COMPARING A SIMPLE STOCHASTIC WEATHER GENERATOR WITH TWO COMMON STATISTICAL TECHNIQUES FOR GAP-FILLING DAILY RAINFALL IN HAWAIʻI

dc.contributor.advisor Chu, Pao-Shin
dc.contributor.author Henry, Julie
dc.contributor.department Atmospheric Sciences
dc.date.accessioned 2022-07-05T19:58:11Z
dc.date.available 2022-07-05T19:58:11Z
dc.date.issued 2022
dc.description.degree M.S.
dc.identifier.uri https://hdl.handle.net/10125/102178
dc.subject Atmospheric sciences
dc.subject Statistics
dc.subject daily rainfall
dc.subject gap-filling
dc.subject stochastic weather generator
dc.title COMPARING A SIMPLE STOCHASTIC WEATHER GENERATOR WITH TWO COMMON STATISTICAL TECHNIQUES FOR GAP-FILLING DAILY RAINFALL IN HAWAIʻI
dc.type Thesis
dcterms.abstract Considerable gaps and breaks of varying lengths are present in many historical daily rainfall records for Hawaiʻi. Countless gap-filling techniques exist, some more complex than others, and all are challenged by the spatially variable, complex terrain in Hawaiʻi. A stochastic weather generator (SWG) attempts to model the randomness and temporal persistence of rainfall and can be used as a gap-filling method. This study compares a simple two-state, first order Markov chain SWG for rainfall occurrence and amount, utilizing the mixed exponential distribution (MEXP), to two common gap-filling methods - quantile mapping (QM) and normal ratio (NR). QM and NR require at least one neighboring (“supporting”) station to gap-fill the target (“primary”) station. Whereas a single-site SWG uses the history of daily rainfall of the station itself, thus eliminating spatial bias. Test years were data masked to various percent coverages using gap types that exist in the record. Gap-filled series were evaluated for accuracy and analyzed for patterns according to station climatology.Overall results show QM is the best performing method when at least one supporting station is available. The SWG model had bias errors at least twice that of QM and NR, had a wider disparity between mean and median absolute error which is an indicator of model performance, and underpredicts monthly mean rainfall more than the other methods. However, it was the best performing method in a few instances, usually at dry stations or during the dry season. Moderate rainfall is possibly underpredicted due to parameters for the MEXP being estimated from only the data masked series. Calculating the transition probabilities and parameter estimations from the climatological record could yield more competitive results. Also, while the MEXP is not intended to model extreme events (avg daily rainfall > 100 mm), it is a good fit for the majority of non-zero daily rainfall. Incorporating an extreme value distribution into the SWG model may address the instances of heavy rainfall. Overall, the SWG method performed with sufficient skill for monthly occurrence and monthly mean rainfall to be considered a good model. If no nearby supporting station is available or missing substantial amounts of data the single-site SWG model is a viable method, but gap-filling any given station for temporal completeness may require a combination of methods to accurately capture the highly variable nature of rainfall in Hawaiʻi
dcterms.extent 74 pages
dcterms.language en
dcterms.publisher University of Hawai'i at Manoa
dcterms.rights All UHM dissertations and theses are protected by copyright. They may be viewed from this source for any purpose, but reproduction or distribution in any format is prohibited without written permission from the copyright owner.
dcterms.type Text
local.identifier.alturi http://dissertations.umi.com/hawii:11428
Files
Original bundle
Now showing 1 - 1 of 1
No Thumbnail Available
Name:
Henry_hawii_0085O_11428.pdf
Size:
3.17 MB
Format:
Adobe Portable Document Format
Description: