Evidence of horizontal urban heat advection in London using six years of data from a citizen weather station network

Abstract

Recent advances in citizen weather station (CWS) networks, with data accessible via crowd-sourcing, provide relevant climatic information to urban scientists and decision makers. In particular, CWS can provide long-term measurements of urban heat and valuable information on spatio-temporal heterogeneity related to horizontal heat advection. In this study, we make the first compilation of a quasi-climatologic dataset covering six years (2015-2020) of hourly near-surface air temperature measurements obtained via 1560 suitable CWS in a domain covering south-east England and Greater London. We investigated the spatio-temporal distribution of urban heat and the influences of local environments on climate, captured by CWS through the scope of Local Climate Zones (LCZ)-a land-use land-cover classification specifically designed for urban climate studies. We further calculate, for the first time, the amount of advected heat captured by CWS located in Greater London and the wider south east England region. We find that London is on average warmer by about 1.0 ^∘C-1.5 ^∘C than the rest of south-east England. Characteristics of the southern coastal climate are also captured in the analysis. We find that on average, urban heat advection (UHA) contributes to 0.22 ± 0.96 ^∘C of the total urban heat in Greater London. Certain areas, mostly in the centre of London are deprived of urban heat through advection since heat is transferred more to downwind suburban areas. UHA can positively contribute to urban heat by up to 1.57 ^∘C, on average and negatively by down to -1.21 ^∘C. Our results also show an important degree of inter- and intra-LCZ variability in UHA, calling for more research in the future. Nevertheless, we already find that UHA can impact green areas and reduce their cooling benefit. Such outcomes show the added value of CWS when considering future urban design.

Keywords: LCZ; London; Netatmo; citizen weather station; crowd-sourcing; urban climate; urban heat advection.

Figure 1.

Map of Local Climate Zones (LCZ) in the larger domain of study (domain 1) at 100 m horizontal resolution. The smaller domain (domain 2) is shown in purple with dashed lines representing the separation between four geographical quadrants for the UHA analysis. Black dots represent the location of quality-checked Netatmo Citizen Weather Stations (CWS) available within the time period from year 2015 to 2020. The Heathrow Airport MetOffice automatic weather station location is shown in fuchsia. LCZ source: Demuzere et al (2019).

Figure 2.

Flowchart diagram showing the different steps undertaken to calculate the downwind urban heat advection (UHA), the upwind-downwind gain/loss heat anomaly ΔUHA and the upwind-downwind temperature anomaly ΔT.

Figure 3.

6-year average (2015–2020) hourly urban heat advection (UHA) per downwind citizen weather station (CWS) in each Local Climate Zone and upwind prevailing winds. Large markers represent the cross-CWS median of the average UHA and triangle whiskers represent the 25$\text{th}$ and 75$\text{th}$ percentiles. The vertical axis has no dimension. Daytime (7AM to 7PM) and nighttime (7PM to 7AM local time) plots are given in figures S9 and S10, respectively.

Figure 4.

Hourly cross-CWS average UHA against wind speed and direction (first two rows above; dotted lines represent the limits for their respective classes), time-mean UHA per CWS against distance to the center of domain 2 (third row) and 6-year-hourly-mean UHA per CWS (fourth row). Each row is subdivided by LCZ to allow for inter- and intra-LCZ comparison.

Figure 5.

6-year average urban heat advection per citizen weather stations overlaid on the Local Climate Zones map of the Greater London area (domain 2). Dashed purple lines represent the quadrants borders. CWS where no UHA could be measured are made transparent.

Figure 6.

Example of the built-up environment’s three-dimensional complexity in the city of London overlaid on the European LCZ map by Demuzere et al (2019). Coordinates are in longitude and latitude.