Uncertainty analysis of artificial neural network based regional flood modelling in New South Wales, Australia

Nilufa Afrin; Sadia T. Mim; Ataur Rahman; Khaled Haddad

doi:10.1007/978-3-032-18708-6_3

Uncertainty analysis of artificial neural network based regional flood modelling in New South Wales, Australia

Nilufa Afrin
, Sadia T. Mim
, Ataur Rahman
, Khaled Haddad

School of Engineering

EnviroWater Sydney

Research output: Chapter in Book / Conference Paper › Chapter › peer-review

Abstract

This is an in-depth uncertainty analysis of Artificial Neural Network (ANN)-based Regional Flood Frequency Analysis (RFFA) in the Australian context. This uses data from 88 gauged stations in New South Wales, Australia. Using eight hydrological and physiographical predictors, six design flood quantiles (Q₂, Q₅, Q₁₀, Q₂₀, Q₅₀ and Q₁₀₀) were modelled. Model robustness was tested through a Monte Carlo simulation technique with random 70/30 train–test splits, producing a distribution of model outcomes across simulation runs. Uncertainty was quantified using the median relative error ratio (RE_r), yielding values of 59.62 ± 10.73% (Q₂), 55.86 ± 10.6% (Q₅), 54.13 ± 10.76% (Q₁₀), 55.86 ± 11.44% (Q₂₀), 57.72 ± 12.37% (Q₅₀), and 59.81 ± 10.58% (Q₁₀₀). Findings show that ANN performance varies with return period, with mid-range quantiles generally achieving lower uncertainty. These results demonstrate both the potential and the limitations of ANN-based RFFA, reinforcing the need for uncertainty analysis before applying AI-driven RFFA models to ungauged catchments.

Original language	English
Title of host publication	4th International Conference on Water and Environmental Engineering: Proceedings of iCWEE 2025
Editors	Ataur Rahman, Taha B. M. J. Ouarda, Muhammad Muhitur Rahman, Dharma Hagare, Zuhaib Siddiqui
Place of Publication	Switzerland
Publisher	Springer
Pages	34-49
Number of pages	16
ISBN (Electronic)	9783032187086
ISBN (Print)	9783032187079
DOIs	https://doi.org/10.1007/978-3-032-18708-6_3
Publication status	Published - 2026
Event	International Conference on Water and Environmental Engineering - Sydney, Australia Duration: 19 Nov 2025 → 21 Nov 2025 Conference number: 4th

Publication series

Name	Lecture Notes in Civil Engineering
Volume	822 LNCE
ISSN (Print)	2366-2557
ISSN (Electronic)	2366-2565

Conference

Conference	International Conference on Water and Environmental Engineering
Abbreviated title	iCWEE
Country/Territory	Australia
City	Sydney
Period	19/11/25 → 21/11/25

UN SDGs

This output contributes to the following UN Sustainable Development Goals (SDGs)

SDG 11 Sustainable Cities and Communities

Keywords

Artificial Intelligence
Artificial Neural Network
Regional Flood Frequency Analysis
Uncertainty Analysis
Ungauged Catchments

Access to Document

10.1007/978-3-032-18708-6_3

Cite this

Afrin, N., Mim, S. T., Rahman, A., & Haddad, K. (2026). Uncertainty analysis of artificial neural network based regional flood modelling in New South Wales, Australia. In A. Rahman, T. B. M. J. Ouarda, M. Muhitur Rahman, D. Hagare, & Z. Siddiqui (Eds.), 4th International Conference on Water and Environmental Engineering: Proceedings of iCWEE 2025 (pp. 34-49). (Lecture Notes in Civil Engineering; Vol. 822 LNCE). Springer. https://doi.org/10.1007/978-3-032-18708-6_3

Afrin, Nilufa ; Mim, Sadia T. ; Rahman, Ataur et al. / Uncertainty analysis of artificial neural network based regional flood modelling in New South Wales, Australia. 4th International Conference on Water and Environmental Engineering: Proceedings of iCWEE 2025. editor / Ataur Rahman ; Taha B. M. J. Ouarda ; Muhammad Muhitur Rahman ; Dharma Hagare ; Zuhaib Siddiqui. Switzerland : Springer, 2026. pp. 34-49 (Lecture Notes in Civil Engineering).

@inbook{272540b9d84a4bcea0cd381f34abae80,

title = "Uncertainty analysis of artificial neural network based regional flood modelling in New South Wales, Australia",

abstract = "This is an in-depth uncertainty analysis of Artificial Neural Network (ANN)-based Regional Flood Frequency Analysis (RFFA) in the Australian context. This uses data from 88 gauged stations in New South Wales, Australia. Using eight hydrological and physiographical predictors, six design flood quantiles (Q2, Q5, Q10, Q20, Q50 and Q100) were modelled. Model robustness was tested through a Monte Carlo simulation technique with random 70/30 train{\textendash}test splits, producing a distribution of model outcomes across simulation runs. Uncertainty was quantified using the median relative error ratio (REr), yielding values of 59.62 ± 10.73\% (Q2), 55.86 ± 10.6\% (Q5), 54.13 ± 10.76\% (Q10), 55.86 ± 11.44\% (Q20), 57.72 ± 12.37\% (Q50), and 59.81 ± 10.58\% (Q100). Findings show that ANN performance varies with return period, with mid-range quantiles generally achieving lower uncertainty. These results demonstrate both the potential and the limitations of ANN-based RFFA, reinforcing the need for uncertainty analysis before applying AI-driven RFFA models to ungauged catchments.",

keywords = "Artificial Intelligence, Artificial Neural Network, Regional Flood Frequency Analysis, Uncertainty Analysis, Ungauged Catchments",

author = "Nilufa Afrin and Mim, \{Sadia T.\} and Ataur Rahman and Khaled Haddad",

year = "2026",

doi = "10.1007/978-3-032-18708-6\_3",

language = "English",

isbn = "9783032187079",

series = "Lecture Notes in Civil Engineering",

publisher = "Springer",

pages = "34--49",

editor = "Ataur Rahman and Ouarda, \{Taha B. M. J.\} and \{Muhitur Rahman\}, Muhammad and Dharma Hagare and Zuhaib Siddiqui",

booktitle = "4th International Conference on Water and Environmental Engineering: Proceedings of iCWEE 2025",

note = "International Conference on Water and Environmental Engineering, iCWEE ; Conference date: 19-11-2025 Through 21-11-2025",

}

Afrin, N, Mim, ST, Rahman, A & Haddad, K 2026, Uncertainty analysis of artificial neural network based regional flood modelling in New South Wales, Australia. in A Rahman, TBMJ Ouarda, M Muhitur Rahman, D Hagare & Z Siddiqui (eds), 4th International Conference on Water and Environmental Engineering: Proceedings of iCWEE 2025. Lecture Notes in Civil Engineering, vol. 822 LNCE, Springer, Switzerland , pp. 34-49, International Conference on Water and Environmental Engineering, Sydney, New South Wales, Australia, 19/11/25. https://doi.org/10.1007/978-3-032-18708-6_3

Uncertainty analysis of artificial neural network based regional flood modelling in New South Wales, Australia. / Afrin, Nilufa; Mim, Sadia T.; Rahman, Ataur et al.
4th International Conference on Water and Environmental Engineering: Proceedings of iCWEE 2025. ed. / Ataur Rahman; Taha B. M. J. Ouarda; Muhammad Muhitur Rahman; Dharma Hagare; Zuhaib Siddiqui. Switzerland : Springer, 2026. p. 34-49 (Lecture Notes in Civil Engineering; Vol. 822 LNCE).

Research output: Chapter in Book / Conference Paper › Chapter › peer-review

TY - CHAP

T1 - Uncertainty analysis of artificial neural network based regional flood modelling in New South Wales, Australia

AU - Afrin, Nilufa

AU - Mim, Sadia T.

AU - Rahman, Ataur

AU - Haddad, Khaled

N1 - Conference code: 4th

PY - 2026

Y1 - 2026

N2 - This is an in-depth uncertainty analysis of Artificial Neural Network (ANN)-based Regional Flood Frequency Analysis (RFFA) in the Australian context. This uses data from 88 gauged stations in New South Wales, Australia. Using eight hydrological and physiographical predictors, six design flood quantiles (Q2, Q5, Q10, Q20, Q50 and Q100) were modelled. Model robustness was tested through a Monte Carlo simulation technique with random 70/30 train–test splits, producing a distribution of model outcomes across simulation runs. Uncertainty was quantified using the median relative error ratio (REr), yielding values of 59.62 ± 10.73% (Q2), 55.86 ± 10.6% (Q5), 54.13 ± 10.76% (Q10), 55.86 ± 11.44% (Q20), 57.72 ± 12.37% (Q50), and 59.81 ± 10.58% (Q100). Findings show that ANN performance varies with return period, with mid-range quantiles generally achieving lower uncertainty. These results demonstrate both the potential and the limitations of ANN-based RFFA, reinforcing the need for uncertainty analysis before applying AI-driven RFFA models to ungauged catchments.

AB - This is an in-depth uncertainty analysis of Artificial Neural Network (ANN)-based Regional Flood Frequency Analysis (RFFA) in the Australian context. This uses data from 88 gauged stations in New South Wales, Australia. Using eight hydrological and physiographical predictors, six design flood quantiles (Q2, Q5, Q10, Q20, Q50 and Q100) were modelled. Model robustness was tested through a Monte Carlo simulation technique with random 70/30 train–test splits, producing a distribution of model outcomes across simulation runs. Uncertainty was quantified using the median relative error ratio (REr), yielding values of 59.62 ± 10.73% (Q2), 55.86 ± 10.6% (Q5), 54.13 ± 10.76% (Q10), 55.86 ± 11.44% (Q20), 57.72 ± 12.37% (Q50), and 59.81 ± 10.58% (Q100). Findings show that ANN performance varies with return period, with mid-range quantiles generally achieving lower uncertainty. These results demonstrate both the potential and the limitations of ANN-based RFFA, reinforcing the need for uncertainty analysis before applying AI-driven RFFA models to ungauged catchments.

KW - Artificial Intelligence

KW - Artificial Neural Network

KW - Regional Flood Frequency Analysis

KW - Uncertainty Analysis

KW - Ungauged Catchments

UR - https://www.scopus.com/pages/publications/105031888620

UR - https://go.openathens.net/redirector/westernsydney.edu.au?url=https://doi.org/10.1007/978-3-032-18708-6_3

U2 - 10.1007/978-3-032-18708-6_3

DO - 10.1007/978-3-032-18708-6_3

M3 - Chapter

AN - SCOPUS:105031888620

SN - 9783032187079

T3 - Lecture Notes in Civil Engineering

SP - 34

EP - 49

BT - 4th International Conference on Water and Environmental Engineering: Proceedings of iCWEE 2025

A2 - Rahman, Ataur

A2 - Ouarda, Taha B. M. J.

A2 - Muhitur Rahman, Muhammad

A2 - Hagare, Dharma

A2 - Siddiqui, Zuhaib

PB - Springer

CY - Switzerland

T2 - International Conference on Water and Environmental Engineering

Y2 - 19 November 2025 through 21 November 2025

ER -

Afrin N, Mim ST, Rahman A, Haddad K. Uncertainty analysis of artificial neural network based regional flood modelling in New South Wales, Australia. In Rahman A, Ouarda TBMJ, Muhitur Rahman M, Hagare D, Siddiqui Z, editors, 4th International Conference on Water and Environmental Engineering: Proceedings of iCWEE 2025. Switzerland : Springer. 2026. p. 34-49. (Lecture Notes in Civil Engineering). doi: 10.1007/978-3-032-18708-6_3

Uncertainty analysis of artificial neural network based regional flood modelling in New South Wales, Australia

Abstract

Publication series

Conference

UN SDGs

Keywords

Access to Document

Other files and links

Fingerprint

Cite this