본문

서브메뉴

Moisture is Coming - Insights Into Predicting Atmospheric Rivers in Our Future
Moisture is Coming - Insights Into Predicting Atmospheric Rivers in Our Future
Moisture is Coming - Insights Into Predicting Atmospheric Rivers in Our Future

상세정보

자료유형  
 학위논문 서양
최종처리일시  
20260202103702
ISBN  
9798291573839
DDC  
551.5
저자명  
Higgins, Timothy Brown.
서명/저자  
Moisture is Coming - Insights Into Predicting Atmospheric Rivers in Our Future
발행사항  
[Sl] : University of Colorado at Boulder, 2025
발행사항  
Ann Arbor : ProQuest Dissertations & Theses, 2025
형태사항  
124 p
주기사항  
Source: Dissertations Abstracts International, Volume: 87-02, Section: B.
주기사항  
Advisor: Subramanian, Aneesh.
학위논문주기  
Thesis (Ph.D.)--University of Colorado at Boulder, 2025.
초록/해제  
요약The US West Coast is prone to high variability of precipitation, largely due to the occurrences of atmospheric rivers (ARs). ARs are long, narrow filamentary plumes of horizontal water vapor transport in the lower Troposphere that can threaten lives and damage property and infrastructure. It can therefore be beneficial to accurately predict and understand AR activity across various scales to help mitigate risk. This dissertation explores several techniques that can help understand and predict ARs in the climate, subseasonal to seasonal, and medium range (decades, weeks, and days into the future, respectively). Understanding changes to rare extreme AR events in varying climate warming scenarios can be challenging, largely due to the computational cost of both running large-ensemble climate models and tracking ARs in climate data with traditional methods. To help address the computational cost of creating a large ensemble, a unique dataset that used computing power from volunteers' computers was used. A machine learning method, "CG-Climate" was used for AR tracking, which substantially reduced the cost of AR tracking and was demonstrated to consistently detect the same events that other common methods did in reanalysis data. The reliability of CG-Climate enabled the analysis of changes to rare extreme AR events in various climate warming scenarios. In the subseasonal range, the predictability of horizontal vapor transport exceeding the 90th percentile, which is often associated with AR activity, was evaluated in comparison to that of precipitation. The relationship between the North Pacific Jet and IVT in the subseasonal range is also examined to better understand the source of potential predictability. Finally, 1000-member large ensemble of horizontal vapor transport fields is created with diffusion, a method of generating images with artificial intelligence. The diffusion model uses a medium-range deterministic forecast as a condition and creates high-quality realistic images by gradually predicting noise in noisy images and removing it. This approach yielded a promising result that could help aid AR predictions in the future.
일반주제명  
Atmospheric sciences
일반주제명  
Hydrologic sciences
일반주제명  
Geophysics
키워드  
Atmospheric rivers
키워드  
Diffusion
키워드  
Ensemble
키워드  
AR extremes
키워드  
Machine learning
키워드  
Predictions
기타저자  
University of Colorado at Boulder Atmospheric and Oceanic Sciences
기본자료저록  
Dissertations Abstracts International. 87-02B.
전자적 위치 및 접속  
로그인 후 원문을 볼 수 있습니다.

MARC

 008260126s2025        us                              c    eng  d
■001000017358231
■00520260202103702
■006m          o    d                
■007cr#unu||||||||
■020    ▼a9798291573839
■035    ▼a(MiAaPQ)AAI32113089
■040    ▼aMiAaPQ▼cMiAaPQ
■0820  ▼a551.5
■1001  ▼aHiggins,  Timothy  Brown.▼0(orcid)0000-0002-2204-9193
■24510▼aMoisture  is  Coming  -  Insights  Into  Predicting  Atmospheric  Rivers  in  Our  Future
■260    ▼a[Sl]▼bUniversity  of  Colorado  at  Boulder▼c2025
■260  1▼aAnn  Arbor▼bProQuest  Dissertations  &  Theses▼c2025
■300    ▼a124  p
■500    ▼aSource:  Dissertations  Abstracts  International,  Volume:  87-02,  Section:  B.
■500    ▼aAdvisor:  Subramanian,  Aneesh.
■5021  ▼aThesis  (Ph.D.)--University  of  Colorado  at  Boulder,  2025.
■520    ▼aThe  US  West  Coast  is  prone  to  high  variability  of  precipitation,  largely  due  to  the  occurrences  of  atmospheric  rivers  (ARs).  ARs  are  long,  narrow  filamentary  plumes  of  horizontal  water  vapor  transport  in  the  lower  Troposphere  that  can  threaten  lives  and  damage  property  and  infrastructure.  It  can  therefore  be  beneficial  to  accurately  predict  and  understand  AR  activity  across  various  scales  to  help  mitigate  risk.  This  dissertation  explores  several  techniques  that  can  help  understand  and  predict  ARs  in  the  climate,  subseasonal  to  seasonal,  and  medium  range  (decades,  weeks,  and  days  into  the  future,  respectively).  Understanding  changes  to  rare  extreme  AR  events  in  varying  climate  warming  scenarios  can  be  challenging,  largely  due  to  the  computational  cost  of  both  running  large-ensemble  climate  models  and  tracking  ARs  in  climate  data  with  traditional  methods.  To  help  address  the  computational  cost  of  creating  a  large  ensemble,  a  unique  dataset  that  used  computing  power  from  volunteers'  computers  was  used.  A  machine  learning  method,  "CG-Climate"  was  used  for  AR  tracking,  which  substantially  reduced  the  cost  of  AR  tracking  and  was  demonstrated  to  consistently  detect  the  same  events  that  other  common  methods  did  in  reanalysis  data.  The  reliability  of  CG-Climate  enabled  the  analysis  of  changes  to  rare  extreme  AR  events  in  various  climate  warming  scenarios.  In  the  subseasonal  range,  the  predictability  of  horizontal  vapor  transport  exceeding  the  90th  percentile,  which  is  often  associated  with  AR  activity,  was  evaluated  in  comparison  to  that  of  precipitation.  The  relationship  between  the  North  Pacific  Jet  and  IVT  in  the  subseasonal  range  is  also  examined  to  better  understand  the  source  of  potential  predictability.  Finally,  1000-member  large  ensemble  of  horizontal  vapor  transport  fields  is  created  with  diffusion,  a  method  of  generating  images  with  artificial  intelligence.  The  diffusion  model  uses  a  medium-range  deterministic  forecast  as  a  condition  and  creates  high-quality  realistic  images  by  gradually  predicting  noise  in  noisy  images  and  removing  it.  This  approach  yielded  a  promising  result  that  could  help  aid  AR  predictions  in  the  future.
■590    ▼aSchool  code:  0051.
■650  4▼aAtmospheric  sciences
■650  4▼aHydrologic  sciences
■650  4▼aGeophysics
■653    ▼aAtmospheric  rivers
■653    ▼aDiffusion
■653    ▼aEnsemble
■653    ▼aAR  extremes
■653    ▼aMachine  learning
■653    ▼aPredictions
■690    ▼a0725
■690    ▼a0388
■690    ▼a0800
■690    ▼a0373
■71020▼aUniversity  of  Colorado  at  Boulder▼bAtmospheric  and  Oceanic  Sciences.
■7730  ▼tDissertations  Abstracts  International▼g87-02B.
■790    ▼a0051
■791    ▼aPh.D.
■792    ▼a2025
■793    ▼aEnglish
■85640▼uhttp://www.riss.kr/pdu/ddodLink.do?id=T17358231▼nKERIS▼z이  자료의  원문은  한국교육학술정보원에서  제공합니다.

미리보기

내보내기

chatGPT토론

Ai 추천 관련 도서


    신착도서 더보기
    최근 3년간 통계입니다.

    소장정보

    • 예약
    • 소재불명신고
    • 나의폴더
    • 우선정리요청
    • 비도서대출신청
    • 야간 도서대출신청
    소장자료
    등록번호 청구기호 소장처 대출가능여부 대출정보
    TF16627 전자도서 대출가능 마이폴더 부재도서신고 비도서대출신청 야간 도서대출신청

    * 대출중인 자료에 한하여 예약이 가능합니다. 예약을 원하시면 예약버튼을 클릭하십시오.

    해당 도서를 다른 이용자가 함께 대출한 도서

    관련 인기도서

    로그인 후 이용 가능합니다.