Understanding Impacts of Excess Heat in Filipino cities#
The Project Development Objective is to empower urban Philippines Local Government Units (LGUs) to take preventative and real-time action to mitigate the impacts of heatwaves through provision of more timely and granular data products that reveal localized occurrences of heatwaves and their impacts on human health and urban economies.
Urban heatwaves have cascading impacts on the economies and well-being of people, including heat-induced hospitalizations, power-grid failure, declines in labor productivity, shortened school days, and others. Unless the impacts of heatwaves are mitigated, Filipino cities, with their high population density, will struggle to meet their full economic potential and ensure the wellbeing of their people.
Until now, cities have been challenged to address heatwaves, because the data and expertise needed to mitigate impacts are typically scattered across many departments. To overcome this barrier, as part of the World Bank’s Technical Deep Dive on Urban Heat (April 24-28, 2023), Filipino authorities have requested support on better measurement of urban heatwaves and their impacts. To meet this demand-driven request, the project team would facilitate access to heat risk and vulnerability information for the most populous cities in the three regions of the Philippines. Through this public data good, urban governments will be able to identify investments and policy interventions that build resilience at a community level and support scaling up the resilience of the city. The knowledge product will be validated through local data collection and stakeholder participation to ensure inclusive design and participation.
In the Philippines, more than 40 percent of the population lives in cities, and many of these cities are at high risk of experiencing heatwaves.Its position as part of Southeast Asia, with high temperatures and humidity already being the norm, and the 7641 islands that make up the country are especially susceptible to climate change. Challenges in exposition, vulnerability, and adaptability to climate made the Philippines become the country with highest extreme weather disaster-related risk in the world according to the 2022 World Risk Report. Additionally, rapid urbanization is expected to continue as “drought and low farm productivity” drive migrants into cities. The consequences of heatwaves to these residents will be profound, and call for new methods to better understand their behaviors and propose better mitigation strategies. This intersects with the high inequality found in Filipino cities: in the first semester of 2021, driven by the Covid-19 pandemic, 23.7% of the country’s overall population was considered poor. Studies have shown strong correlation between poverty and vulnerability to heatwaves, and low-income housing has poor thermal performance and high thermal discomfort levels across the globe. Even without taking into consideration socioeconomic status, a vast majority of Filipinos experience heat stress severely (as self-reported), and this increases with population density.
The serious consequences are already being felt, with recent months showcasing extreme temperatures. In April 2023, for example, dangerous heat was observed throughout the country.
In this notebook, we visualize the areas within Filipino cities that are adversely impacted by heat using multiple data sources, and try to understand some of the aspects that have led to this upending crisis.
Population of Philippines#
The data for population is taken from Meta High Resolution Population Maps
Calculate population aggregated to different admin regions
Population of the Philippines as of 2020#
Show code cell source
import matplotlib.pyplot as plt
fig, ax = plt.subplots(1, 3, figsize=(36, 6))
plt.rcParams["font.family"] = "cursive"
PHILIPPINES_ADM1.plot(column="phl_general_2020", legend=False, ax=ax[0])
PHILIPPINES_ADM2.plot(column="phl_general_2020", legend=False, ax=ax[1])
PHILIPPINES_ADM3.plot(column="phl_general_2020", legend=True, ax=ax[2])
for i in range(0, 2):
ax[i].spines["top"].set_visible(False)
ax[i].spines["bottom"].set_visible(False)
ax[i].spines["right"].set_visible(False)
ax[i].spines["left"].set_visible(False)
ax[i].set_xticks([])
ax[i].set_yticks([])
[]
Observation
At the Admin 3 level, Cebu looks more populated than Manila. However, Manila comprises of 4 admin 3 regions making it a lot more densely populated
Measure historical heatwaves and heat index trends in the Philippines using NASA GLDAS data#
Heatwaves are generally understood to be the combination of extreme temperatures over multiple days. A more quantitative definition, however, is very contested and there is no definitive answer, with different characterizations being useful at different times or contexts.
Heat Wave Duration Index: HWDI or heatwave duration index shows the number of heatwaves according to multiple modifications of the WMO definitions of heatwaves. Here, a heatwave is defined as 2 consecutive days where the mean temperature is more than 5 degrees of the mean temperature of the baseline from 1948 to 2014.
Heat Index: Heat index aka apparent temperature is based on work carried out by Robert G. Steadman in 1979 (“An Assessment of Sultriness, Parts I and II”) where he discussed factors that would impact how hot a person would feel under certain conditions. It incorporates 21 parameters and assumptions: body mass (147.7 lbs), height (5’7”), actively walking (3.1 mph), clothing (pants and short sleeve shirt), heat tolerance, in the shade, etc. This formula became the “heat index”. It is the traditional measurement of heat stress due to high temperatures and high humidity. Heat index is calculated for shady areas The inputs to this are - temperature, relative humidity Equation: Heat Index = -42.379 + 2.04901523T + 10.14333127RH - .22475541TRH - .00683783TT - .05481717RHRH + .00122874TTRH + .00085282TRHRH - .00000199TTRHRH
T = ambient dry temperature (in Fahrenheit)
R = relative humidity (percentage)
output_notebook()
show(
get_line_plot(
phl_cities_hwdi_mean2,
measure="heat_waves_per_time_period",
title="Heatwaves per time period",
subtitle="heatwave is defined as 2 consecutive days where the mean temperature is more than 5 degrees of the mean temperature of the baseline from 1948 to 2014",
source="GLDAS data with analysis from the Data Lab, World Bank",
)
)
phl_cities_heatindex_year = (
phl_cities_heatindex[["city", "time", "heat_index"]]
.groupby([pd.Grouper(key="time", freq="Y"), "city"])
.agg({"mean", "max", "min"})
.reset_index()
)
phl_cities_heatindex_year.set_index(["time", "city"], inplace=True)
phl_cities_heatindex_year.columns = phl_cities_heatindex_year.columns.droplevel(0)
phl_cities_heatindex_year.reset_index(inplace=True)
phl_cities_heatindex["time"] = phl_cities_heatindex["time"].apply(
lambda x: datetime.strptime(x, "%Y-%m-%d %H:%M:%S")
)
output_notebook()
from bokeh.models import Panel
bokeh.core.validation.silence(EMPTY_LAYOUT, True)
bokeh.core.validation.silence(MISSING_RENDERERS, True)
tabs = []
for value in list(["min", "max", "mean"]):
tabs.append(
Panel(
child=get_line_plot(
phl_cities_heatindex_year.reset_index(),
"Heat Index Per Year",
"Source: GLDAS NASA - Heat index calculated using NWS (2011)",
measure=value,
subtitle="Heat index is calculated using both temperature and relative humidity",
),
title=value.capitalize(),
)
)
tabs = Tabs(tabs=tabs, sizing_mode="scale_both")
show(tabs, warn_on_missing_glyphs=False)
Comparing heatwaves and heat index from different sources for April 2023#
Heat Index from PAGASA: Philippine Atmospheric, Geophysical and Astronomical Services Administration (PAGASA).
Show code cell source
output_notebook()
show(
get_line_plot(
pd.concat([manila_heatindex_tomorrowio_max, phl_cities_heatindex_pagasa_max]),
measure="Maximum Heat Index",
title="Max Heat Index",
subtitle="heat index calculated from Tomorrow.io",
source="Tomorrow.io",
)
)
Show code cell source
output_notebook()
show(
get_line_plot(
pd.concat([manila_heatindex_tomorrowio_mean, phl_cities_heatindex_pagasa]),
measure="Average Heat Index",
title="Mean Heat Index",
subtitle="heat index calculated from Tomorrow.io",
source="Tomorrow.io",
)
)
Observations#
The mean heat index from PAGSA and tomorrow.io aligns almost exactly but the max values are far apart from each other
Plotting heat index at 5km grids#
By plotting the heat index on resolution 7 hex grids, we noticed that there is difference in temperatures in different regions of Manila
Next Steps#
Compare Heat Index in Manila with other cities
Analyse change in visitations to Points of Interest with changing heat index (using a methodology similar to that used in Turkiye Rapid Needs Assessment)