Home / Current Issue / Paper 1722674
Field Condition Assessment of Roadside Drainage and Culvert Structures on Urban Arterials in Southwestern and North Central Nigeria
Subject area: Science,Engineering and Technology · Area of research: Highway and Drainage Engineering
Abstract
Roadside drainage carries the flood risk of Nigerian cities and its condition is neither inventoried nor rated, so an agency that would prioritize maintenance has nothing to work on. The available protocols were written for closed culvert barrels and piped sewers, not for the open lined channels of a tropical urban arterial. This paper develops a five-point rating protocol for such assets, paired with a design that rates a fifth of the inventory twice and reports agreement disaggregated. Both were applied to an inventory of 949 assets along 23.99 km of arterial alignment in Osogbo and Ilorin. Pooled agreement looks usable: Cohen’s κ is 0.561, quadratically weighted κ is 0.841 and Krippendorff’s α is 0.842, with the teams within one grade on 98.0 percent. Disaggregated it does not. Agreement on covered drains falls to 0.277 against 0.682 for open lined channels, and cutting the statistic by city rather than asset type invites the wrong diagnosis entirely. The general result is arithmetic: a pooled statistic can conceal the subgroup in which an instrument has failed, and a pooled spatial test can too. We recommend that no condition dataset be published without a disaggregated agreement statistic, that surface-derived grades be used in bands rather than at a point, and that covered assets be coded as unassessed.
Keywords
drainage condition assessment, culvert inspection, inter-rater reliability, ordinal regression, urban flooding, asset management, nigeria
How to cite this paper
@article{1722674,
author = {Olayemi Bolaji, Oladimeji Basit Alaka, Oluwabusola Ajayi, Ridwan Tiamiyu},
title = {Field Condition Assessment of Roadside Drainage and Culvert Structures on Urban Arterials in Southwestern and North Central Nigeria},
journal = {Iconic Research And Engineering Journals},
year = {2019},
volume = {3},
number = {1},
pages = {663-683},
issn = {2456-8880},
url = {https://www.irejournals.com/formatedpaper/1722674.pdf},
abstract = {Roadside drainage carries the flood risk of Nigerian cities and its condition is neither inventoried nor rated, so an agency that would prioritize maintenance has nothing to work on. The available protocols were written for closed culvert barrels and piped sewers, not for the open lined channels of a tropical urban arterial. This paper develops a five-point rating protocol for such assets, paired with a design that rates a fifth of the inventory twice and reports agreement disaggregated. Both were applied to an inventory of 949 assets along 23.99 km of arterial alignment in Osogbo and Ilorin. Pooled agreement looks usable: Cohen’s κ is 0.561, quadratically weighted κ is 0.841 and Krippendorff’s α is 0.842, with the teams within one grade on 98.0 percent. Disaggregated it does not. Agreement on covered drains falls to 0.277 against 0.682 for open lined channels, and cutting the statistic by city rather than asset type invites the wrong diagnosis entirely. The general result is arithmetic: a pooled statistic can conceal the subgroup in which an instrument has failed, and a pooled spatial test can too. We recommend that no condition dataset be published without a disaggregated agreement statistic, that surface-derived grades be used in bands rather than at a point, and that covered assets be coded as unassessed.},
keywords = {drainage condition assessment, culvert inspection, inter-rater reliability, ordinal regression, urban flooding, asset management, nigeria},
month = {July},
}