Statistical Downscaling along the US Eastern Coast by Two Methods with Application on Intensity-Duration-Frequency curve Changes

Statistical Downscaling along the US Eastern Coast by Two Methods with Application on Intensity-Duration-Frequency curve Changes

Bibliographic Details
Main Author: Wang, Yaoping
Language:English
Published: The Ohio State University / OhioLINK 2015
Subjects:
Environmental Science
Environmental Engineering
precipitation
Rglimclim
quantile mapping
downscaling
eastern US
intensity-duration-frequency
Online Access:http://rave.ohiolink.edu/etdc/view?acc_num=osu1417824178
id ndltd-OhioLink-oai-etd.ohiolink.edu-osu1417824178
record_format oai_dc
collection NDLTD
language English
sources NDLTD
topic Environmental Science
Environmental Engineering
precipitation
Rglimclim
quantile mapping
downscaling
eastern US
intensity-duration-frequency
spellingShingle Environmental Science
Environmental Engineering
precipitation
Rglimclim
quantile mapping
downscaling
eastern US
intensity-duration-frequency

Wang, Yaoping
Statistical Downscaling along the US Eastern Coast by Two Methods with Application on Intensity-Duration-Frequency curve Changes
author Wang, Yaoping
author_facet Wang, Yaoping
author_sort Wang, Yaoping
title Statistical Downscaling along the US Eastern Coast by Two Methods with Application on Intensity-Duration-Frequency curve Changes
title_short Statistical Downscaling along the US Eastern Coast by Two Methods with Application on Intensity-Duration-Frequency curve Changes
title_full Statistical Downscaling along the US Eastern Coast by Two Methods with Application on Intensity-Duration-Frequency curve Changes
title_fullStr Statistical Downscaling along the US Eastern Coast by Two Methods with Application on Intensity-Duration-Frequency curve Changes
title_full_unstemmed Statistical Downscaling along the US Eastern Coast by Two Methods with Application on Intensity-Duration-Frequency curve Changes
title_sort statistical downscaling along the us eastern coast by two methods with application on intensity-duration-frequency curve changes
publisher The Ohio State University / OhioLINK
publishDate 2015
url http://rave.ohiolink.edu/etdc/view?acc_num=osu1417824178
work_keys_str_mv AT wangyaoping statisticaldownscalingalongtheuseasterncoastbytwomethodswithapplicationonintensitydurationfrequencycurvechanges
_version_ 1719437448147107840
spelling ndltd-OhioLink-oai-etd.ohiolink.edu-osu14178241782021-08-03T06:28:36Z Statistical Downscaling along the US Eastern Coast by Two Methods with Application on Intensity-Duration-Frequency curve Changes Wang, Yaoping Environmental Science Environmental Engineering precipitation Rglimclim quantile mapping downscaling eastern US intensity-duration-frequency Climate change is now a known fact and is unlikely to be reversible in any short term. The global temperature increase is accompanied by changes in atmospheric circulation and intensification of the hydrologic cycle, causing changes in local precipitation characteristics. The expected characteristics and impacts of precipitation change is different for different parts of the US. Therefore, an indispensable tool in local climate impact assessment is downscaling, which translates large-scale climate model projections to higher resolutions that can be fed into impact models on the local scale. This study conducts statistical downscaling at twelve weather stations along the eastern coast of the US for the historical and future climate scenarios from an ensemble of 18 Global Climate Models (GCMs). Two downscaling methods, quantile-mapping and Rglimclim, are tested using cross-validation during the historical period. A GCM-weighting method, Reliability-Ensemble-Averaging [Filippo Giorgi and Mearns, 2002] is tested on its ability to reduce GCM-related uncertainty in downscaling results. Precipitation characteristic changes are examined on mean, frequency, persistence, extreme statistics, and Intensity-Duration-Frequency curves. The following results are found. The quantile-mapping method, after a priori bias-correction on the grid-level precipitation and calibration on a monthly basis, performed well on most precipitation statistics, but overestimated the length of dry and wet spells at the southeastern US locations. The overestimation is stronger for the more southerly location, reaching, respectively, about 10% and 25% at the Florida location. The Rglimclim downscaling performance is also good except for small underestimation of the length of wet- and dry-spells and severe underestimation of the 90% precipitation intensities, the latter being sometimes as much as -50%. The most important factor in Rglimclim performance is predictor choice. Air temperature, specific humidity, and geopotential height are found to be unsuitable predictors for downscaling future precipitation with Rglimclim. Uncertainty reduction through GCM-weighting was found to be unproductive in the context of downscaling to local precipitation change, especially if weighting is based on final downscaling performance. The projected future precipitation change are mostly statistically insignificant due to large spread (+/- 50% to >=+/-100%) among GCMs, but exceptionally good agreement is reached on increased winter precipitation volume at the northeastern locations and increased extreme precipitation at all locations, where about 90-100% of the GCMs agree on the sign of change. The quantile-mapping downscaled changes between 1961-1990 and 2070-2099 under RCP4.5 and RCP8.5 are about 20-40%, and Rglimclim downscaled changes mostly between 0-10%, with sporadic negative changes. Daily to 10-day precipitation intensities at various frequencies are projected to increase between 5-15% based on downscaling results at all locations. While this level of increase is not drastic, engineering designs should best take climate change into account to prevent undesirable capacity exceedances and flood damage. Future work is needed to obtain higher-resolution downscaled precipitation so as to construct intensity-duration-frequency relationships that are more relevant to the urban environment. 2015-05-15 English text The Ohio State University / OhioLINK http://rave.ohiolink.edu/etdc/view?acc_num=osu1417824178 http://rave.ohiolink.edu/etdc/view?acc_num=osu1417824178 unrestricted This thesis or dissertation is protected by copyright: all rights reserved. It may not be copied or redistributed beyond the terms of applicable copyright laws.

Similar Items