Between 1980 and 2017, 14 different “billion-dollar” winter storms inflicted a total of $43.9 billion in damages on the United States and caused 1,013 deaths.1 Untold further damage, disruption, injury, and death likely can be attributed to the vast majority of winter storms that did not reach the billion-dollar impact level. During the same time period, the EM-DAT database identifies 355 winter weather disasters accounting for over 19,000 fatalities globally.2 Due to population concentrations and the frequency and magnitude of winter events, winter storms particularly impact the northeastern United States (Kocin and Uccellini 2004a,b). In the Northeast, winter storms impact virtually all economic sectors, including transportation systems, by disrupting the national air space through flight cancellations and routing decisions (Kulesa 2002; Koetse and Rietveld 2009; Steiner et al. 2015), impeding ground transportation (Strong et al. 2010); adversely affecting energy generation (Hines et al. 2009) and distribution, and increasing energy demand on regional scales (Hong et al. 2013; Troccoli et al. 2014); shifting or curtailing consumer behavior in retail and recreational activities (AHUA 2014); and even impacting financial services (Brown et al. 2017).
Facing extreme weather events, including winter storms, weather information allows decision-makers to mitigate the socioeconomic impacts of hydrometeorological events. Building on existing U.S. National Weather Service (NWS) efforts to provide consistent messaging to the meteorological community and as part of the NWS Roadmap’s focus on Building a Weather-Ready Nation (WRN), the NWS has increased their efforts to provide decision support services to core partners. Two recent National Academy of Sciences, Engineering, and Medicine (NAS) reports noted the importance of a more integrated warning system that is built on social science research and ensures full communication between all actors throughout the entire emergency management process (NAS 2018a,b).3 In 2011, the NWS formalized their approach to impact-based decision support services (IDSS; sidebar “What is IDSS?” describes the NWS IDSS in more detail). Beyond traditional forecast creation and dissemination, IDSS is the provision of relevant information and interpretative services that enable partners to prepare for and respond to, as planned, extreme weather, water, and climate events for the protection of life and property (Uccellini and Ten Hoeve 2019). In part, IDSS serves to improve communication to connect information with decision-making. By enhancing IDSS, the NWS hopes to help core partners better understand, rely on, and use NWS forecasts and warnings in their decision processes to undertake better informed actions in the face of upcoming extreme and life-threatening events. The fundamental purpose of IDSS is to better connect forecasts and warnings to critical decision points of core partners “for the protection of life and property and enhancement of the national economy.”4 This study integrates economics and social science analysis to assess and improve IDSS.
We posit that formal IDSS provides core government partners with better information and support to take actions that reduce socioeconomic impacts during extreme winter storms. To illustrate the potential benefits of IDSS, we compare two storms in the New York City (NYC) area with similar characteristics but differing in their implementation of IDSS: the December 2010 storm that occurred before the implementation of formal IDSS and the January 2016 storm that occurred after the implementation of formal IDSS.5 Specific research questions for this study included 1) “What are the economic impacts from winter storms that impacted the NYC area in December 2010 and January 2016?” and 2) “Were the impacts mitigated by IDSS?” In this paper, we hope to provide the meteorological science community with a foundation for valuing benefits associated with IDSS with winter storms as an example. We further aim to show how building a WRN will facilitate connecting forecasts and warnings to decision-makers through IDSS.
Recognizing that a comparison of only two cases represents a limited sample and likely significant uncertainty in benefit estimates, we conclude with recommendations that future studies compare a larger number of storms across different areas and storm types to improve our understanding of the socioeconomic benefits of IDSS.
The socioeconomic value of IDSS
We developed a conceptual model of how IDSS works and how it can benefit decision-makers (Fig. 1). In this model, weather is shown as prestorm weather conditions that are observed and modeled, leading to forecasts and warnings of impending winter storm events. Prior to IDSS, forecasts and warnings were disseminated through conventional channels. With the advent of the NWS IDSS, additional capabilities and procedures are implemented to improve the dissemination and communication of forecasts and warnings to core partners. With IDSS, processes are put in place well before the forecast such that the forecast and warnings will connect to the decision process in a collaborative way. This ensures the information is more relevant to decision-makers to enhance the protection of property and health and safety. The benefits discussed here are focused mainly on improvements in dissemination and use of warning information more so than improvements in observations, modeling, or forecasts themselves.
Value chain of IDSS information.
Citation: Bulletin of the American Meteorological Society 101, 5; 10.1175/BAMS-D-18-0153.1
At the bottom of the Fig. 1, weather is shown again as the actualization of the event—the “storm event.” The actual weather outcome occurs within the context of the defensive investments and mitigating actions that are designed to lead to improved social and economic outcomes. Based on prior mitigating actions and procedures, and event-specific information from the NWS, the interaction of defensive investments and mitigating actions with the actualization of the storm event leads to social and economic impacts with potential asset damages, service interruptions, and human health impacts in all sectors. We focus on aviation, ground transportation, and energy.
Actual impacts in each sector are highly dependent on the structure of the sector and the decision processes. For instance, we expect a larger portion of the impacts in aviation to be in service interruptions than in human health impacts, whereas in road transportation there may be a significant component of human health impacts if extreme winter weather leads to more traffic accidents. Emergency management (EM) cuts across all sectors as a core partner and as an intermediary in communicating and implementing public responses to extreme weather events. To simplify the figure, we do not include other intermediaries, such as broadcast and private sector meteorologists; however, these intermediaries also play critical roles in forecast, warning, and decision processes.
The value of weather forecast information is zero unless it may be used to improve a societal outcome (Williamson et al. 2002). The goal of forecasts and warnings is to prompt mitigating actions that reduce the social and economic impacts of extreme weather. IDSS helps emergency managers interpret the forecasts and warnings more accurately, and understand the nature and magnitude of mitigating actions that are needed to reduce societal impacts.
We also developed an illustration of the methodological approach used to compare the two case study storms; this illustration is not drawn to scale (Fig. 2). On the horizontal axis of Fig. 2, the Northeast Snowfall Impact Scale (NESIS) indicates the storm’s potential impact based on the storm magnitude and population affected (Kocin and Uccellini 2004a). The NESIS characterizes and ranks snowstorms in the Northeast urban corridor. NESIS scores are a function of the area affected by the snowstorm, the amount of snow, and the number of people living in the path of the storm. This is thus a measure of vulnerability and not impact outcomes per se.6 The vertical axis on Fig. 2 indicates the observed impacts. These impacts can be qualitatively described, for instance, in terms of minor or major impacts, or from low to high impacts, as shown in Fig. 2. With appropriate empirical information, impacts may also be quantified in terms of lives lost, flights cancelled, utility customers without electricity, recovery time for ground transportation systems, and other impacts. Based on available data, impacts could also be quantified using monetary measures of these impacts (e.g., value of a statistical life, cost to airlines and passengers, dollar damages or willingness-to-pay to avoid power outages; Lazo 2010; Katz and Lazo 2011; Palmer and Richardson 2014; WMO et al. 2015). We drew the “damage curves” on Fig. 2 as a function of whether or not IDSS is implemented during a storm with damages presumably being less with IDSS adequately implemented. With the provision of IDSS, core partners and deep-relationship core partners (see sidebar “What is IDSS?”) have better information and support to take actions that reduce impacts.
Conceptual model on the value of IDSS.
Citation: Bulletin of the American Meteorological Society 101, 5; 10.1175/BAMS-D-18-0153.1
What is IDSS?
IDSS is the provision of relevant information and interpretative services to enable core partners’ decisions when weather, water, or climate has a direct impact on the protection of lives and livelihoods. It is a critical component of broader efforts of the NWS to build a WRN, where communities are ready, responsive, and resilient to increasing vulnerabilities from extreme weather, water, and climate events, as well as environmental hazards.
IDSS builds on the basic level of service that the NWS provides to general partners and the public, which includes:
Provision of general forecasts/warnings via standard NWS dissemination media
Response to calls from the public
Education about how to access and interpret NWS data/products
Outreach/interaction for preparedness
Interaction as part of the NWS StormReady/TsunamiReady programs and NOAA WRN Ambassador Initiative
The NWS provides IDSS to their core partners, which include government and nongovernment entities with direct responsibilities for preparing for and disseminating weather, water, and climate information that supports decision-making, to increase the usefulness and impact of weather information. Some IDSS are provided within the context of an ongoing weather event (Episodic IDSS), including event-focused webinars, instant messaging programs (NWSChat), mobile support (iNWS), and onsite or remote support. Other IDSS are provided on an ongoing basis (Recurring IDSS), including training; supporting improved interactions of Integrated Warning Teams that include staff from the NWS, emergency managers, and broadcast media; pre-event/scenario planning; water use/contingency forecasts and planning; tabletop exercises used to plan actions and procedures; after-action reviews; and daily coordination for routine high-value decisions such as aviation operations or reservoir releases.
In the provision of IDSS, the NWS utilizes everything available and can tailor what they provide the core partner based on the situation and their need. The NWS currently does not have a consistent tool set for providing event-specific tailored forecasts for IDSS, so the information generated and disseminated is left to the capabilities of staff at the offices. (R. S. Bandy, NWS Decision Support Integration Branch, 2019, personal communication).
For more information about IDSS, see the NWS Service Description Document (SDD) Impact-Based Decision Support Services for NWS Core Partners, April 2018 (www.nws.noaa.gov/im/IDSS_SDD_V1_0.pdf).
The vertical difference between the two damages curves indicates the potential societal benefits of implementing IDSS. These curves show little difference for very minimal storms, increasing to a maximum difference for major storms where improved decision-making from IDSS may maximize societal benefits and then level off at extreme storms where impacts are so severe that improved decision-making cannot prevent many types of impacts. We expect that IDSS leads to more proactive decision making by core partners to help reduce social and economic impacts for storms at any given NESIS level. The empirical question and the one we begin to examine in our case study comparison is, what is the magnitude of the difference in observed impacts in monetary or other socioeconomic measure with and without IDSS?
Methods
To begin quantifying the socioeconomic benefits of IDSS, we conducted a focused evaluation in three sectors where IDSS has been integrated to some extent over the last several years: aviation, ground transportation, and energy (Abt Associates 2017). We undertook a modified sector expert elicitation approach (Morgan 2014; Hemming et al. 2017) to understand 1) protocols for responding to extreme winter weather, 2) the types of economic and social impacts associated with extreme winter storms, and 3) the extent to which winter storms’ impacts were or could be mitigated by IDSS. First, we interviewed and gathered data from several agencies and organizations about the impacts of extreme winter storms to understand the extent to which IDSS affects social and economic impacts of extreme winter storms, including the following:
Storm characteristics and IDSS: NWS Eastern Region, Performance and Evaluation Branch, Science and Technology Services Division, Integrated Dissemination Program, Geographic Information Systems, Objective Lead for Evolve, NOAA National Climatic Data Center
EM: New York Governor’s Office of Storm Recovery, New York City Economic Development Corporation, New York City Emergency Management, Federal Emergency Management Agency
Aviation: Federal Aviation Administration, The MITRE Corporation, Port Authority of New York and New Jersey, NOAA/NWS Aviation Weather Center
Ground transportation: Federal Highway Administration, New York City Emergency Management, New York Department of Transportation, Metropolitan Transportation Authority
Energy: Office of Infrastructure Security and Energy Restoration, U.S. Department of Energy, Consolidated Edison Inc., Public Service Electric and Gas Company (PSEG)
Building on this information, we then used a case study analysis to compare two storms with similar characteristics but differing in their implementation of IDSS: one storm without formal IDSS and one storm with formal IDSS. In attributing benefits to IDSS, to the extent feasible, we attempt to account for concurrent societal and institutional changes that may have also enhanced benefits in the time period between the two events. As indicated in Fig. 1, socioeconomic impacts extend from asset damage to service interruptions to human health impacts (e.g., life and safety), and beyond (e.g., impacts on ecosystem services). In this study, we focus primarily on service interruptions for the aviation, ground transportation, and energy sectors. We assessed storm impacts in these three sectors and, where we had appropriate data, we monetized impacts
To monetize impacts in the aviation sector, we used BTS (2017) data to estimate the number of flight cancellations due to weather for each storm event. For each storm, we included 4 days of flight cancellations: 1 or 2 days of the actual event and 2–3 days of recovery. We then estimated the cost of flight cancellations to airlines and airline passengers separately. For airlines, we applied Marks (2014) average cost per cancelled flight segment for uncontrollable events, which are largely due to weather and airspace, as opposed to controllable events, which are largely due to crew and maintenance.7 Marks (2014) estimated that the cost per flight segment depending on the size of the aircraft ranges from $730 to $13,450, with an average cost per flight segment of $5,450 for uncontrollable events. We multiplied the average cost of $5,450 by the number of flight cancellations per storm event to monetize the costs of cancellations to airlines. Separately, we estimated the value of airline passenger disruptions. We determined the number of passengers disrupted by multiplying the number of cancelled flights by the average number of passengers per flight during the month of the storm (BTS 2017). We then estimated the lost time per passenger, by multiplying 1) the number of disrupted passengers, 2) the length of delay per cancellation as reported in Xiong and Hansen (2009; one cancellation = 165.9 min or 2.77 h),8 and 3) the average value of this lost passenger time at about $47.10 per person per lost hour (DOT 2016).9 The December 2010 storm occurred over the holidays with more scheduled flights than the January 2016 storm; therefore, we were unable to compare the number of flight cancellations during the two storm events. Instead, to estimate the potential benefit of IDSS in the aviation sector, we applied the 2016 storm cancellation rate (of 50.2% of flights) to the 2010 storm and to flight cancellations due to weather during the 2010/11 winter season, and the 2010 storm cancellation rate (of 40.1% of flights) to the 2016 storm and to flight cancellations due to weather during the 2015/16 winter season.
To determine the economic impacts in the energy sector, we estimated the number of customers whose service was interrupted from power outages and the power outage duration for our two storm events using data provided by the local utilities (C. Viemeister, Consolidated Edison, 2017, personal communication; T. Lupski, PSEG, 2017, personal communication). We then applied data from Sullivan et al. (2013) to monetize the cost of power outages to customers. Sullivan et al. (2013) provided estimated average electric customer interruption costs by customer type—including residential, small commercial, and medium and large commercial—and duration of power outages. We used the U.S. Energy Information Administration number of retail customers by state for the State of New York to obtain a distribution of customer types, and we assumed that the power outages affect customer types in a manner consistent with the distribution of customer types across the state. Because we received power outage duration data from just one utility, we assumed that the power outage duration was consistent across utilities, such that the average power outage duration was 12.21 h for all customers affected by the December 2010 storm and 4.81 h for all customers affected by the January 2016 storm. For the December 2010 storm, we generated costs for a 12-h power outage by multiplying the estimated average electric customer interruption cost for an 8-h outage based on Sullivan et al. (2013) by 1.5 (12 h divided by 8 h). For the January 2016 storm, we used the estimated average electrical interruption costs for a 4-h power outage. To monetize power outage interruption costs to customers, we distributed power outages across customer types, and multiplied the number of power outages by customer type by the estimated average electricity customer interruption cost.
Overview of the winter storm case studies
In this section we summarize the December 2010 and January 2016 winter storms, IDSS during each storm, and associated mitigating actions undertaken in each sector by core partners.10 Table 1 provides summary information for each storm, indicating IDSS and mitigating actions.
Overview IDSS and mitigating actions during winter storm case studies.
December 2010 winter storm.
From Saturday, 25 December, to Monday, 27 December 2010, a large blizzard impacted the Northeast United States and severely affected the NYC metropolitan area (Soltow 2010). The storm first appeared in the NWS Hazardous Weather Outlook on Tuesday, 21 December; however, the forecasts had much uncertainty and subsequent day-to-day challenges. Early forecasts predicted light to moderate snowfall, as many expected that the storm would stay offshore (Kocin et al. 2010; Weinstein and Funk 2011). European, Canadian, and U.S. forecast models did not converge on a consistent forecast scenario until 36–48 h before the onset of the heaviest snow (Kocin et al. 2010). On Friday, 24 December, models suddenly converged on solutions that indicated the event would bring heavy snow to the NYC area on Sunday, 26 December. By Saturday, 25 December, confidence grew that NYC would indeed experience a major winter storm and a blizzard warning was issued later that afternoon. The blizzard warning predicted snowfall between 11 and 16 inches (1 in. = 2.54 cm), a forecast that ultimately underestimated the actual snowfall (Weinstein and Funk 2011).
Snowfall began across the greater NYC area in the early afternoon on Sunday, 26 December, and the storm rapidly intensified. Manhattan and other parts of NYC reported heavy moisture-laden snow on Sunday and in the early hours of Monday, accompanied by high winds (Kocin et al. 2010). Snow accumulations totaled 20–30 in. across NYC and the Lower Hudson Valley, and 10–20 in. across Long Island. Snow fell at up to 1–2 in. h‒1 and winds gusted 30–45 miles per hour (mph; 1 mph = 0.45 m s‒1), occasionally exceeding 60 mph (Kocin et al. 2010). This winter storm was a category 3 event on the NESIS, with an NESIS score of 4.92.
Although IDSS was not formalized in 2010, the New York Weather Forecast Office (WFO) provided informal offsite IDSS using emails, phone conversations, and WebEx (R. Dickman, NWS, 2017, personal communication). As the storm forecasts showed a strengthening storm, the WFO hosted a live webinar for additional visibility (R. Dickman, NWS, 2017, personal communication).
Forecast uncertainty did not provide a solid basis for storm preparations until just before the onset of the storm (Kocin et al. 2010). Any uncertainty was also compounded by having to bring road crews and other city workers into work on Christmas Eve and Christmas Day, at increased expense. NYC agencies deployed more than 3,030 pieces of snow-removal equipment throughout the city, but many of these vehicles were not equipped with proper chains on their tires. In addition, the city did not ensure that sufficient private contractors were on call to assist with plowing, towing, and shoveling during the storm (Weinstein and Funk 2011). NYC decided against declaring a snow emergency and did not issue other hazardous weather advisories (Weinstein and Funk 2011), based in part on the sudden upgrade of the forecast severity after days of forecasts that had the storm just missing the NYC area. Mitigation was largely reactive and often hampered by lack of preparation.
January 2016 winter storm.
From Friday, 22 January, to Monday, 25 January 2016, a record-breaking storm dropped heavy snow from Louisiana to Maine. The storm first appeared in the NWS Hazardous Weather Outlook on Monday, 18 January, more than 4 days before the event developed (WFO OKX 2016). There was high certainty that the storm would impact the Mid-Atlantic region; however, the forecast showed a sharp northern boundary that kept NYC out of the high impact area (L. Uccellini, NWS, 2018, personal communication). On Thursday morning, 21 January, the NWS issued the first blizzard and winter storm watches. The next day, the NWS issued the first winter storm warnings (WFO OKX 2016), forecasting 8–12 in. of snow. Late that day, the forecast showed the sharp boundary located north of NYC, which put NYC in the high-impact area 1 day before event (L. Uccellini, NWS, 2018, personal communication). On Saturday, 23 January, the severity of the forecast was upgraded to Blizzard Warnings with predicted snowfall between 24 and 28 in. in the NYC area.
Heavy snow and strong winds created blizzard conditions along the New Jersey, New York, and Connecticut coastlines, and near-blizzard conditions elsewhere (WFO OKX 2016). Snow fell at a rate of 3 in. h‒1 during periods of the snow, causing whiteout conditions (DSNY 2016). The NYC area experienced snowfall accumulation of 20–30 in., with unofficial totals of 34 in. reported in Queens, wind gusts up to 50 mph across NYC (Uccellini 2016), and accumulations up to 22 in. across Long Island (Fanning 2016). Accumulations in Central Park reached 27.5 in., making it the largest snow storm recorded by the NWS at Central Park since 1869. The storm also caused widespread erosion and local coastal flooding during high tides (WFO OKX 2016). The winter storm was a category 4 event on the NESIS, with an NESIS score of 7.66, making it the fourth most impactful storm along the East Coast compared to other storms since the Blizzard of 1988 (Kocin and Uccellini 2004a).
Like the December 2010 storm, the forecast for the severity of the weather was upgraded just before the storm struck NYC (L. Uccellini, NWS, 2018, personal communication ). However, unlike the December 2010 storm, the provision of IDSS began a week before the storm’s onset, with telephone briefings to New York City Emergency Management (NYCEM) that lasted into early the following week during recovery efforts. NWS staff spent about 80 h providing telephone weather briefings, servicing media requests, and answering spotter calls or calls from storm spotter volunteers (WFO OKX 2016). The NWS provided separate email briefings to coastal partners on coastal flooding hazards and impacts for Nassau and Suffolk counties (Uccellini 2016), and to aviation partners with a “high impact” summary email on potential snowfall and disruptive surface travel for airport employees (WFO OKX 2016). NWS staff also participated in executive-level conference calls with Governor Chris Christie, Governor Andrew Cuomo, and Mayor Bill de Blasio (Uccellini 2016; WFO OKX 2016). Furthermore, embedded in the NYC Emergency Operations Center (EOC), NWS meteorologists provided onsite IDSS from Friday, 22 January, through Sunday, 24 January (WFO OKX 2016). The NWS Facebook reach was over 1.2 million, an increase of 242% from the previous week, and NWS tweets about the storm earned 3.3 million impressions (WFO OKX 2016).
NYC and the State of New York issued several advisories and actions to reduce storm impacts. Governor Cuomo and Mayor de Blasio declared a state of emergency and a winter weather emergency the morning of Saturday, 23 January. Later that afternoon, the Governor ordered a shutdown of NYC in response to the worsening winter storm. As state, city, and local government agencies took these mitigating actions, the NWS continued to provide extensive, formal IDSS activities, especially within the various EOCs where they remained onsite. This was only feasible as the NWS had built trust with NYCEM and other local partners over the preceding five years by providing and refining IDSS offerings and participation in various tabletop exercises led by the EM community.
Impacts
In this section, we compare the impacts of the December 2010 winter storm (without formal IDSS) and the January 2016 winter storm (with formal IDSS) for each sector. We examine the differences in impacts with and without IDSS as an indication of the value of IDSS, acknowledging that there are several other factors leading to differences in impacts (see “Discussions and Conclusions” section). Based on our review of the literature and event reports, expert interviews (see sidebar “Comments from core partners and sector experts on IDSS”), and economic monetization, we assess the storm impacts in the aviation, ground transportation, and energy sectors and, where reasonable, monetize impacts.
Comments from core partners and sector experts on IDSS
Aviation
According to Ralph Tamburro, the Delay Reduction Program Manager at the Port Authority and previously with the Federal Aviation Administration (FAA), the January 2016 storm resulted in much more snow, but recovery in the aviation sector was faster, suggesting that there was better planning and preparation by the airlines and a more effective response by the Port Authority to keep the runways open (R. Tamburro 2017, personal communication).
Kevin Johnston at the FAA indicated that in 2012, the FAA national operations manager (NOM) was not interested in bringing the NWS into the FAA, but the NOM now says bringing NWS meteorologists back into the Command Center was the best decision made by FAA leadership at the Command Center (K. Johnston 2017, personal communication). According to Johnston, meteorologists with knowledge about aviation operations can provide impact-based forecasts that allow the FAA to make critical decisions proactively.
Ground transportation
According to NYCEM, IDSS allows for a quicker and more complete update on the forecast, which allows NYCEM to relay the information to the appropriate agencies to make better-informed decisions in undertaking mitigating actions. For example, if forecasted snowfall totals are reduced, NYCEM can adjust staffing levels to reflect reduced risks.
According to MTA, IDSS provides weather forecasts and information easily understood by an emergency manager and relevant for decision-makers. MTA’s director of emergency management and operations support does not need to translate NWS weather forecasts and information for briefings to leadership, and the MTA actively uses IDSS to make decisions that affect 15.3 million people (A. McMahan, MTA, 2017, personal communication).
During the December 2010 storm, there were widespread flight cancellations and delays due largely to a reactive approach by many airlines. Approximately 3,760 flight cancellations due to weather cost airlines $20.5 million and passengers $40.3 million.11 During the January 2016 storm, airports also experienced extensive cancellations and delays, but for the most part these were related to a much more proactive approach. Approximately 2,420 flight cancelations due to weather cost airlines $13.2 million and passengers $30.1 million. Applying storm cancellation rates (described in “Methods” section), we find that IDSS along with proactive decision-making provides a potential value of approximately $17.5 million during these two events, and, aggregating across a typical winter, approximately up to $35.3 million annually in the NYC area. For the aviation sector, proactive flight cancellations resulted in a lower percentage of cancellations and increased passenger safety during the extreme event. Compared to reactive cancellations, proactive cancellations reduce impacts across the national airspace—a benefit we have not evaluated in the current study.
Ground transportation impacts during the 2010 storm were significantly greater than during the 2016 storm. In 2010, the city failed to procure and position sufficient private resources—including private contractors for towing, plowing, and shoveling services—before the storm (Weinstein and Funk 2011). NYC did deploy 3,030 snow-removal vehicles; however, the reactive nature of vehicle deployment was largely ineffective. During the 2010 storm, public transportation was severely hampered: Long Island Rail Road (LIRR) services were suspended during the event, Metro North Railroad services were reduced, the Metropolitan Transportation Authority (MTA) experienced bus delays, and the MTA subway experienced delays and interruptions to aboveground service. In addition, severe snow conditions blocked roadways, resulting in disabled vehicles on arterial roadways and local roadways. In contrast, snow removal and other actions were carried out more quickly for the January 2016 storm event (DSNY 2016). Large spreaders began spreading salt on roadways when the snow began; these spreaders were the first line of defense against snow and ice conditions (DSNY 2016). Although vehicle deployment was proactive, there were delays in plowing tertiary streets in Queens. In addition, bus, MTA subway, and LIRR and Metro-North railroad services were proactively suspended during the 2016 storm. The Port Authority of New York and New Jersey bridges and tunnels, and MTA bridges and tunnels were also proactively closed to nonessential vehicles. The proactive approach to the 2016 storm in the transportation sector led to significantly fewer blocked roads and delays, and reduced recovery time for ground transportation from 7 to 2 days. Figure 3 provides visual evidence of the difference in the results of mitigation efforts between 2010 and 2016.
Difference in 2010 and 2016 road conditions.
Citation: Bulletin of the American Meteorological Society 101, 5; 10.1175/BAMS-D-18-0153.1
In the energy sector, during the December 2010 storm there were approximately 76,930 customer power interruptions with an average duration of 12.2 h and an estimated interruption cost of $106.8 million. During the January 2016 storm, approximately 30,700 customers lost power for an average duration of 4.8 h and an estimated interruption cost of $14.7 million. Between the two events the energy sector experienced significantly fewer customer outages with IDSS in place with an estimated reduction in customer service interruption costs of over $92.1 million. Although we see a significant decrease in the number of power outages and outage durations between the December 2010 and January 2016 storms (from 12.2 to 4.8 h), we do not have enough information to attribute this decrease in power outage durations entirely to IDSS. Between these two storms, a regional utility made significant investments in storm hardening after Hurricane Sandy. In particular, the utility installed smart switches for overhead lines, which automatically disconnect segments of the electric grid that are experiencing problems and allow for power to flow to other areas that are not interrupted while making repairs to problem areas (Consolidated Edison 2013). These investments in storm hardening likely resulted in significant reductions in power outages from the December 2010 storm to the January 2016 storm; however, we believe that IDSS also likely contributed to the decrease in power outage durations between the two storm events.
IDSS improved decision-making and response to extreme winter weather in these three sectors, which are each considered critical infrastructure (DHS 2014). Because sectors are highly interdependent, it is likely that IDSS provided significant benefits to virtually every other sector in the region as well.
Discussion and conclusions
Through IDSS, the NWS provides forecast advice and interpretative services to help core partners make better, more informed decisions. The purpose of this study was to explore methods to qualitatively and quantitatively assess the socioeconomic benefits of IDSS and lay the groundwork for future studies to further assess IDSS. Using a case study comparison of two extreme winter storms in the NYC region, we find strong indications that IDSS has improved decision-making, enhanced communication between the NWS and core partners, and reduced societal and economic impacts of extreme winter weather events. Table 2 shows the qualitative and quantitative measures of impacts during both storms in the three sectors and the “improvement” (i.e., reduced impacts) between the storms. In summary, between these two storms, improved preparation and decision-making, and the implementation of IDSS reduced costs by over $17 million in aviation (using lower-bound estimates), reduced recovery time by 5 days for ground transportation, and reduced costs by over $90 million in the energy sector. The uncertainty ranges in these benefit estimates are likely significant. In addition, while it is difficult to characterize an uncertainty range with case studies, we feel the estimates are of the correct sign and nonzero.
Change in service interruptions between the December 2010 and January 2016 case study storm events.
It is important to note that the case study and expert assessment approach have limitations, which future research should work to address. Specifically, it is not possible to attribute all the improvements in socioeconomic outcomes to IDSS as there are differences both between the 2010 and 2016 storms and in society’s preparations, capacities, and processes to deal with such events. For example, the 2010 storm was over the Christmas holidays, which may have increased some impacts (e.g., on aviation) and reduced others (e.g., fewer people missed work). In addition, following the 2010 storm, responsible authorities undertook many new procedures to deal with winter storms, such as implementing driving restrictions and proactive cancellations of flights. To a certain extent, these improved processes are supported by improved weather forecasting, warnings, and enhanced IDSS. Furthermore, given the time period since the 2010 winter storm case, there may be memory bias on the part of interviewees that can inflate or deflate empirical estimates (Schacter et al. 2003; Morgan 2014). Although these differences make it challenging to precisely estimate the value of IDSS that influence decisions across a wide spectrum of government and commercial sectors and their related response to impending extreme weather events, these case studies play an essential role in understanding how emergency managers and other core partners use IDSS to reduce social and economic impacts from extreme events.
As the NWS and the larger Weather, Water, and Climate Enterprise continues to develop, implement, and improve IDSS and build a WRN, we provide several recommendations for how to assess the socioeconomic benefits of the more proactive approach to forecasting and warning products and services. To build a stronger business case for IDSS, we recommend more case studies to control for differences in storms, forecasts, and societal and institutional changes between storms. We would prefer a larger dataset with time series of extreme events measured on all relevant factors. Unfortunately, this dataset does not exist because, by definition, extreme events are few and far between and because of the difficulty of measuring the diverse socioeconomic factors that influence outcomes. To move in that direction, we recommend examining the implementation of IDSS for other meteorological events from hurricanes, tornadoes, floods, and droughts, as well as other phenomena. To realize the full value of conducting case studies that focus on the benefits of IDSS, we also recommend integrating case study results into the IDSS process, collecting information during each extreme event on what does and does not work with respect to the needs of core partners and impacted sectors, and better quantifying and characterizing potential changes in outcomes and enhanced socioeconomic benefits. In fact, evaluating the effectiveness of IDSS from socioeconomic perspectives is the appropriate approach to verification of IDSS efforts (i.e., user-relevant verification). Finally, we feel it would also be worthwhile to frame evaluation of IDSS in the broader context of the social sciences such as public choice theory (e.g., Mueller 2003) and communication (Fischer et al. 2016). As others note, these tasks involve complex challenges (Palmer 2017), but are necessary to solidify the connections between the weather community and those making decisions to protect life and property during the extreme weather events that impact us all.
Acknowledgments
The authors thank two anonymous reviewers and the journal editors. This work greatly benefited from information provided by many participants in the research project, including those from the NWS and NOAA, and experts from the emergency management, human health, aviation, ground transportation, and energy sectors, as noted in the full project report (Abt Associates 2017). We are grateful to Dr. Louis W. Uccellini for valuable comments and feedback. The views expressed in this article are solely those of the authors and do not necessarily represent the views of their employers. This work was funded by the National Weather Service (NWS). The NWS is a component of the National Oceanic and Atmospheric Administration (NOAA). NOAA is an Operating Unit of the U.S. Department of Commerce.
References
Abt Associates, 2017: Social and economic effects of severe winter storms: New York case study. Final Rep., 66 pp., https://vlab.ncep.noaa.gov/web/nws-social-sciences/winter-weather.
AHUA, 2014: Economic costs of disruption from a snowstorm. American Highway User Alliance, 9 pp., www.highways.org/wp-content/uploads/2014/02/economic-costs-of-snowstorms.pdf.
Brown, J. R., M. Gustafson, and I. Ivanov, 2017: Weathering cash flow shocks. SSRN, 45 pp., https://doi.org/10.2139/ssrn.2963444.
BTS, 2017: Airline on-time statistics. Bureau of Transportation Statistics, Department of Transportation, accessed 7 March 2017, www.transtats.bts.gov/DL_SelectFields.asp?Table_ID=236&DB_Short_Name=On-Time.
Consolidated Edison, 2013. Post Sandy enhancement plan. Consolidated Edison Co. of New York, Orange and Rockland Utilities, 114 pp., accessed 26 October 2017, www.coned.com/-/media/files/coned/documents/services-outages/post_sandy_enhancement_plan.pdf.
DHS, 2014: Critical infrastructure security and resilience note: Winter storms and critical infrastructure. Department of Homeland Security, 12 pp., www.npstc.org/download.jsp?tableId=37&column=217&id=3277&file=OCIA_Winter_Storms_and_Critical_Infrastructure_141215.pdf.
DOT, 2016: The value of travel time savings: Departmental guidance for conducting economic evaluations. U.S. Department of Transportation, 28 pp., www.transportation.gov/sites/dot.dev/files/docs/vot_guidance_092811c.pdf.
DSNY, 2016: 2016–2017 snow plan for the borough of the Bronx: Pursuant to Local Law 28 of 2011. The City of New York Department of Sanitation, 13 pp., www1.nyc.gov/assets/dsny/docs/2016-2017BoroughSnowPlansBronx.pdf.
Fanning, A., 2016: Historic eastern U.S. winter storm, 22–25 January, 2016. NOAA/NCEP/Weather Prediction Center, 5 pp., www.wpc.ncep.noaa.gov/winter_storm_summaries/event_reviews/2016/Historic_EasternUS_WinterStorm_Jan2016.pdf.
Fischer, D., O. Posegga, and K. Fischbach, 2016. Communication barriers in crisis management: A literature review. AIS Research Paper 168, 19 pp., https://aisel.aisnet.org/ecis2016_rp/168.
Hemming, V., M. A. Burgman, A. M. Hanea, M. F. McBride, and B. C. Wintle, 2017: A practical guide to structured expert elicitation using the IDEA protocol. Methods Ecol. Evol., 9, 169–180, https://doi.org/10.1111/2041-210X.12857.
Hines, P., J. Apt, and S. Talukdar, 2009: Large blackouts in North America: Historical trends and policy implications. Energy Policy, 37, 5249–5259, https://doi.org/10.1016/j.enpol.2009.07.049.
Hong, T., W.-K. Chang, and H.-W. Lin, 2013: A fresh look at weather impact on peak electricity demand and energy use of buildings using 30-year actual weather data. Appl. Energy, 111, 333–350, https://doi.org/10.1016/j.apenergy.2013.05.019.
Hosterman, H. R., J. K. Lazo, J. M. Sprague-Hilderbrand, and J. E. Adkins, 2019: Using the National Weather Service’s impact based decision support services to prepare for extreme winter storms. J. Emerg. Manage., 17, https://doi.org/10.5055/jem.2019.0439.
Katz, R. W., and J. K. Lazo, 2011: Economic value of weather and climate forecasts. The Oxford Handbook of Economic Forecasting, M. P. Clements and D. F. Hendry, Eds., Oxford University Press, https://doi.org/10.1093/oxfordhb/9780195398649.013.0021.
Kocin, P. J., and L. W. Uccellini, 2004a: A snowfall impact scale derived from Northeast storm snowfall distributions. Bull. Amer. Meteor. Soc., 85, 177–194, https://doi.org/10.1175/BAMS-85-2-177.
Kocin, P. J., and L. W. Uccellini, 2004b: Northeast Snowstorms. Vol. 1, Meteor. Monogr., No. 54, Amer. Meteor. Soc., 296 pp.
Kocin, P. J., L. W. Uccellini, J. Alpert, B. Ballish, D. Bright, R. Grumm, and G. Manikin, 2010: The blizzard of 25–27 December 2010: Forecast assessment. 57 pp., www.wpc.ncep.noaa.gov/winter_storm_summaries/event_reviews/2010/December25_27_2010_Blizzard.pdf.
Koetse, M. J., and P. Rietveld, 2009: The impact of climate change and weather on transport: An overview of empirical findings. Transp. Res. D, 14, 205–221, https://doi.org/10.1016/j.trd.2008.12.004.
Kulesa, G., 2002: Weather and aviation: How does weather affect the safety and operations of airports and aviation, and how does FAA work to manage weather-related effects? The Potential Impacts of Climate Change on Transportation Workshop, USDOT Center for Climate Change and Environmental Forecasting, Washington, DC, 10 pp., http://climate.dot.gov/documents/workshop1002/kulesa.pdf.
Lazo, J. K., 2010: The costs and losses of integrating social sciences and meteorology. Wea. Climate Soc., 2, 171–173, https://doi.org/10.1175/2010WCAS1086.1.
Marks, J., 2014: Updating airline cancellation costs and customer disruption. AGIFORS 54th Annual Symp ., Dubai, United Arab Emirates, Airline Group of the International Federation of Operational Research Societies, 39 pp., https://airinsight.com/wp-content/uploads/2014/10/Updating-airline-cancellation-costs-and-customer-disruption.pdf.
Morgan, M. G., 2014: Use (and abuse) of expert elicitation in support of decision making for public policy. Proc. Natl. Acad. Sci. USA, 111, 7176–7184, https://doi.org/10.1073/pnas.1319946111.
Mueller, D. C., 2003. Public Choice III. Cambridge University Press, 768 pp.
NAS, 2018a: Integrating Social and Behavioral Sciences within the Weather Enterprise. National Academies Press, 198 pp., https://doi.org/10.17226/24865.
NAS, 2018b: Emergency Alert and Warning Systems: Current Knowledge and Future Research Directions. National Academies Press, 142 pp., https://doi.org/10.17226/24935.
Palmer, T., 2017: The primacy of doubt: Evolution of numerical weather prediction from determinism to probability. J. Adv. Model. Earth Syst., 9, 730–734, https://doi.org/10.1002/2017MS000999.
Palmer, T., and D. Richardson, 2014: Decisions, decisions...! ECMWF Newsletter, No. 141, ECMWF, Reading, United Kingdom, 12–14, www.ecmwf.int/sites/default/files/elibrary/2014/14584-newsletter-no141-autumn-2014.pdf.
Schacter, D. L., J. Y. Chiao, and J. P. Mitchell, 2003: The seven sins of memory. Implications for self. Ann. N. Y. Acad. Sci., 1001, 226–239, https://doi.org/10.1196/annals.1279.012.
Soltow, M., 2010: Event review: December 25–27, 2010 winter storm, eastern United States. NOAA/NCEP, 13 pp., www.wpc.ncep.noaa.gov/winter_storm_summaries/event_reviews/2010/Eastern_US_WinterStorm_December_2010.pdf.
Sridar, B., and S. S. M. Swei, 2006: Relationship between weather, traffic and delay based on empirical methods. Sixth Conf. on AIAA Aviation Technology, Integration and Operations Conference (ATIO), Wichita, KS, AIAA, AIAA 2006-7760, https://doi.org/10.2514/6.2006-7760.
Steiner, M., A. Anderson, S. Landolt, S. Linden, and B. R. J. Schwedler, 2015: Coping with adverse winter weather: Emerging capabilities in support of airport and airline operations. J. Air Traffic Control, 57, 36–45.
Strong, C. K., Z. Ye, and X. Shi, 2010: Safety effects of winter weather: The state of knowledge and remaining challenges. Transp. Rev., 30, 677–699, https://doi.org/10.1080/01441640903414470.
Sullivan, M. J., M. G. Mercurio, J. A. Schellenberg, and J. H. Eto, 2013: How to estimate the value of service reliability improvements. Tech. Rep., 5 pp., www.ourenergypolicy.org/wp-content/uploads/2013/08/REPORT-lbnl-3529e.pdf.
Troccoli, A., L. Dubus, and S. E. Haupt, Eds., 2014: Weather Matters for Energy. Springer Science, 528 pp.
Uccellini, L. W., 2016: The historic nor’easter of January 2016: Service assessment. NOAA/NWS, 83 pp., http://www.weather.gov/media/publications/assessments/16Northeast_Blizzard.pdf.
Uccellini, L. W., and J. Ten Hoeve, 2019: Evolving the National Weather Service to build a weather-ready nation. Part I: Connecting observations, forecasts, and warnings to decision makers through impact-based decision support services. Bull. Amer. Meteor. Soc., 100, 1923–1942, https://doi.org/10.1175/BAMS-D-18-0159.1.
Weinstein, E., and S. Funk, 2011: Preliminary review of the city’s response to the December 2010 blizzard: Report and recommendations to Mayor Michael R. Bloomberg. Mayor’s Office of Operations and Mayor’s Office of Citywide Emergency Communications, 9 pp., www.nyc.gov/html/om/pdf/2011/review_of_2010_blizzard_response_01-10-11.pdf.
WFO OKX, 2016: January 22–23, 2016 blizzard after-action review: WFO OKX Event Review Outline. National Weather Service Forecast Office, https://www.weather.gov/okx/Blizzard_Jan2016.
Williamson, R. A., H. Hertzfeld, J. Cordes, and J. Logsdon, 2002: The socioeconomic benefits of earth science and applications research: Reducing the risks and costs of natural disasters in the USA. Space Policy, 18, 57–65, https://doi.org/10.1016/S0265-9646(01)00057-1.
WMO, WBG, GFDRR, and USAID, 2015: Valuing weather and climate: Economic assessment of meteorological and hydrological services. WMO-1153, 286 pp., www.gfdrr.org/valuing-weather-and-climate-economic-assessment-meteorological-and-hydrological-services.
Xiong, J., and M. Hansen, 2009: Value of flight cancellation and cancellation decision modeling: Ground delay program postoperation study. Transp. Res. Rec., 2106, 83–89, https://doi.org/10.3141/2106-101.
Based on summary data from the National Oceanic and Atmospheric Administration’s (NOAA’s) Billion-Dollar Weather and Climate Disasters database (www.ncdc.noaa.gov/billions/summary-stats).
The EM-DAT database uses a specific classification to include events as disasters (see www.emdat.be/emdat_db/). We searched EM-DAT globally for “cold waves” and “severe winter conditions.”
As noted in NAS (2018a, p. 18), “The growing emphasis on IDSS also points to the need to focus beyond just forecast and warning products towards services that support decisions for ‘end to end’ integrated planning and for building resiliency throughout the full cycle of preparedness and mitigation; monitoring, assessment, and forecasting; dissemination of warnings and recommended actions; response efforts of institutions and individuals; and post-event assessment and recovery efforts” (boldface is included in original).
NWS Mission (www.weather.gov/about/#).
We have not attempted to estimate or characterize costs of the implementation of IDSS in part as the NWS has not specifically determined the costs of policy and personnel changes related to IDSS (R. S. Bandy, NWS Decision Support Integration Branch, 2019, personal communication).
Five categories are associated with NESIS values: category 1 is notable with an NESIS value of 1.00–2.49, category 2 is significant with an NESIS value of 2.50–3.99, category 3 is major with an NESIS value of 4.00–5.99, category 4 is crippling with an NESIS value of 6.00–9.99, and category 5 is extreme with an NESIS value greater than 10.00. For more information, see Kocin and Uccellini (2004a).
Based on our review of Marks (2014), the average costs for uncontrollable events does not include costs to the airlines for passenger reaccommodation, such as meals and lodging; however, it may include some airline costs associated with rebooking passengers.
Figures in Abt Associates (2017) reflect the use of the average of values from Xiong and Hansen (2009) and Sridar and Swei (2006) where one cancellation = 600 min, for an average cancellation time of 383 min. However, because of the disparity in these estimates, the current paper uses the value from Xiong and Hansen, which may constitute a lower bound.
The estimate is also known as the Value of Travel Time Savings, which values lost travel time for intercity flights at $36.10 per person per hour for personal travel, and $63.20 per person per hour for business trips. Noting a distribution between personal and business travel that is 59.6% to 40.4%, the weighted average value equates to about $47.10 per person per lost hour (or $0.785 per minute).
For more detail on these two storms, the IDSS during the storms, and the storm impacts, see Abt Associates (2017) and Hosterman et al. (2019).
We provide all monetized value estimates in 2016 dollars. Note also that we report only lower-bound estimates here and thus estimates may be different than those shown in Abt Associates (2017).
