Skip to main content
U.S. flag

An official website of the United States government

This article is part of the Fall 2022 issue of the Earth Science Matters Newsletter. 

structural equation model diagram
Figure 1: A Structural Equation Model (SEM) describing the interactions of urban land covers and urban temperatures. Numbers aligned with each arrow represent the partial regression coefficient of the connect relationship, with larger numbers indicating a stronger relationship and dotted lines indicated a non-significant relationship. Blue arrows represent a cooling effect, while red arrows represent a warming effect. In the upper panel representing daytime, the largest coefficient number (strongest relationship) between air temperature and a land cover category is on the blue arrow from tree canopy cover, meaning that during the day tree cover has the strongest cooling effect on air temperature of the land cover categories. In the bottom, nighttime panel, the largest coefficient number of the land cover categories is on the blue arrow between turf cover and air temperature, meaning that at night turf cover has the strongest cooling effect on air temperature. Credit – Ibsen, PC., et al, (2022), Science of the Total Environment

Urban natural resource managers often aim to achieve nature-based solutions to improve the resistance and resilience of urban ecosystems to climate change, and extreme heat waves. A major management concern requiring a nature-based solution is, how much urban tree canopy will it take to mitigate extreme heat and where should new tree plantings go to best serve city residents? To address that question, a team of USGS researchers developed a study to examine how air temperature is modulated by replacing one type of urban land cover with another. The research generated spatial maps of urban heat to help locate where tree plantings will provide the greatest cooling benefits.

The project had three major objectives:

  1. To understand how different urban land covers interact to determine local air temperatures 
  2. To quantify the heating and cooling effects of land cover to air temperature during normal days and days of extreme heat 
  3. To model local air temperature over an urban extent, and thus create maps of daytime and nighttime air temperature within the city 

Air temperature was selected as the key variable because it determines how people in cities experience heat. Much of the current urban climate research uses land surface temperature, which is easy to measure with satellite remote sensing, but does not directly measure heat as experienced by people. In addition, Denver, Colorado was selected because it is a semi-arid city, a missing piece in urban climate modelling that is dominated by studies in either humid or arid cities. Climate change is increasing aridity in currently humid regions of the U.S., so quantifying the land cover drivers in semi-arid Denver, helps resolve our understanding of climate impacts on urban heat.

The USGS team found that four types of urban land covers (trees, turf, impervious surfaces, and buildings) determine 25% of local air temperature during nighttime hours, and 17% during the day, with the remaining percentages determined by local weather conditions. Overall, the cooling magnitude of vegetated urban land covers (tree canopy and turf) is larger than the warming magnitude of urban impervious surfaces and buildings meaning that planting trees, particularly in very warm parts of the city could have a substantial impact on day-time temperatures. For example, replacing 25% of impervious surface with an urban tree canopy in a 60 m2 area can reduce local air temperatures by almost 1 °C during the day (from Figure 1 - every percent increase in tree canopy results in directly reducing air temperature by .026 °C by direct and indirect means, and every percent decrease in impervious surface results in reducing air temperature by .013 °C by direct and indirect means). Interestingly, air cooling at night was driven more by the presence of irrigated turf rather than trees, whereas during the day, turf had a minimal cooling effect. The team interpreted these results as tree shading and leaf transpiration drive daytime cooling, while irrigation and turf transpiration control nighttime cooling. 

These results were derived through a mixed methods approach, both in the data collection and statistical analysis. Scientists collected reference local air temperature through a network of approximately 100 microclimate sensors distributed over the study region in the summer of 2018 and obtained the land cover data through a local public dataset. Regional climate conditions were derived from satellite data and regional airport meteorological data. Data were analyzed with a Structural Equation Model (Figure 1), which provided the contributions of land cover to air temperatures, and the data were also put into a machine learning model, which predicted air temperature values over the entire study extent.

The resulting paper “Urban landcover differentially drives day and nighttime air temperature across a semi-arid city” was recently published in the journal Science of the Total Environment.

modeled day and nighttime urban air temperature in Denver, CO
Figure 2: Modeled day and nighttime urban air temperature in the city of Denver (footprint cutout in red in lower left corner), during a potential heat wave. Local temperature used for daytime was 35 °C, and for nighttime was 22° C. Blue colors on the map indicate areas that are cooler than the citywide mean air temperature, while red areas represent areas that are warmer than the citywide average temperature at that time. As we found tree canopy has the strongest cooling effect during the day, and impervious surface has the strongest warming effect at night, these maps can provide a target for where best to provide heat mitigation during the day and night through land cover change (planting trees and removing impervious surfaces). Credit – Ibsen, PC., et al, (2022), Science of the Total Environment

Related Content