Download PDF Spring Bridge Issue on Engineering and Climate Change March 15, 2020 Volume 50 Issue 1 The seven articles in this issue cannot cover all engineering-related aspects of climate change, but they highlight several areas of concern. Predictability of Hydrometeorological Extremes and Climate Impacts on Water Resources in Semiarid Zones: Expectations and Reality Monday, March 16, 2020 Author: Soroosh Sorooshian, Vesta Afzali Gorooh, Negin Hayatbini, Mohammed Ombadi, Mojtaba Sadeghi, Phu Nguyen, and Kuolin Hsu The following two questions are often asked by practicing professionals and the public: How will climate change affect water availability and precipitation variability and change at regional scales? Can changes in precipitation trends and variability be predicted? In this article we consider these questions in the context of observations and modeling to predict water availability and precipitation variability. Globally, the amount of fresh water on average remains constant, but variability and changing trends at the continental, regional, and local scales are critical factors for planning and management. Water is needed for domestic consumption, agriculture, industry, and ecosystem services, and as the world’s population increases, so does the demand for freshwater supplies. Global population more than doubled from 3.7 billion in 1970 to over 7.5 billion today and it is projected to reach 10 billion by 2050.[1] In the Western United States, which is largely semiarid to arid, the population during 1970–2018 more than doubled, from nearly 34 million to 76 million.[2] And more densely populated and expanding urban developments spreading to flood plains and areas near inland waters such as lakes and rivers increase vulnerability to flooding associated with hydrologic extremes. Focusing on the precipitation component of the hydrologic cycle, we organize the rest of this article along two lines: First, what do historical recorded observations of precipitation reveal? Second, what do predictive models indicate about future trends and patterns of precipitation? What Do Historical Precipitation Observations Reveal? Historical observations can be categorized as instrumental observations or proxy and reconstruction records. Instrumental Observations: Gauges, Radars, and Satellites Instrumental observations come from three sources: rain gauges, radars, and, more recently, satellites. Rain gauges are the source of the longest precipitation records—dating back to the late 1800s—and have served as the backbone for most of the information needs of operational and water resource engineering communities as well as hydrologic services around the world. The Global Precipitation Climatology Centre in Germany (operated by the country’s national meteorological service, Deutscher Wetterdienst, under the auspices of the World Meteorological Organization), collects and archives rain gauge information provided by member nations, reported as monthly aggregates from some 6,000 gauges.[3] Its collection of global rainfall information started in 1891.[4] US meteorological observations using rain gauges date from the 1880s. Cleveland Abbe (1888) described the standards for weather (rain) gauges to be used by the US Army Signal Corps, and the 8² diameter gauge is still in use by many offices of the National Weather Service (NWS) and cooperative weather observers both nationally and internationally.[5] For the United States, an excellent source of information about extreme rainfall events and statistical precipitation frequency for any location/region of the country is the online Hydrometeorological Design Studies Center (HDSC) of the National Oceanic and Atmospheric Administration (NOAA).[6] Weather radars are an important source of information about precipitation measurement, but they have limitations in mountainous and remote regions. Remotely sensed precipitation will likely become the dominant source of information in the future, although the value of ground-based rainfall measurements from gauges will never diminish. We focus here on precipitation estimated from satellite observations that yield near-global estimates for continents and oceans, followed by specific illustrations of extreme precipitation and flooding. Remotely Sensed Precipitation Observations Although shorter in length than gauge observations, satellite data make it possible to observe and analyze precipitation patterns at high resolution over oceans and continents down to country and regional scales. NOAA’s Climate Data Record (CDR) program has released a relatively high resolution (daily, 0.25°) precipitation database, PERSIANN (Precipitation Estimation from Remotely Sensed Information Using Artificial Neural Networks)–CDR, covering 60°N–60°S from 1983 to almost the present (the data are updated quarterly).[7] The data make it possible to examine historical patterns and trends at the watershed scale. The same data are also provided through a CHRS website (the algorithm and methods used to produce the data are described in Ashouri et al. 2015).[8] Figure 1 A recent examination of global precipitation volume and variability over the oceans and continents for a 33-year period (1983–2015; figure 1A-B) shows no statistically significant trends (according to the Mann-Kendall test; Nguyen et al. 2018). But the analyses show a different picture when the data are examined at the country level: The warmer colors (reds) in figure 1C indicate a downward trend in precipitation while the cooler colors (blues) indicate an upward trend (solid colors indicate significant changes; changes in hashed, lighter-colored areas are not statistically significant). In addition to NOAA’s CDR program, the Global Precipitation Measurement (GPM) mission is an international satellite program providing observations of rain and snow worldwide every 3 hours. A joint mission of NASA and the Japan Aerospace Exploration -Agency (JAXA), the GPM Core Observatory satellite[9] was launched in February 2014. The Geostationary Operational Environmental Satellite Program (GOES),[10] a joint effort of NASA and NOAA, provides critical high resolution (in both time and space) data for precipitation estimation. Other countries have similar meteorological satellite programs and international collaborations that make it possible to map meteorological events and estimate precipitation at the global scale. In all of these efforts, advances in machine learning are playing a crucial role in processing vast amounts of satellite data to improve the capability of artificial intelligence algorithms for precipitation retrieval. Evidence of Extreme Precipitation and Floods NOAA’s National Centers for Environmental Information (NCEI; formerly the National Climatic Data Center) are the steward and archive center for most precipitation and other meteorological records. Analysis of US rain gauge information over the 1901–2015 period reported in the fourth National Climate Assessment (NCA4) report (Easterling et al. 2017) pinpoints two key findings: Annual US precipitation averaged across the country has increased approximately 4 percent, albeit with regional and seasonal variations. The intensity of extreme precipitation as indicated by several metrics (e.g., 5-year maximum daily precipitation, 99th percentile daily precipitation) has increased. The second finding appears to support conclusions about global warming and is illustrated by the following examples of extreme precipitation events across the globe: Typhoon Hagibis in October 2019 resulted in a record-breaking amount of rainfall over Japan—more than 3.3 feet in 24 hours. The storm’s severe impacts on infrastructure and houses led to the evacuation of 8 million people and about $10 billion in insured losses (Freedman 2019). Cyclones Idai and Kenneth in March–April 2019 brought unprecedented rainfall and flooding in Mozambique, Zimbabwe, and Malawi, killing hundreds of people.[11] Also in March–April 2019, extreme precipitation in Eastern Iran after a multiyear drought caused extensive loss of life and property (Asanjan et al. 2019). In 2014 the United Kingdom experienced its wettest winter in 250 years (Vaughan 2014). In the Eastern United States and Atlantic coastal region, Hurricanes Florence and Michael in September and October 2018 resulted in rainfall of 20 to 30 inches, which produced catastrophic flooding (NCEI). Hurricane Harvey in 2017 was another record--breaking event, in both peak intensity and geographical extent its maximum precipitation exceeded 60 inches in 24 hours near Houston.[12] In January–February 2017, heavy precipitation delivered by a number of back-to-back atmospheric -rivers (narrow corridors of concentrated moisture in the atmosphere) caused major flooding in Northern -California, which experienced its wettest winter in almost a century (NCEI).[13] In May 2015 Texas and Oklahoma experienced unprecedented amounts of rainfall that resulted in major flooding (NCEI). These record-breaking events highlight the need for strategies to mitigate and adapt to such extremes.[14] Proxy and Reconstruction Records for Droughts At the other extreme, droughts have brought devastation and hardship to many regions of the world. While drought prediction remains a challenge, historical observations provide insight into their frequency and persistence. With respect to the United States, observations since 1895 have shown pronounced multi-year to multidecadal variability, but no evidence of long-term trends toward more or fewer droughts (figure 2A).[15] Figure 2 What about drought evidence over much longer historical periods (i.e., thousands of years)? One way to address this question is to use reconstruction time series data from either isotopic studies of dried lake deposits or tree ring proxies of precipitation or river flow. Figure 2B illustrates nearly 2,000 years (1–1980 AD) of hydroclimate history (Laird et al. 1996; Woodhouse and Overpeck 1998) over the US Great Plains based on analysis of North Dakota’s Moon Lake salinity record. The figure shows deviations from mean log -salinity -values; negative values indicate low salinity and therefore wet conditions, positive values indicate high -salinity and dry conditions. This reconstructed proxy history shows a number of megadroughts—-lasting 100 years or more—before 1200 AD, and a shift since then to relatively wetter conditions. The Dust Bowl of the mid-1930s pales in comparison to the earlier periods. Analysis of reconstructed tree ring time series in the Western United States shows similar multidecadal patterns of drought in the same time frame (Cook et al. 2004). What Do Predictive Models Reveal about Future Trends and Patterns of Precipitation? We now provide an overview of modeling used for generating forecasts and projections of precipitation, to support prudent water resource planning and decision making. Predictions are categorized as short, medium, and long range (figure 3). Figure 3 Short-Range Forecasts These are intended for flashflood and general flood forecasting and require models and observations within hours or days. They depend on close cooperation between the hydrologic modeling and weather forecast communities. Advanced modeling and geographic information system tools allow for the development of models at very fine resolution (meters). However, for short-range forecasts they face important challenges in meeting parameteriza-tion and calibration requirements, and, more importantly, limited availability of high-resolution quantitative precipitation estimates from in situ observations and quantitative precipitation forecasts from numerical weather prediction models (current research and development efforts may improve the reliability of the latter). Such inputs are needed for the National Water Model, a hydrologic modeling framework under development that involves multiple governmental agencies and academic researchers and is intended to “provide high-resolution forecasts of soil moisture, surface runoff, snow water equivalent, and other parameters.”[16] Mid-Range Forecasts Also known as seasonal forecasts, these cover periods from weeks to 3 months. A number of climate modeling and numerical weather prediction centers globally provide seasonal forecasts for operational purposes such as reservoir management. Official US seasonal forecasts are provided by NOAA’s NWS Climate Prediction Center.[17] Figure 4 Other centers, such as the International Research Institute for Climate and Society (IRI), provide probabilistic seasonal precipitation forecasts based on NOAA’s North American Multi-Model Ensemble project. Figure 4 illustrates an example of the IRI’s probabilistic seasonal forecast generated in December 2018 for January–March 2019. It shows two areas (circled in red) that experienced extreme precipitation not forecasted by the probabilistic multimodel system. This points to some limitations of the models and should provide an appreciation for the probabilistic nature of these predictions. Seasonal forecast products are under continuous development and their consideration for any application and decision making should be approached with caution and appreciation of their probabilistic nature. Long-Range Forecasts The third category of required hydrometeorological predictions covers periods from years to decades. Such information is critical for many applications, especially water resource system planning, infrastructure design, and operations. Over the past two decades much emphasis in the literature has been on results from the application of state-of-the-art climate models to examine future spatial and temporal precipitation patterns and trends and to evaluate potential climate change impacts on water resources in various regions of the world. The Intergovernmental Panel on Climate Change (IPCC) Fourth Assessment Report (AR4; Easterling et al. 2017) indicates trends of decreasing precipitation across most of the subtropics and increasing precipitation in tropical regions. Examination of changes in precipitation trends in the Southwestern United States has resulted in numerous publications generally agreeing with the IPCC findings, expecting a drier region in this century and raising concerns about the future of southwestern water supplies (e.g., Cayan et al. 2010; Seager et al. 2007). The article by Seager and colleagues (2007) captured much attention although the authors acknowledged (in their abstract) “if these models are correct…,” recognizing the limitations of the models’ abilities and accuracy. The studies were all based on low-spatial-resolution general circulation model simulations, which do not capture the topography’s influence on precipitation in the mountainous regions of the western states. To investigate further the level of accuracy of the climate model projections, we examined outputs from the North American Regional Climate Change Assessment Program (NARCCAP; Mearns et al. 2012), an international effort to produce simulations for (i) investigating uncertainties in regional-scale projections (for 2040–70) of climate and (ii) generating future climate scenarios for use in impact research. NARCCAP ran regional climate models (RCMs) at a spatial resolution of 50 km, driven by atmosphere-ocean general circulation models (AOGCMs) covering the contiguous United States and most of Canada. The AOGCMs were forced with the IPCC Special Report on Emissions Scenarios (SRES) A2, describing a very heterogeneous world for the 21st century (Nakicenovic et al. 2000). Simulations with these models were produced for the period 1971–2000. It is important to note that climate model projections do not attempt to predict the timing of meteorological events such as storms and droughts. FIgure 5 The results show substantial differences in the six pairs of RCM/AOGCM regional climate projections over the Western United States (figure 5). Half of the simulations indicate that precipitation will increase in the 2041–70 period compared with 1971–2000 under the SRES A2 emissions scenario, whereas the other half indicate that precipitation will decrease. This result demonstrates the inability of the models to agree on precipitation trends in the water-scarce western states and underscores the need for improvement of climate models. We conclude that even state-of-the-art climate -models hardly provide useful information about potential changes in precipitation trends in the future. Therefore, the best practice for water planning in this region should be to design resilient water systems that can cope with a wide range of scenarios of precipitation variability. The NARCCAP study was reported in 2012, and in 2013 the IPCC Fifth Assessment Report (AR5; IPCC 2014) was released along with climate model simulations known as CMIP5 (Coupled Model Inter-comparison Project Phase 5).[18] The CMIP5 outputs were produced for over 20 models from 20 modeling groups around the world. Four of the CMIP5 models were randomly selected and their projections of precipitation over the Western United States through the end of 21st century were compared. As shown in figure 6, the -models differ substantially in their projections: some show increasing trends and others decreasing trends over the same areas. Figure 6 Since the NARCCAP study and release of the CMIP5 model runs, research by modeling centers and the scientific community has yielded advances in both AOGCM and RCM, with better resolutions. When the CMIP6 climate model simulations for the Sixth IPCC Assessment Report are made available in 2021, it will be possible to assess (i) the models’ performance with respect to their ability to capture both precipitation patterns and trends in retrospective studies against available observations and (ii) agreement between models and their projections. Conclusions We offer four key observations about models of future hydrologic extremes and the needs of the water resources community: Despite advances, prediction of hydroclimate variables such as precipitation remains a major challenge. The accuracy of hydroclimate models falls short of meeting requirements for water resources planning and decision making. Nature is complex, and efforts to observe and model its nonlinear behavior are imperfect. Therefore, one should exercise caution and a willingness to question the reliability and credibility of information generated by models. Long-term and sustained observation programs are critical for model development and testing. Without model testing against extensive observations, the full potential of models in operational settings will not be achieved. As called for in ASCE (2018), building resiliency into water resource system design and operation is the best approach to meet water demand and cope with and adapt to the hazards of extreme floods and droughts. References Abbe C. 1888. Treatise on meteorological apparatus and methods. In: Annual Report of the Chief Signal Officer for 1887, appendix 46. Washington: Government Printing Office. ASCE [American Society of Civil Engineers]. 2018. Climate-Resilient Infrastructure: Adaptive Design and Risk Management, ed Ayyub BM. ASCE Manual of Practice 140. Reston VA. Asanjan AA, Faridzad M, Hayatbini N, Gorooh VA, Sadeghi M, Shearer EJ, Sorooshian S, Nguyen P, Hsu K, Taghian M. 2019. An assessment of the unprecedented extreme precipitation events over Iran: From satellite perspective. Online at http://chrs.web.uci.edu/articles/iran_rainfall.pdf. Ashouri H, Hsu KL, Sorooshian S, Braithwaite DK, Knapp KR, Cecil LD, Nelson BR, Prat OP. 2015. -PERSIANN-CDR: Daily precipitation climate data record from multisatellite observations for hydrological and climate studies. Bulletin of the American Meteorological Society 96(1):69–83. Ayyub BM, Hill AC. 2019. Climate-resilient infrastructure: Engineering and policy perspectives. The Bridge 49(2):8–15. Baecher G, Bensi M, Reilly A, Phillips B, Link LE, Knight S, Galloway G. 2019. Resiliently engineered flood and hurricane infrastructure: Principles to guide the next generation of engineers. The Bridge 49(2):26–33. Cayan DR, Das T, Pierce DW, Barnett TP, Tyree M, -Gershunov A. 2010. Future dryness in the southwest US and the hydrology of the early 21st century drought. Proceedings, National Academy of Sciences 107(50):21271–76. Cook ER, Woodhouse CA, Eakin CM, Meko DM, Stahle DW. 2004. Long-term aridity changes in the Western United States. Science 306(5698):1015–18. Easterling DR, Kunkel KE, Arnold JR, Knutson T, LeGrande AN, Leung LR, Vose RS, Waliser DE, Wehner MF. 2017.Precipitation change in the United States. In: Climate Science Special Report: Fourth National Climate Assessment, Vol I, eds Wuebbles DJ, Fahey DW, Hibbard KA, Dokken DJ, Stewart BC, Maycock TK. Washington: US Global Change Research Program. Freedman A. 2019. Why Typhoon Hagibis packed such a -deadly, devastating punch in Japan. Washington Post, Oct 14. IPCC [Intergovernmental Panel on Climate Change]. 2014. Climate Change 2014: Synthesis Report, eds Pachauri RK, Meyer LA. Contribution of Working Groups I, II, and III to the Fifth Assessment Report. Geneva. Laird KR, Fritz SC, Maasch KA, Cumming BF. 1996. Greater drought intensity and frequency before AD 1200 in the northern Great Plains, USA. Nature 384:552–54. Mearns LO, Arritt R, Biner S, Bukovsky MS, McGinnis S, Sain S, Caya D, Correia J Jr, Flory D, Gutowski W, and 10 others. 2012. The North American Regional Climate Change Assessment Program: Overview of phase I results. Bulletin of the American Meteorological Society 93(9):1337–62. Nakicenovic N, Alcamo J, Grubler A, Riahi K, Roehrl RA, Rogner HH, Victor N. 2000. Special Report on Emissions Scenarios (SRES). Working Group III of the -Inter-governmental Panel on Climate Change. Cambridge University Press. Nguyen P, Thorstensen A, Sorooshian S, Hsu K, -Aghakouchak A, Ashouri H, Tran H, Braithwaite D. 2018. Global precipitation trends across spatial scales using satellite observations. Bulletin of the American Meteorological Society 99(4):689–97. Seager R, Ting M, Held I, Kushnir Y, Lu J, Vecchi G, Huang HP, Harnik N, Leetmaa A, Lau NC, Li C. 2007. Model p-rojections of an imminent transition to a more arid climate in southwestern North America. Science 316(5828):1181–84. Vaughan A. 2014. England and Wales hit by wettest winter in nearly 250 years. The Guardian, Feb 27. Woodhouse CA, Overpeck JT. 1998. 2000 years of drought variability in the central United States. Bulletin of the American Meteorological Society 79(12):2693–714. [1] World Population Perspective, United Nations Department of Economics and Social Affairs population division (https://-population.un.org/wpp) [2] United States Census Bureau (https://www.census.gov) [3] The number of rain gauges changes depending on funding, political environment, and other factors. [4] https://opendata.dwd.de/climate_environment/GPCC/html/ download_gate.html [5] https://www.weather.gov/iwx/coop_8inch [6] https://www.nws.noaa.gov/oh/hdsc/index.html [7] https://www.ncdc.noaa.gov/cdr/atmospheric/precipitation- persiann-cdr [8] http://rainsphere.eng.uci.edu/ [9] www.nasa.gov/mission_pages/GPM/spacecraft/index.html [10] https://www.nasa.gov/content/goes-overview/index.html [11] “Mozambique flooding ‘worse than thought’: UN agency,” BBC News, Apr 28, 2019 (https://www.bbc.com/news/-world-africa-48087906) [12] Precipitation Frequency Data Server, HDSC, NOAA (https://hdsc.nws.noaa.gov/hdsc/pfds) [13] https://www.ncei.noaa.gov [14] Two articles in the summer 2019 issue of The Bridge address the importance of climate-resilient infrastructures capable of withstanding floods and hurricanes (Ayyub and Hill 2019; Baecher et al. 2019). [15] US Drought Monitor (https://droughtmonitor.unl.edu) [16] NOAA Office of Water Prediction, https://water.noaa.gov/about/nwm [17] https://www.cpc.ncep.noaa.gov [18] https://pcmdi.llnl.gov/mips/cmip5 About the Author:Soroosh Sorooshian (NAE) is a Distinguished Professor in the Departments of Civil and Environmental Engineering and Earth System Science and director of the Center for Hydrometeorology and Remote Sensing (CHRS), Henry Samueli School of Engineering, University of California, Irvine. Vesta Afzali Gorooh, Negin Hayatbini, Mohammed Ombadi, and Mojtaba Sadeghi are PhD students; Phu Nguyen is an assistant adjunct professor; and Kuolin Hsu is a professor, all at CHRS.