Abstract
Urbanisation is a recognized driver of changes in catchment river flow. However, quantifying the urban influence remains a major challenge, due to the brevity of land cover records and the challenge of isolating this signal from other drivers. This study assesses the contribution of urbanisation to changes in river discharge across different seasons and quantiles (low, median, high, mean, and peak flows). Twelve catchments (21–1660 km2) are selected after screening all gauged UK catchments for minimal human influences other than significant changes in urban land cover. Generalized Additive Models for Location, Scale and Shape (GAMLSS) are developed using long (40–63 years) historical records of precipitation, temperature, urban land cover, and daily river discharge (m3/s). Model coefficients reveal that increased urban area is associated with a rise in discharge across all flow quantiles and seasons, on average, and the contribution of urbanisation to non-stationarity is stronger for low flows and average flows than it is for high flows. For every 1 % increase in urban land cover there is an associated increase in the median of 1.9 % ±2.8 % (1 s.d.) for low flow, 0.9 % ±2.3 % (1 s.d.) for median flow, 0.9 % ±1.9 % (1 s.d.) for mean flow, 1.1 % ±2.0 % (1 s.d.) for high flow, and 0.5 % ±2.2 % (1 s.d.) for seasonal maximum flow across seasons. The urbanisation-flow signal tends to be greatest in catchments with less initial urban extent and low bedrock permeability.
Original language | English |
---|---|
Article number | 128417 |
Number of pages | 16 |
Journal | Journal of Hydrology |
Volume | 613 |
Issue number | Part A |
Early online date | 28 Aug 2022 |
DOIs | |
Publication status | Published - Oct 2022 |
Keywords
- Urbanisation
- River flow
- Non-stationary
- Urban extent
- Statistical modelling
- Attribution