contributions make this index better suited for community-based applications over the EVI, which carries strong climate-based characteristics (e.g., climate divisional-scale aggregation of data, very high thresholds established for soil moisture, use of long-term monthly data over daily for detection of extremes). To cater more effectively to community and public engagement, we have made several key improvements in the development of the CERI, including 1) incorporating a more comprehensive resilience index (BRIC) to encompass the broad impacts of climatic extremes; 2) implementing a county-scale aggregation to improve usability for target audiences; 3) enhancing the measurement of soil moisture extremes to better reflect their potential impact on communities, such as agriculture and human health; 4) developing separate indices specifically tailored to state and national levels of resilience, accounting for variations in scale and context; and 5) creating a new web-based tool to help promote public and community accessibility. Despite the number of published climate-risk indices, limited accessibility of these tools reduces their overall impact. By creating a platform where this index can be easily accessed and shared, we demonstrate that indices can be more widely utilized and effectively integrated into decision-making processes. To increase the impact of new tools and research, we suggest that the development of future tools and indices be conducted with accessibility and public engagement in mind.
SPI/SPEI at a 3-month accumulation period.
SPI/SPEI calculated at a 30-day accumulation period.
Washington, Oregon, Idaho, and Montana.
Acknowledgments.
The authors thank Dr. Kevin Ash at the University of Florida for his assistance in developing the weighting system for CERI. They also appreciate the valuable feedback provided by the editor and anonymous reviewers, which significantly improved the quality of this paper.
Data availability statement.
This work uses public-domain temperature and precipitation data that are available from the PRISM Climate Group (https://prism.oregonstate.edu/) and the Earth Engine Catalog (https://developers.google.com/earth-engine/datasets). Drought data (gridMET) are available from the Climatology Lab at the University of California Merced (https://www.climatologylab.org/gridmet.html) and the Earth Engine Catalog. Resilience data (BRIC) are available from the Hazards and Vulnerability Institute at the University of South Carolina (https://sc.edu/study/colleges_schools/artsandsciences/centers_and_institutes/hvri/data_and_resources/bric/index.php). These data are available as a Microsoft Excel file and were joined to 2016 TIGER county and state shapefiles (https://www.census.gov/cgi-bin/geo/shapefiles/index.php) as well as the Earth Engine Catalog. Note that the BRIC scaling used in this work differs from that developed by Cutter et al. (2014).
REFERENCES
Abatzoglou, J. T., 2012: Development of gridded surface meteorological data for ecological applications and modelling. Int. J. Climatol., 33, 121–131, https://doi.org/10.1002/joc.3413.
Alam, I., S. Otani, A. Nagata, M. S. Khan, T. Masumoto, H. Amano, and Y. Kurozawa, 2022: Short- and long-term effects of drought on selected causes of mortality in northern Bangladesh. Int. J. Environ. Res. Public Health, 19, 3425, https://doi.org/10.3390/ijerph19063425.
Brázdil, R., M. Trnka, J. Mikšovský, L. Řezníčková, and P. Dobrovolný, 2015: Spring-summer droughts in the Czech Land in 1805–2012 and their forcings. Int. J. Climatol., 35, 1405–1421, https://doi.org/10.1002/joc.4065.
Burns, J., C. Zuzak, and J. Rozelle, 2018: Introducing FEMA’s National Risk Index: A baseline multi-hazard risk assessment for the United States and an interactive web platform to visualize it. 2018 Fall Meeting, Washington, DC, Amer. Geophys. Union, Abstract NH52B-04.
Cheng, T., and M. Adepeju, 2014: Modifiable temporal unit problem (MTUP) and its effect on space-time cluster detection. PLOS ONE, 9, e100465, https://doi.org/10.1371/journal.pone.0100465.
Clarke, B., F. Otto, R. Stuart-Smith, and L. Harrington, 2022: Extreme weather impacts of climate change: An attribution perspective. Environ. Res. Climate, 1, 012001, https://doi.org/10.1088/2752-5295/ac6e7d.
Çöltekin, A., S. De Sabbata, C. Willi, I. Vontobel, S. Pfister, M. Kuhn, and M. Lacayo, 2011: Modifiable temporal unit problem. ICC2011 Workshop, Paris, France, ICA GeoVisualization, https://doi.org/10.5167/uzh-54263.
Cutter, S. L., and S. Derakhshan, 2020: Temporal and spatial change in disaster resilience in US counties, 2010–2015. Environ. Hazards, 19, 10–29, https://doi.org/10.1080/17477891.2018.1511405.
Cutter, S. L., B. J. Boruff, and W. L. Shirley, 2003: Social vulnerability to environmental hazards. Soc. Sci. Quart., 84, 242–261, https://doi.org/10.1111/1540-6237.8402002.
Cutter, S. L., L. Barnes, M. Berry, C. Burton, E. Evans, E. Tate, and J. Webb, 2008: A place-based model for understanding community resilience to natural disasters. Global Environ. Change, 18, 598–606, https://doi.org/10.1016/j.gloenvcha.2008.07.013.
Cutter, S. L., C. G. Burton, and C. T. Emrich, 2010: Disaster resilience indicators for benchmarking baseline conditions. J. Homel. Secur. Emerg. Manage., 7, 1, https://doi.org/10.2202/1547-7355.1732.
Cutter, S. L., K. D. Ash, and C. T. Emrich, 2014: The geographies of community disaster resilience. Global Environ. Change, 29, 65–77, https://doi.org/10.1016/j.gloenvcha.2014.08.005.
Davies, R., 2021: USA—Evacuations after flash floods in New MexiCounty FloodList, accessed 14 September 2022, https://floodlist.com/america/usa/floods-new-mexico-june-2021.
Denton, F., and Coauthors, 2014: Climate-resilient pathways: Adaptation, mitigation, and sustainable development. Climate Change 2014: Impacts, Adaptation, and Vulnerability. Part A: Global and Sectoral Aspects, C. B. Field et al., Eds., Cambridge University Press, 1101–1131, https://www.ipcc.ch/site/assets/uploads/2018/02/WGIIAR5-Chap20_FINAL.pdf.
Di Liberto, T., 2021: Record-breaking June 2021 heatwave impacts the U.S. west. NOAA Climate, accessed 29 June 2022, https://www.climate.gov/news-features/event-tracker/record-breaking-june-2021-heatwave-impacts-us-west.
Ebi, K. L., and Coauthors, 2021: Extreme weather and climate change: Population health and health system implications. Annu. Rev. Public Health, 42, 293–315, https://doi.org/10.1146/annurev-publhealth-012420-105026.
Eckstein, D., V. Künzel, and L. Schäfer, 2021: Global climate risk index 2021. Who suffers most from extreme weather events, 2000–2019. Germanwatch Briefing Paper, 52 pp., https://www.developmentaid.org/api/frontend/cms/file/2021/03/Global-Climate-Risk-Index-2021_1.pdf.
Feldmeyer, D., D. Wilden, A. Jamshed, and J. Birkmann, 2020: Regional climate resilience index: A novel multimethod comparative approach for indicator development, empirical validation and implementation. Ecol. Indic., 119, 106861, https://doi.org/10.1016/j.ecolind.2020.106861.
FEMA, 2021: The National Risk Index. FEMA, https://hazards.fema.gov/nri/.
Gilbert, P.-L., 2021: Palmer’s Z-index as an underlying index for a weather-based crop insurance policy: Saskatchewan study case. M.S. thesis, Dept. d’Économie Appliquée, École des Hautes Études Commerciales de Montréal, 94 pp., https://biblos.hec.ca/biblio/memoires/gilbert_pierreluc_m2021.pdf.
Gleason, K. L., J. H. Lawrimore, D. H. Levinson, T. R. Karl, and D. J. Karoly, 2008: A revised U.S. Climate Extremes Index. J. Climate, 21, 2124–2137, https://doi.org/10.1175/2007JCLI1883.1.
Gonzalez, R., and Coauthors, 2017: Community-driven climate resilience planning: A framework, version 2.0. National Association of Climate Resilience Planners, 64 pp., https://movementstrategy.org/resources/community-driven-climate-resilience-planning-a-framework/.
Hallegatte, S., J. Rentschler, and J. Rozenberg, 2020: Adaptation Principles: A Guide for Designing Strategies for Climate Change Adaptation and Resilience. World Bank, 200 pp., http://hdl.handle.net/10986/34780.
Hansen, J., M. Sato, and R. Ruedy, 2012: Perception of climate change. Proc. Natl. Acad. Sci. USA, 109, E2415–E2423, https://doi.org/10.1073/pnas.1205276109.
Inostroza, L., M. Palme, and F. de la Barrera, 2016: A heat vulnerability index: Spatial patterns of exposure, sensitivity and adaptive capacity for Santiago de Chile. PLOS ONE, 11, e0162464, https://doi.org/10.1371/journal.pone.0162464.
IPCC, 2013: Summary for policymakers. Climate Change 2013: The Physical Science Basis, T. F. Stocker et al., Eds., Cambridge University Press, 1–29.
Isard, S. A., and W. E. Easterling, 1989: Predicting large-area corn yield with a weighted Palmer Z-index. J. Climate, 2, 248–252, https://doi.org/10.1175/1520-0442(1989)002<0248:PLACYW>2.0.CO;2.
Joerin, J., R. Shaw, Y. Takeuchi, and R. Krishnamurthy, 2014: The adoption of a climate disaster resilience index in Chennai, India. Disasters, 38, 540–561, https://doi.org/10.1111/disa.12058.
Karl, T. R., 1986: The sensitivity of the Palmer drought severity index and Palmer’s Z-index to their calibration coefficients including potential evapotranspiration. J. Climate Appl. Meteor., 25, 77–86, https://doi.org/10.1175/1520-0450(1986)025<0077:TSOTPD>2.0.CO;2.
Karl, T. R., R. W. Knight, D. R. Easterling, and R. G. Quayle, 1996: Indices of climate change for the United States. Bull. Amer. Meteor. Soc., 77, 279–292, https://doi.org/10.1175/1520-0477(1996)077<0279:IOCCFT>2.0.CO;2.
Kaur, V., and R. Behl, 2010: Grain yield in wheat as affected by short periods of high temperature, drought and their interaction during pre- and post-anthesis stages. Cereal Res. Commun., 38, 514–520, https://doi.org/10.1556/CRC.38.2010.4.8.
Keyantash, J., and J. A. Dracup, 2002: The quantification of drought: An evaluation of drought indices. Bull. Amer. Meteor. Soc., 83, 1167–1180, https://doi.org/10.1175/1520-0477-83.8.1167.
Knapp, P. A., 1998: Spatio-temporal patterns of large grassland fires in the intermountain west, U.S.A. Global Ecol. Biogeogr. Lett., 7, 259–272, https://doi.org/10.2307/2997600.
Kusumastuti, R. D., V. Viverita, Z. A. Husodo, L. Suardi, and D. N. Danarsari, 2014: Developing a resilience index towards natural disasters in Indonesia. Int. J. Disaster Risk Reduct., 10, 327–340, https://doi.org/10.1016/j.ijdrr.2014.10.007.
Leandro, J., K.-F. Chen, R. R. Wood, and R. Ludwig, 2020: A scalable flood-resilience-index for measuring climate change adaptation: Munich city. Water Res., 173, 115502, https://doi.org/10.1016/j.watres.2020.115502.
Luber, G., and J. Lemery, 2015: Global Climate Change and Human Health: From Science to Practice. Jossey-Bass, 638 pp.
Marzi, S., J. Mysiak, A. H. Essenfelder, M. Amadio, S. Giove, and A. Fekete, 2019: Constructing a comprehensive disaster resilience index: The case of Italy. PLOS ONE, 14, e0221585, https://doi.org/10.1371/journal.pone.0221585.
McMichael, A. J., D. H. Campbell-Lendrum, C. F. Corvalán, K. L. Ebi, A. K. Githeko, J. D. Scheraga, and A. Woodward, 2003: Climate change and human health: Risks and responses. World Health Organization, 333 pp., https://iris.who.int/bitstream/handle/10665/42742/924156248X_eng.pdf.
Minnesota Department of Natural Resources, 2022: The drought of 2021. Accessed 14 September 2022, https://www.dnr.state.mn.us/climate/journal/drought-2021.html.
Multnomah County, 2022: Final report: Health impacts from excessive heat events in Multnomah County, Oregon, 2021. Final Heat Rep. 2021, 26 pp., https://multco-web7-psh-files-usw2.s3-us-west-2.amazonaws.com/s3fs-public/20220624_final-heat-report-2021_SmallFile-2.pdf.
NCEI, 2016: December 2015 drought report. NOAA, accessed 13 September 2022, https://www.ncei.noaa.gov/access/monitoring/monthly-report/drought/201512.
NCEI, 2021: June 2021 drought report. NOAA, accessed 18 May 2023, https://www.ncei.noaa.gov/access/monitoring/monthly-report/drought/202106.
NCEI, 2022a: U.S. Climate Extremes Index definition. NOAA, accessed 17 January 2023, https://www.ncei.noaa.gov/access/monitoring/cei/definition.
NCEI, 2022b: U.S. billion-dollar weather and climate disasters, 1980–present. NOAA, accessed 15 December 2022, https://doi.org/10.25921/stkw-7w73.
NCEI, 2022c: Storm Events Database. NOAA, accessed 28 June 2022, https://www.ncdc.noaa.gov/stormevents/.
NCEI, 2022d: Annual 2021 drought report. NOAA, accessed 4 October 2022, https://www.ncei.noaa.gov/access/monitoring/monthly-report/drought/202113.
NWS Birmingham, 2021: Central Alabama impacts from Claudette. NOAA/National Weather Service, accessed 21 August 2022, https://www.weather.gov/bmx/event_06192021.
NWS Great Falls Montana, 2021: Montana drought/precipitation summary. NOAA/National Weather Service, 7 pp., https://www.weather.gov/media/tfx/wxsummary/montanawx_2021_07.pdf.
NWS Lake Charles, 2021: Tropical Storm Claudette 2021. NOAA/National Weather Service, accessed 14 September 2022, https://www.weather.gov/lch/2021Claudette-PSHLIX.
NWS Pendleton, 2021: Climatological data for Pendleton area, OR—June 2021. NOAA/National Weather Service, accessed 14 September 2022, https://www.weather.gov/wrh/climate?wfo=pdt.
Openshaw, S., 1981: The modifiable areal unit problem. Quantitative Geography: A British View, Routledge and Kegan Paul, 60–69.
Palmer, T. N., 1986: Influence of the Atlantic, Pacific and Indian Oceans on Sahel rainfall. Nature, 322, 251–253, https://doi.org/10.1038/322251a0.
Palmer, W. C., 1965: Meteorological drought. U.S. Weather Bureau Research Paper 45, 58 pp., https://www.droughtmanagement.info/literature/USWB_Meteorological_Drought_1965.pdf.
Pauline, E. L., J. A. Knox, L. Seymour, and A. J. Grundstein, 2021: Revising NCEI’s climate extremes index and the CDC’s social vulnerability index to analyze climate extremes vulnerability across the United States. Bull. Amer. Meteor. Soc., 102, E84–E98, https://doi.org/10.1175/BAMS-D-19-0358.1.
Popovich, N., and W. Choi-schagrin, 2021: Hidden toll of the northwest heat wave: Hundreds of extra deaths. New York Times, 11 August, https://www.nytimes.com/interactive/2021/08/11/climate/deaths-pacific-northwest-heat-wave.html.
Pörtner, H. O., and Coauthors, 2023: Technical summary. Climate Change 2022: Impacts, Adaptation and Vulnerability. Cambridge University Press, 37–118, https://doi.org/10.1017/9781009325844.002.
PRISM Climate Group, 2022: Daily temperature and precipitation data. Accessed 20 October 2022, https://prism.oregonstate.edu.
Quiring, S. M., and T. N. Papakryiakou, 2003: An evaluation of agricultural drought indices for the Canadian prairies. Agric. For. Meteor., 118, 49–62, https://doi.org/10.1016/S0168-1923(03)00072-8.
Salvador, C., R. Nieto, C. Linares, J. Díaz, and L. Gimeno, 2020: Short-term effects of drought on daily mortality in Spain from 2000 to 2009. Environ. Res., 183, 109200, https://doi.org/10.1016/j.envres.2020.109200.
Samset, B. H., C. W. Stjern, M. T. Lund, C. W. Mohr, M. Sand, and A. S. Daloz, 2019: How daily temperature and precipitation distributions evolve with global surface temperature. Earth’s Future, 7, 1323–1336, https://doi.org/10.1029/2019EF001160.
Schiermeier, Q., 2021: Climate change made North America’s deadly heatwave 150 times more likely. Nature News, 8 July, https://doi.org/10.1038/d41586-021-01869-0.
Soulé, P. T., 1992: Spatial patterns of drought frequency and duration in the contiguous USA based on multiple drought event definitions. Int. J. Climatol., 12, 11–24, https://doi.org/10.1002/joc.3370120103.
Summers, K., and Coauthors, 2020: Development of a Cumulative Resilience Screening Index (CRSI) for natural hazards: An assessment of resilience to acute meteorological events and selected natural hazards. U.S. Environmental Protection Agency, EPA/600/R-20/274, 2020, 287 pp.
Svoboda, M., and B. A. Fuchs, 2016: Handbook of drought indicators and indices. Integrated Drought Management Tools and Guidelines Series 2, WMO/TD-1173, 52 pp., https://www.droughtmanagement.info/literature/GWP_Handbook_of_Drought_Indicators_and_Indices_2016.pdf.
Szép, I. J., J. Mika, and Z. Dunkel, 2005: Palmer drought severity index as soil moisture indicator: Physical interpretation, statistical behaviour and relation to global climate. Phys. Chem. Earth, 30A/B/C, 231–243, https://doi.org/10.1016/j.pce.2004.08.039.
Thompson, V., and Coauthors, 2022: The 2021 western North America heat wave among the most extreme events ever recorded globally. Sci. Adv., 8, eabm6860, https://doi.org/10.1126/sciadv.abm6860.
Tian, L., S. Yuan, and S. M. Quiring, 2018: Evaluation of six indices for monitoring agricultural drought in the south-central United States. Agric. For. Meteor., 249, 107–119, https://doi.org/10.1016/j.agrformet.2017.11.024.
Trenberth, K. E., 2011: Changes in precipitation with climate change. Climate Res., 47, 123–138, https://doi.org/10.3354/cr00953.
United Nations Office for Disaster Risk Reduction, 2020: The human cost of disasters: An overview of the last 20 years 2000–2019. Accessed 2 October 2022, https://www.undrr.org/publication/human-cost-disasters-overview-last-20-years-2000-2019.
U.S. Department of Energy, 1996: United States Historical Climatology Network (U.S. HCN) monthly temperature and precipitation data (revision 3). U.S. Historical Climatology Network, accessed 16 September 2022, https://cdiac.ess-dive.lbl.gov/ndps/ushcn/ushcn.
van der Wiel, K., and R. Bintanja, 2021: Contribution of climatic changes in mean and variability to monthly temperature and precipitation extremes. Commun. Earth Environ., 2, 1, https://doi.org/10.1038/s43247-020-00077-4.
Vasiliades, L., and A. Loukas, 2009: Hydrological response to meteorological drought using the Palmer drought indices in Thessaly, Greece. Desalination, 237, 3–21, https://doi.org/10.1016/j.desal.2007.12.019.
Wang, J., Y. Guan, L. Wu, X. Guan, W. Cai, J. Huang, W. Dong, and B. Zhang, 2021: Changing lengths of the four seasons by global warming. Geophys. Res. Lett., 48, e2020GL091753, https://doi.org/10.1029/2020GL091753.
Wang, Y., Y. Xie, W. Dong, Y. Ming, J. Wang, and L. Shen, 2017: Adverse effects of increasing drought on air quality via natural processes. Atmos. Chem. Phys., 17, 12 827–12 843, https://doi.org/10.5194/acp-17-12827-2017.
Zuzak, C., E. Goodenough, C. Stanton, M. Mowrer, N. Ranalli, D. Kealey, and J. Rozelle, 2023: National Risk Index. FEMA Tech. Doc., 415 pp., https://www.fema.gov/sites/default/files/documents/fema_national-risk-index_technical-documentation.pdf.
