The Impact Of Global Warming On Regional Climate And Storm Surge Activities Over U.S. East Coastal Waters

Under the global warming condition, a rising sea level over U.S. east coast is expecteddue to the thermal expansions of ocean and glacial melt, as well as a dynamical adjustmentof sea level to the possible slowdown of Atlantic Meridional Overturning Circulation(AMOC). This sea level rising trend makes the U.S. east coast even more sensitive andvulnerable to storm-surge floodings, which make it necessary and meaningful to assess theimpact of climate change over the U.S. east coastal waters.Although the global coupled models from the IPCC have provided valuable informationit can not be used directly to infer the impact of climate change in the regional ocean due toits coarse resolution (typical2.5°for atmosphere and1°for ocean). Thus it is necessaryto downscale the IPCC data before any applications for the regional study. Here amultivariate statistical downscaling method is developed to produce regional,high-resolution, coastal surface wind fields based on the global model predictions for theU.S. east coast and Gulf of Mexico. The statistical relationship is built upon linearregressions between the empirical orthogonal functions(EOF) space of a cross-calibrated,multi-platform (CCMP), multi-instrument ocean surface wind velocity data set (predictand)and that of the global NCEP wind reanalysis (predictor) covering a10-year period from2000to2009.The validation shows a higher correlation of0.9between modeled andobserved data, with its explained variance exceeding80%in most region and thediscrepancy between modeled and observed data varies from2.0~3.0m/s: Besides, thecomparisons between model and buoys data show a good agreement both in phase andamplitude, and cross-validation also confirms the insensitivity of statistical model to thechoice of predictand/predictor datasets. The successful reproduction of main patterns andfeatures over the U.S. east coastal waters indicates the capability of our statistical model indiagnosing the impact of climate change over our study region. The resulting predictand-predictor relationship is then applied to IPCC GFDL CM2.1output (2.5°×2.5°) to downscale coastal wind at0.25°×0.25°resolution for0141-0180(PI scenario) and2001-2100(A1FI scenario) respectively. The temporal andspatial variability of “future” wind speeds and wind energy potential over the U.S. eastand gulf coastal waters are quantified: in winter, the decrease of mean wind speed is foundover Atlantic between32°N-40°N with its maxima of more than-0.56m/s while a slightincrease exists in Gulf of Mexico; The maximum of wind speed is found to be reducedover the whole U.S. east waters, especially off the west of Florida Peninsula where itsreduction reaches-5m/s. The95percentile of wind speed presents a significant decreaseextending southeastward from New England region while a slight increase exists in theGulf of Mexico and off the Florida peninsula. In summer, a region with remarkabledecrease of mean wind speed can be found extending southwestward, with its maxima of-0.86m/s and core location nearby68°W, when the decrease trend can be seen in most ofstudy region. The maximum wind speed presents a significant decrease trend towardsouthwestward along the U.S. east coast with its amplitude reaching-4.0m/s; The95percentile of wind speed also show a weak increase of0.5-1.5m/s, with its core locationnearby64°W，32°N.The results of analysis for wind energy indicate: the significant decrease of mean windenergy mainly occurs to the north of32°N with a amplitude of-178W/m~2when a generalreduction of more than-1500W/m~2can be seen except in the northwest of Gulf of Mexico,where a increase of1000W/m~2exist. A reduction of300-600W/m~2for95percentile canbe found between32°N-42°N, toward southeastward from Cape cod, with its maxima of-759W/m~2.In summer, the decrease of mean wind energy is primarily found nearby64°W,38°N with a amplitude of-50W/m~2and maxima of-77W/m~2while maximum windenergy shows a general deceased pattern along the whole U.S. east coast with a amplitudeof more than-800W/m~2. And the95percentile shows a significant reduction ranging from100to250W/m~2nearby64°W,38°N with its maxima reaching-272W/m~2.In order to further assess the impact of climate change on the storm surge, a3Dregional surge model is thus built based on ROMS with large-scale circulation and tidaleffects included and then the statistical downscaling wind plus SLP filed are used to drive this model to explore the changes of storm surge activities over the U.S. east waters.The results of99percentile of surge height indicate: A remarkable decrease of surgeheight can be seen over the whole MAB region with its relative change basically greaterthan15%while the maxima of about40%occurs in the middle of MAB, decreasinggradually toward the end. Similar spatial pattern can also be found in the SAB but withweaker amplitude when a slight increase of less than6%appears in the Gulf of Mexico.The analysis of maximum surge height points out a great reduction in the MAB regionwith its amplitude of more than30%and similar spatial pattern with a peak value of60%;The SAB region is different from the MAB, with a increase ranging from20%to30%while no any significant changes occurs in the remaining regions.The analysis of histogram for the surge height shows a consistent pattern in the MABregion, suggesting a great decease in both amplitude and frequency for the surge floodings,especially in the extreme events, which further shows a negative shift of surge height in thehistogram. The thing is kind of complicated in the SAB since no any pattern or trend ispresented in this region. Moving southward into Gulf of Mexico a significant increaseappears in both the amplitude and frequency while no any changes is found for theremaining.The accumulated extreme curves of surge height also confirm the remarkable decreasein the MAB region, with a reduction of more than10cm, while an opposite trend is foundin the northwest of Gulf of Mexico.To investigate the significance of high surge events, the Wilcoxton rank sum test isapplied to the high surge events, and the P-values in the MAB and northwest of Gulf ofMexico region are found to be less than0.03, which confirm the changes of surge height issignificant in a statistical meaning while most of stations are no-significant.