Advancements in modelling and mapping the urban atmosphere in a temperate maritime climate

Sytse Koopmans

Research output: Thesisinternal PhD, WU

Abstract

The overarching goal of this thesis is to develop methods and models to create better insights in urban weather and climate. The approaches are diverse, ranging from data assimilation of crowdsourced urban weather observations to adjust a complex numerical weather prediction model for the urban environment, to a more scalable practical application of statistical urban morphological and meteorological relations in detailed urban heat maps. The specific goals are twofold: Firstly, to offer a realistic representation of the weather for different urban districts and their near surroundings, for both the current and the future climate. Secondly, to better understand the involved physical processes and to interpret the remaining uncertainties in modelling urban weather and climate.

In monitoring climate change long historical records of weather stations are essential. Changes in land-use in their surroundings through time can affect the observed climatological trend. Urbanization around weather stations is a world-wide phenomenon. Using the Weather Research and Forecasting (WRF) mesoscale model (Chapter 2), we estimate the impact of urbanization on the long 2-m temperature record of WMO station De Bilt (the Netherlands). However, the nearby city of Utrecht and its suburbs expanded during the 20th century. We find this urbanization during the 20th century has resulted in a 100-year temperature trend of 0.22 ± 0.06 K. This is about 20% of the observed 20th century climate trend at WMO station De Bilt. This has been done by comparing WRF mesoscale model simulations for the land-use conditions for the years 1900 and 2000 for the near full range of weather regimes.

Chapter 3 focuses on an improved urban climate model representation for a large city with Amsterdam as a test case. In the urban reanalysis for Amsterdam various observations and model fields are combined to produce a complete and coherent collection of consistent meteorological gridded data. This is achieved through data assimilation, which computes an intermediate atmospheric state based on the weighted uncertainties of model and observations. Here we utilize the WRF mesoscale model with 3D variational data assimilation for July 2014 as a testbed. The data assimilation module is fed with volume radar data and WMO routine weather station data. The novel part of our approach is the nudging of temperatures of (crowdsourced) personal weather station data in the city. To assimilate effectively urban temperatures a statistical scheme was developed to discard unreliable crowdsourced data, and a statistical fit was created based on genuine personal weather station observation-model differences and explanatory weather variables. Heat storage in the urban fabric is a major contribution to the Urban Heat Island (UHI) effect. Therefore, to effectively nudge urban model temperatures with personal weather station data, we adjust the urban fabric, which includes walls and roads. A model evaluation with independent urban observations reveals this urban nudging technique reduces the temperature biases, especially the cold bias seen at night.

Chapter 4 deals with urban heat effects of the additional housing issue in the Netherlands. 1 million houses must be built before 2040. A large part of this assignment has to be realized within the existing city borders. The densification of urban areas is expected to raise the UHI and has adverse effects on human health. Different urban planning strategies are investigated and evaluated on the impact on the UHI at the time of the minimum temperature for demarcated urban districts in The Hague, the Netherlands. Thereto, a validated diagnostic equation was applied to estimate the UHI based on routine meteorological observations and straightforward urban morphological properties as sky-view factor and vegetation fraction. The urban planning strategies differ in replacing low- and mid-rise buildings with high-rise buildings, and construction of low-rise buildings on neighbouring green areas which affects sky-view factor and vegetation fraction differently. They have in common that a fixed amount of population was housed in the urban districts. We find that, in most cases, the vegetation fraction is a more critical parameter than the sky-view factor to minimize the extra heat stress incurred when densifying a neighbourhood. This means that an urban planning strategy consisting of high-rise buildings and preserved green areas is often the optimal solution. In addition, meteorological observations are transformed according to four different KNMI climate scenarios in 2050. The impact of climate change on heat stress using these climate scenarios is clearly larger than the imposed urban densification.

Finally, Chapter 5 continues by investigating and providing a manual for a heat map with a suitable index that mimic more closely how people experience heat with the Physiological Equivalent Temperature (PET). The PET includes factors which affect the heat release of the human body by transpiration and ventilation by wind with dry air. For rural areas these required meteorological data are available from weather stations. However, for urban areas diagnostic meteorological and geographical models are required to calculate best estimates for urban specific weather. For example, the surface wind is typically reduced in urban areas and can be estimated by the height and frontal density of buildings and trees. The PET model is trained with 10 street configurations and 15-year climatological data with the advanced 3D energy model Rayman©. With a regression technique the PET was estimated based on weather station data and real geographical inputs. In the afternoon averaged PET heat map for Wageningen (the Netherlands) sun and shadow differences can be accurately determined with the used 1-m resolution. The resulting PET differences are therefore a useful and a step forward in mapping urban heat. However, verifications against traverse bike measurements reveal the largest uncertainties arise from radiation differences due to the size and transmissivity of tree crowns.

Original languageEnglish
QualificationDoctor of Philosophy
Awarding Institution
  • Wageningen University
Supervisors/Advisors
  • Holtslag, Bert, Promotor
  • Steeneveld, Gert-Jan, Co-promotor
Award date14 Sept 2021
Place of PublicationWageningen
Publisher
Print ISBNs9789463959339
DOIs
Publication statusPublished - 14 Sept 2021

Fingerprint

Dive into the research topics of 'Advancements in modelling and mapping the urban atmosphere in a temperate maritime climate'. Together they form a unique fingerprint.

Cite this