國立虎尾科技大學 |

An Advanced Approach to Assessing Flood Risk in Urban Environments.

紀錄類型:	書目-語言資料,手稿 : Monograph/item
正題名/作者:	An Advanced Approach to Assessing Flood Risk in Urban Environments./
作者:	Gong, Siming.
面頁冊數:	1 online resource (280 pages)
附註:	Source: Dissertations Abstracts International, Volume: 85-11, Section: B.
Contained By:	Dissertations Abstracts International85-11B.
標題:	Geographic information science. -
電子資源:	click for full text (PQDT)
ISBN:	9798382765525

An Advanced Approach to Assessing Flood Risk in Urban Environments.
Gong, Siming.

An Advanced Approach to Assessing Flood Risk in Urban Environments. - 1 online resource (280 pages)

Source: Dissertations Abstracts International, Volume: 85-11, Section: B.

Thesis (Ph.D.)--University of Technology Sydney (Australia), 2024.

Includes bibliographical references

Urban areas face heightened flood vulnerability due to expanding impervious surfaces, which require consistent research to enhance flood forecasting and management systems. Catchment modelling is increasingly popular in flood risk management. However, modelling urban catchments is challenging due to heterogeneous land use and land cover (LULC) patterns, especially struggle with the extensive interpretation and parameterisation of spatial datasets using conventional data processing methods, leading to inconsistent simulations. Deep Neural Networks (DNNs) have shown promising results in processing catchment data, learning complex spectral and textural relationships for LULC category identification and segmentation. Therefore, implementing DNNs to interpret intricate spatial data and estimate parameters holds significant research value.The project has primarily focused on devising a methodology to effectively apply DNNs in analysing LULC data to estimate parameters in urban catchment models. The assessments of approaches' adaptability, stability and consistency were advanced together as the sub-projects. An end-to-end data assimilation and processing strategy was proposed by integrating DNNs and the clustering algorithm to generate a pixel-based LULC map on high-resolution satellite imagery. A reproducible framework was developed to estimate parameters for the hydrological component of the catchment model at the subcatchment scale by grouping the generated LULC features. The cumulative likelihood and Kolmogorov-Smirnov (KS) methods were adopted to describe the distribution of the results and assess goodness-of-fitting. Then, the expanded LULC samples were quantified using information entropy (IE) to evaluate prediction stability, where administrative data defined the sampling scope and sample impervious fraction was translated into polynomial functions to analyse occurrence likelihood distribution of samples' impervious fraction.The integrated DNNs approach achieved excellent classification performance, where the MeanShift+UNet has the highest accuracy on the test set. The suitability assessment illustrates that all three methods are more suitable for semi-distributed modelling systems. Parameter estimation results suggest that LULC has a limited effect on subcatchment flow length, and the proposed approach can estimate parameters for initialising catchment modelling systems. The probability-fitting study of impervious samples reflects the distribution differential of impervious fractions under the attributes of administrative data. The IE stability test result shows a robust model that clarifies the different confident ranges of imperviousness estimation based on land zoning information. A model of the Alexandra Canal Catchment (1379.8 ha) in Sydney, Australia, was built and configured by generated parameters, continuous rainfall and tidal data to validate the approaches' performance in the ungauged catchment. The modelling result shows great consistency compared to historical models.

Electronic reproduction.
Ann Arbor, Mich. :
ProQuest,
2024

Mode of access: World Wide Web

ISBN: 9798382765525Subjects--Topical Terms:

1372545
Geographic information science.
Index Terms--Genre/Form:

554714
Electronic books.

An Advanced Approach to Assessing Flood Risk in Urban Environments.
LDR:04338ntm a22003737 4500 001 1148428
005 20240924101923.5
006 m o d
007 cr bn ---uuuuu
008 250605s2024 xx obm 000 0 eng d
020 $a 9798382765525
035 $a (MiAaPQ)AAI31355341
035 $a (MiAaPQ)TechSydney178219
035 $a AAI31355341
040 $a MiAaPQ $b eng $c MiAaPQ $d NTU
100 1 $a Gong, Siming. $3 1474384
245 1 3 $a An Advanced Approach to Assessing Flood Risk in Urban Environments.
264 0 $c 2024
300 $a 1 online resource (280 pages)
336 $a text $b txt $2 rdacontent
337 $a computer $b c $2 rdamedia
338 $a online resource $b cr $2 rdacarrier
500 $a Source: Dissertations Abstracts International, Volume: 85-11, Section: B.
500 $a Advisor: Ball, James.
502 $a Thesis (Ph.D.)--University of Technology Sydney (Australia), 2024.
504 $a Includes bibliographical references
520 $a Urban areas face heightened flood vulnerability due to expanding impervious surfaces, which require consistent research to enhance flood forecasting and management systems. Catchment modelling is increasingly popular in flood risk management. However, modelling urban catchments is challenging due to heterogeneous land use and land cover (LULC) patterns, especially struggle with the extensive interpretation and parameterisation of spatial datasets using conventional data processing methods, leading to inconsistent simulations. Deep Neural Networks (DNNs) have shown promising results in processing catchment data, learning complex spectral and textural relationships for LULC category identification and segmentation. Therefore, implementing DNNs to interpret intricate spatial data and estimate parameters holds significant research value.The project has primarily focused on devising a methodology to effectively apply DNNs in analysing LULC data to estimate parameters in urban catchment models. The assessments of approaches' adaptability, stability and consistency were advanced together as the sub-projects. An end-to-end data assimilation and processing strategy was proposed by integrating DNNs and the clustering algorithm to generate a pixel-based LULC map on high-resolution satellite imagery. A reproducible framework was developed to estimate parameters for the hydrological component of the catchment model at the subcatchment scale by grouping the generated LULC features. The cumulative likelihood and Kolmogorov-Smirnov (KS) methods were adopted to describe the distribution of the results and assess goodness-of-fitting. Then, the expanded LULC samples were quantified using information entropy (IE) to evaluate prediction stability, where administrative data defined the sampling scope and sample impervious fraction was translated into polynomial functions to analyse occurrence likelihood distribution of samples' impervious fraction.The integrated DNNs approach achieved excellent classification performance, where the MeanShift+UNet has the highest accuracy on the test set. The suitability assessment illustrates that all three methods are more suitable for semi-distributed modelling systems. Parameter estimation results suggest that LULC has a limited effect on subcatchment flow length, and the proposed approach can estimate parameters for initialising catchment modelling systems. The probability-fitting study of impervious samples reflects the distribution differential of impervious fractions under the attributes of administrative data. The IE stability test result shows a robust model that clarifies the different confident ranges of imperviousness estimation based on land zoning information. A model of the Alexandra Canal Catchment (1379.8 ha) in Sydney, Australia, was built and configured by generated parameters, continuous rainfall and tidal data to validate the approaches' performance in the ungauged catchment. The modelling result shows great consistency compared to historical models.
533 $a Electronic reproduction. $b Ann Arbor, Mich. : $c ProQuest, $d 2024
538 $a Mode of access: World Wide Web
650 4 $a Geographic information science. $3 1372545
650 4 $a Water resources management. $3 1178858
650 4 $a Environmental engineering. $3 557376
650 4 $a Hydrologic sciences. $3 1179176
650 4 $a Neural networks. $3 1011215
650 4 $a Floods. $3 972039
650 4 $a Stormwater management. $3 1474385
650 4 $a Geographic information systems. $3 554796
655 7 $a Electronic books. $2 local $3 554714
690 $a 0800
690 $a 0370
690 $a 0595
690 $a 0775
690 $a 0388
710 2 $a University of Technology Sydney (Australia). $3 1465278
710 2 $a ProQuest Information and Learning Co. $3 1178819
773 0 $t Dissertations Abstracts International $g 85-11B.
856 4 0 $u http://pqdd.sinica.edu.tw/twdaoapp/servlet/advanced?query=31355341 $z click for full text (PQDT)