Climate Change
What natural catastrophes cost, how warm the planet is, how dry and how flooded the ground, how much burns — and what is being built against it
Thesis
Is climate change starting to show up in the economy?
- The planet is about 1.5 °C warmer than before industry.
- Drought, floods and fires trigger more emergencies each year.
- Public construction adapts, but no faster than the losses grow.
Index
Six indicators, standardized and equal-weighted, updated monthly — 100 at the start of the common sample, higher means more climate pressure
Climate Pressure Index
128.6
1 mo
↓ -1.0
1 yr
↑ +8.3
Source: Liquiqui, from Copernicus / NOAA / Census
Where each constituent stands today, in standard deviations from its own average over the sample
- Global temperatureMonthly · Sep 1, 2026+2.0
- US drought (Palmer, inv.)Monthly · Sep 1, 2026+1.9
- Storm & fire activationsMonthly · Sep 1, 2026+1.3
- Flood activationsMonthly · Sep 1, 2026+0.0
- US burned area, 52 weeksWeekly · Sep 1, 2026-1.1
- Public construction growthMonthly · Aug 1, 2026-0.4
Notes
- What is happening
- The index stands at 128.6 in September 2026, down from 129.6 in August, against a base of 100 in March 2013 and a record of 135.4 in December 2023. It climbed in steps: 112 on average in 2019, 130 in 2023, 129 in 2024, 124 in 2025; it had fallen back to 120 in January 2026 before rising again. Today temperature (+2.0 standard deviations) and drought (+1.9) carry the index, followed by extreme events (+1.3); floods sit at their average (0.0), and burned area (−1.1) and public construction (−0.4) pull down.
- Why it matters
- Climate change does not show in one series: temperature says what the planet is doing, drought and floods what it does to the ground, emergency activations and fires what it inflicts on people, public construction what is set against it. This index combines six measures at equal weight — the global temperature anomaly, the Palmer index inverted, Copernicus activations for storms, fires and landslides over twelve months, flood activations, US burned area over 52 weeks, and public construction growth — each standardised against its own history since 2013.
- What's driving it
- Temperature near record and the deepest US drought since the 1930s carry the index; a quiet fire year pulls the other way.
- What to watch
- Two constituents say most of it: temperature, which does not come back down, and drought, which is cyclical — 2012 was followed by 2019, the wettest year in the series. An index holding above 125 once the American drought breaks would mean extremes and fires have taken over. Public construction is the one constituent policy controls: growth below 2% with insured losses at $100 billion a year says adaptation is not keeping up.
Signals
Each indicator scored by how far it has moved — and the arrow points the way the pressure is heading
| Criterion | Current Reading | Trend | Comment |
|---|---|---|---|
Catastrophe lossestable only What do natural catastrophes cost, and how much of it is insured? | Moderate | ↓ | $224 billion of overall losses in 2025 on Munich Re's count, $108 billion of it insured, after 320 and 140 in 2024. Insured losses topped $100 billion for the fourth time in five years, against once before 2017. The first half of 2026 is quiet: 112 billion, 44 insured, below the ten-year average — and the Atlantic season had produced no hurricane by September 21st. |
Global temperaturein the index How much warmer is the planet than before the industrial era? | High | ↑ | 1.70°C above 1850-1900 in September 2026, after 1.73°C in August — the joint-warmest month ever measured, level with July 2023 on Copernicus's count — below the monthly record of 1.81°C of September 2023, and 1.78°C on October 11th. On twelve months, 1.49°C, against a peak of 1.64°C in mid-2024 and 1.21°C in 2019. The 1.5°C threshold was crossed on a running mean for most of 2024 and 2025; a very strong El Niño is forming for the winter. |
Extreme eventsin the index How many storms, wildfires and landslides trigger emergency mapping? | Moderate | ↑ | 62 Copernicus activations over the twelve months to September 2026, against 42 a year in 2019 and a record of 75 in May 2023. In the United States, NOAA's extremes index covered 31.7% of the country in 2025 and 45.9% in 2024, the second-highest value since 1910: the 2020s average 33% against 17% in the 2000s. |
Droughtin the index Is American soil drier than usual? | High | ↑ | Palmer index at −5.9 in September 2026, averaging −6.3 since January: drier than the 2012 drought (−5.1 in August) and at 1930s levels. The Drought Monitor puts 55.2% of the contiguous United States in drought on October 6th 2026 and 31.5% in severe drought, down from 59.4% in mid-September and the April peak of 63%; in 2019 the average was 10%. |
Floodsin the index How many floods trigger satellite mapping? | Moderate | ↓ | 25 flood activations over the twelve months to September 2026, against a record 44 in October 2024, 35 a year in 2024 and 34 in 2025, and 23 in 2019. The retreat from the peak reflects the absence of a season like 2024's in Central Europe and West Africa; the level remains above any year before 2023. |
Wildfirein the index How much land is burning? | Low/Moderate | → | 3.84 million hectares burned in the United States over the 52 weeks to early October 2026 on GWIS's count, up from the 3.2 million trough of October 2025 but far from the 9.4 million record of August 2021. NIFC counts 9.0 million acres since January 1st, nearly double the same period of 2025 and more than all of 2024. In the EU, 0.74 million hectares: the Iberian fires of August 2025 have left the window. |
Public constructionin the index How much does the public sector build, and how fast is it growing? | Moderate | ↓ | $548 billion at an annual rate in August 2026, a record, $104 billion of it on water, sewage and conservation — double 2019. But growth is only +2.5% on the year, against +19% on average in 2023 and +12% in 2024: the federal programmes of 2021 have finished ramping up. |
Raw data
Natural Catastrophe Losses
Overall losses (Munich Re, US$ bn)
224 B$
1 yr
↓ -96 B$
Insured losses (Munich Re, US$ bn)
108 B$
1 yr
↓ -32 B$
H1 2026 overall losses (US$ bn)
112 B$
H1 2026 insured losses (US$ bn)
44 B$
Source: Munich Re NatCatSERVICE (transcribed)
Economic damages by type of disaster, worldwide, US$ bn at current prices (EM-DAT)
Source: EM-DAT, CRED / UCLouvain, via Our World in Data
Notes
- What is happening
- $224 billion of overall losses in 2025, $108 billion of it insured, after 320 and 140 in 2024 — 368 and 147 once restated for inflation. The year was dominated by the Los Angeles fires: 53 billion, 40 insured, the costliest wildfire in history, which alone explains EM-DAT's record $54 billion of wildfire damages. Insured losses topped $100 billion for the sixth time, four of them in the last five years; before 2017, only once, in 2011. The first half of 2026 comes in at 112 billion, 44 insured, against 131 and 80 a year earlier.
- Why it matters
- Natural catastrophe losses are climate change measured in dollars, exposure included: they grow because events worsen and because there is more to destroy in their path. Munich Re publishes each January the overall and insured losses of the year just ended; the figures here are those announced at the time, in current dollars, because the inflation-restated series is not public. EM-DAT, the Louvain university database, adds damages by type of disaster since 1970.
- What's driving it
- More property in harm's way — coasts, wildland edges, floodplains — and storms, fires and floods that hit harder.
- What to watch
- The figure that matters is the ten-year average of insured losses, $107 billion at 2025 prices: that is the level reinsurers price. A year below 80 billion would bring premiums down; two years above 150 would make some states uninsurable. The second half carries the hurricanes, and the Atlantic season had produced only six tropical storms and no hurricane by September 21st, a first since 1914; what is left to decide whether 2026 stays a quiet year is the very strong El Niño forming for the winter, which Munich Re flags as the second-half risk.
Global Temperature Anomaly
Latest day, °C above 1850-1900
1.78
Monthly anomaly, °C above 1850-1900
1.70
1 mo
↓ -0.03
1 yr
↑ +0.16
12-month mean, °C above 1850-1900
1.49
1 mo
↑ +0.01
1 yr
↓ -0.02
Source: Copernicus C3S / ECMWF, ERA5
Notes
- What is happening
- 1.78°C above 1850-1900 on October 11th 2026 and 1.70°C on average in September, after 1.73°C in August — the joint-warmest month ever measured, level with July 2023 on Copernicus's count — and below the monthly anomaly record of 1.81°C of September 2023. On twelve months, 1.49°C, against a peak of 1.64°C in June 2024 at the end of El Niño, 1.54°C on average in 2025 and 1.21°C in 2019. The running mean dropped back below 1.5°C in spring 2026 after spending most of 2024 and 2025 above it.
- Why it matters
- The global temperature anomaly is the climate signal itself, without intermediary. Copernicus computes it every day from the ERA5 reanalysis — air temperature at two metres, averaged over the whole globe — and publishes it a day late. The reference is 1850-1900, the pre-industrial era of the Paris Agreement; the twelve-month mean smooths seasons and El Niño.
- What's driving it
- Greenhouse gases at a record, less cooling from shipping aerosols since 2020, and a new El Niño strengthening since the spring, after the one of 2023-2024.
- What to watch
- The threshold is 1.5°C on a long average — twenty years in the IPCC's definition, which the planet will reach around 2030 at the current pace of 0.2°C a decade. A twelve-month mean back above 1.6°C outside El Niño would mean warming is accelerating, as several teams have suspected since 2023. That El Niño is here: NOAA gives a better than 90% chance of a very strong event this winter, and the anomaly record usually comes the year after the peak — 2027 is the year to watch.
Extreme Weather Events
Activations, trailing 12 months
62
1 mo
↑ +4
1 yr
↑ +12
Activations this month
0
1 mo
↓ -5
1 yr
↓ -1
US Climate Extremes Index (% of area)
31.7%
1 yr
↓ -14.2%
Source: Copernicus EMS
NOAA Climate Extremes Index — the share of the contiguous US experiencing extremes of temperature, precipitation, drought or tropical systems, annual
Source: NOAA NCEI
Notes
- What is happening
- 62 activations over the twelve months to September 2026, against 42 a year in 2019, a record 75 in May 2023 — the year of the Canadian fires — and a dozen in 2014. The series doubled between 2016 and 2019, then settled between 45 and 60. NOAA's index covered 31.7% of the United States in 2025 and 45.9% in 2024, the second-highest value since 1910 after 2012; the 2020s average 33%, the 2010s 31%, the 2000s 17%.
- Why it matters
- Copernicus's emergency service maps disasters by satellite at the request of authorities: each activation is an event serious enough for a government to ask for help. Counting activations for storms, wildfires and landslides gives a global, immediate measure of weather extremes, with one bias: the service is used more than at its start in 2012. NOAA's extremes index, annual since 1910, gives the long measure for the United States.
- What's driving it
- A warmer atmosphere holds more water and dries soils faster: heavier storms and longer fire seasons from the same weather.
- What to watch
- The level to watch on activations is 70 over twelve months, reached only in 2023: above it, several continents are in crisis at once, not one season. For NOAA, a 2020s decade ending above 35% on average would say 2012 and 2024 are no longer exceptions but the regime.
Drought
Palmer Drought Severity Index (US)
-5.87
1 mo
↑ +0.25
1 yr
↓ -3.49
Area in drought, D1 or worse
53.7%
1 mo
↓ -4.9%
1 yr
↑ +8.7%
Area in severe drought, D2 or worse
30.3%
1 mo
↓ -2.9%
1 yr
↑ +6.8%
Source: NOAA NCEI
US Drought Monitor — share of the contiguous US in drought (D1-D4) and in severe drought or worse (D2-D4), weekly
Source: US Drought Monitor
Notes
- What is happening
- The Palmer index stands at −5.9 in September 2026, after −6.1 in August, and −6.3 on average since January: drier than 2012 (−5.1 in August) and than any year since the 1930s, when it touched −8.4. The Drought Monitor puts 55.2% of the contiguous United States in drought on October 6th and 31.5% in severe drought or worse, against 59.4% and 33.8% in mid-September and an April 21st peak of 63% — the record is 65% in September 2012. In 2019, the wettest year in the series, 10% of the country was in drought and the Palmer index at +4.7.
- Why it matters
- Drought is the slowest climate extreme and the costliest for agriculture and rivers. The Palmer index combines precipitation and temperature into a soil water balance for the contiguous United States, every month since 1895: negative is dry, −4 is extreme. The US Drought Monitor, weekly since 2000, says what share of the country is in drought by category, from D1 (moderate) to D4 (exceptional).
- What's driving it
- A dry winter over the Plains and the West, heat that evaporates what little fell, and an El Niño that is strengthening but waters the South only from late autumn.
- What to watch
- Drought is cyclical and the Palmer scale reads in duration: 2012 was erased in two winters. The threshold that would matter is an index below −4 for two summers running, which the series has seen only in the 1930s and 1950s. Next winter decides, and a very strong El Niño usually means a wet one for the South: a Drought Monitor still above 40% in March 2027 would put the drought into farm prices and the flows of the Colorado and the Mississippi.
Floods
Flood activations, trailing 12 months
25
1 mo
↓ -3
1 yr
↓ -2
Flood activations this month
1
1 mo
↑ +1
1 yr
↓ -4
Source: Copernicus EMS
Flood activations per month
Source: Copernicus EMS
Notes
- What is happening
- 25 flood activations over the twelve months to September 2026, against a record 44 in October 2024 — the year of the Valencia, Central European and Sahel floods — 35 a year on average in 2024, 34 in 2025 and 23 in 2019. The retreat from the peak is real but the level remains above any year before 2023. Floods account for 46% of the service's 1,061 activations since 2012, far ahead of fires.
- Why it matters
- Floods are the most frequent catastrophe and the least insured: in Munich Re's figures they weigh most in the gap between overall and insured losses. The measure is the number of Copernicus emergency-service activations for floods — every request for satellite mapping of a flood anywhere in the world since 2012 — per month and on trailing twelve months. It is the only global flood series available in real time without a key or a licence.
- What's driving it
- Heavier rainfall on hardened ground and expanding cities in floodplains, from Valencia to the Sahel.
- What to watch
- Activations depend on who asks: Europe asks more than Asia, and the service opened to the whole world in 2015. The threshold to watch is 40 over twelve months — exceeded only at the end of 2024 — which means a flood season on several continents at once. In EM-DAT, flood damages exceeded $45 billion in 2020, 2021, 2022 and 2024: four of the five largest years in the series.
Wildfire Activity
US burned area, 52 weeks (thousand ha)
3,832
1 mo
↓ -97
1 yr
↑ +631
EU burned area, 52 weeks (thousand ha)
740
1 mo
↑ +10
1 yr
↓ -226
US acres burned year to date (millions)
9.03
Same period last year (millions of acres)
4.75
Source: Copernicus GWIS (MODIS)
NIFC — acres burned in the United States per year, millions, since 1983
Source: NIFC
Notes
- What is happening
- 3.84 million hectares burned in the United States over the 52 weeks to early October 2026, up from the 3.2 million trough of October 2025 but far from the 9.4 million record of August 2021 and the 5.2 million average of 2019. NIFC counts 9.0 million acres since January 1st 2026, nearly double the 4.7 million of the same period of 2025, more than all of 2025 (5.1 million) and already above 2024's 8.9 million; the annual records are 10.1 million in 2015 and 2020. In the EU, 0.74 million hectares over 52 weeks, down from 0.94 million in mid-August: the Iberian fires of August 2025 have left the window.
- Why it matters
- Wildfire is the extreme that climate change amplifies most directly: hotter and drier means more fuel and a longer season. Copernicus's Global Wildfire Information System (GWIS) measures burned area by satellite every week, here on trailing 52 weeks for the United States and the European Union; NIFC, the US coordination centre, counts acres burned since 1983 and publishes the year's running total every day.
- What's driving it
- Fuel built up by the wet 2019 and 2023 seasons, now dried by the 2026 drought and heat.
- What to watch
- Burned area says little about cost: Los Angeles's 15,000 hectares cost $53 billion, 2021's 9 million hectares far less. The figure to watch is the share of wildfire in insured losses, up from 2% in 2019 to 37% in 2025 on Munich Re's count. A NIFC year above 10 million acres — there have been three, 2015, 2017 and 2020 — would need another million: September and early October added half a million, and October rarely brings more — 2026 ends above 2024, not above 10.
Public Construction Spending
Public construction, total
548 B$
1 mo
↑ +1 B$
1 yr
↑ +13 B$
Water, sewage and conservation
104 B$
1 mo
↑ +1 B$
1 yr
↑ +5 B$
Public construction, year on year
2.5%
1 mo
↑ +0.2%
1 yr
↓ -1.1%
Source: US Census Bureau, via FRED
Total public construction spending, change on a year earlier
Source: US Census Bureau, via FRED
Notes
- What is happening
- $548 billion at an annual rate in August 2026, a record, against 344 on average in 2019 and 509 in 2024. Water, sewage and conservation weigh 104 billion, double 2019 (53). But growth is only +2.5% on the year, after +19% on average in 2023, +12% in 2024 and +6% in 2025: the $550 billion of the 2021 infrastructure law has finished ramping up, and state budgets are absorbing the rise in construction costs.
- Why it matters
- Public construction is the measurable answer to climate: levees, water networks, treatment plants, roads rebuilt stronger. The Census measures it every month at an annual rate; the water, sewage and conservation-and-development categories — dams, levees, flood control — are the ones that build directly against water and weather. The year-on-year growth rate feeds the index, because the level trends up.
- What's driving it
- The 2021 infrastructure law's water and resilience money, now spent; state budgets absorbing higher construction costs.
- What to watch
- The figure to compare is insured losses: $100 billion a year of damage against $104 billion of construction for water and climate, before inflation. Public construction growth turning negative in 2027, as the federal credits expire, would say adaptation is being financed by the catastrophes themselves — after the fact, through FEMA — rather than before.
Methodology
1Sources
Copernicus (ERA5 temperature, EMS activations, GWIS burned area), NOAA (Palmer drought index), Census via FRED (public construction spending); Munich Re and EM-DAT losses charted, outside the index.
2Normalisation
Temperature as a monthly anomaly, activations and burned area as trailing sums, public construction as year-on-year growth; weekly burned area read at each month's start.
3Z-scores
Each constituent minus its mean since 2013, divided by its standard deviation; capped at ±3. The Palmer index enters inverted so that drier reads higher.
4Weighting
Six constituents, equal weights.
5Index
Average of the six z-scores × 20 points, rebased to 100 in March 2013. Updated monthly.
6Interpretation
Higher means more climate pressure. Temperature does not come back down; drought and fires are cyclical — the breakdown separates the two.