Arctic Annual Surface Air Temperature
Data Source: NASA GISTEMP V4 Analysis| Linear Trend Analysis
This chart is interactive. Rollover each data point to see details about the chart data. Toggle on or off graphic elements by clicking on them. Click the three lines to the top left of the graph to download, print or share the chart.
Rate of Change
- >.18°C
Annual, Autumn and Winter Arctic temperature increased per decade from 1900-2025
Key Takeaway
In addition to the warmest water year on record, autumn 2024 was the warmest since 1900 at 2.28°C above the 1991–2020 mean, while the winter and summer 2025 temperatures were the second and third warmest on record at 2.41°C and 0.83°C above the 1991–2020 average.
Methods and Data
- The NASA Goddard Institute for Space Studies surface temperature analysis version 4 (GISTEMP v4) is used to describe long-term Arctic (60–90°N) and Global (90°S–90°N) surface air temperatures since 1900 (Fig. 1). GISTEMP v4 air temperatures over lands are obtained from the NOAA Global Historical Climatology Network version 4 (GHCN v4) dataset and ocean surface temperatures are taken from the NOAA Extended Reconstructed Sea Surface Temperature version 5 (ERSST v5) dataset. The GISTEMP product is described in detail in Hansen et al. (2010) and Lenssen et al. (2019).
- Decadal trends for the full period at annual and seasonal timesteps are calculated using linear regression and statistical significance is determined where p⩽0.05.
Precipitation
Data Source: ERA5 Reanalysis (ECMWF)
This chart is interactive. Rollover each data point to see details about the chart data. Toggle on or off graphic elements by clicking on them. Click the three lines to the top left of the graph to download, print or share the chart.
Rate of Change
- 0.75%
+0.75% per decade
Key Takeaway
Despite a dry summer, precipitation averaged across the Arctic for water year 2024/25 (October 2024–September 2025) was at a record high, and precipitation totals for the winter, spring, and autumn seasons were each in the top five highest, reinforcing positive trends in Arctic precipitation over the 1950–2025 period. An intensifying water cycle with more extreme precipitation events threatens to further disrupt Arctic ecosystems, infrastructure, and communities.
Methods and Data
- Analyses were performed for the period 1951-August 2025 on monthly gridded precipitation fields from the ERA5 reanalysis (https://www.ecmwf.int/en/forecasts/dataset/ecmwf-reanalysis-v5) and daily precipitation records from gauging station (land only) compiled by the Global Precipitation Climatology Centre (GPCC) (https://opendata.dwd.de/climate_environment/GPCC/html/fulldata-daily_v2022_doi_download.html).
- Spatial field of seasonal precipitation anomalies for the 2024/2025 water year relative to a 1991-2020 baseline were compiled and plotted from ERA5 data for the region poleward of 60 degrees north
- Plots were compiled for the region poleward of 60 degrees north of ranks of maximum 5-day precipitation for each season during the 2024/25 water year relative to the period 1951–2023
- Precipitation from each product was averaged by water year season and year from the 1951/1952 to the 2024/2025 water year (weighted by latitude) and linear trends were computed and plotted for the region poleward of 60 degrees north, expressed as a percentage of values relative to a 1991-2020 baseline
- A plot for the region poleward to 60 degrees north was compiled of the spatial field of precipitation trends from ERA5 from 1950-2025 in units of centimeters per year
Terrestrial Snow Cover
Data Source: NOAA Climate Data Record of Northern Hemisphere Snow Cover Extent (SCE)
This chart is interactive. Rollover each data point to see details about the chart data. Toggle on or off graphic elements by clicking on them. Click the three lines to the top left of the graph to download, print or share the chart.
Rate of Change
- −8.7%
June snow cover extent (SCE) declined at −8.7% per decade; May SCE declined at −2.5% per decade.
Key Takeaway
For the Arctic as a whole, May SCE has declined nearly 15% since 1967 while June SCE has declined more than 50% since 1967.
Methods and Data
- Compute trends from May monthly snow cover extent (theil-sen) and compute fitted trend line (yfit)
- Convert to percent change by first dividing all snow cover extent values and all yfit values by yfit value in 1967 and then multiplying all values by 100. This depicts changes relative to values near the start of period (yfit in 1967 is 100%).
- Repeat for June monthly snow cover extent
Greenland Ice Sheet
Data Source: GRACE Satellite Measurements
This chart is interactive. Rollover each data point to see details about the chart data. Toggle on or off graphic elements by clicking on them. Click the three lines to the top left of the graph to download, print or share the chart.
Rate of Change
- -219Gt/yr
– 219 Gt/yr (recent rate)
Key Takeaway
Ongoing ice mass loss with regional variations.
Methods and Data
The GRACE (Gravity Recovery and Climate Experiment, 2002–2017) and GRACE-FO (Follow On, 2018–present) satellite missions detect gravity anomalies to measure changes in total ice mass (GRACE/GRACE-FO Level-2: JPL RL06.1 doi:10.5067/GFL20-MJ061; Technical Notes 13 & 14: https://podaac.jpl.nasa.gov/gravity/gracefo-documentation). We apply a regional averaging kernel (Wahr et al., 1998) to the GRACE/GRACE-FO Level-2 products that is consistent with the JPL and GSFC mascon solutions (Watkins et al., 2015; Loomis et al., 2019). The GRACE/GRACE-FO source data include peripheral glaciers and ice masses that are not part of the Greenland Ice Sheet. We scale these numbers by 0.84 to approximate changes on the ice sheet only (Colgan et al. 2015).
Weather data are obtained from 20 Danish Meteorological Institute (DMI) land-based weather stations with records starting from 1784 (Nuuk), 11 Mittarfeqarfiit stations, and Summit Station (from DMI over 1991–2019 and provided by NOAA GEOSummit since 2019). Temperature and surface ablation measurements come from ten automatic weather station transects from the Programme for Monitoring of the Greenland Ice Sheet (PROMICE) at the Geological Survey of Greenland and Denmark (GEUS), following Van As et al. (2016).
Surface melt duration and extent are derived from daily Special Sensor Microwave Imager/Sounder (SSMIS) 37 GHz horizontally polarized passive microwave radiometer satellite data (Mote 2007). This detects the presence of melt, but not the magnitude of it.
PROMICE combines ice thickness estimates with ice velocity measurements to approximate solid-ice discharge over Greenland (Mankoff et al. 2020). The approximation arises in the areas between the measurement sites and the downstream calving fronts, as the product does not measure accumulation or ablation in these areas, nor does it account for short-term ice front advance or retreat.
Changes in ice surface elevation measured by ICESat-2 reflect ice mass gain or loss as well as changes in density, snow accumulation, and melt. ICESat-2 mass-difference estimates were calculated by correcting ICESat-2 elevation measurements (Smith et al. 2023) for these anomalies (Medley et al. 2022) following the processing strategy for ICESat-2 level-3B products (Smith, 2023). The largest error comes from the density, snow accumulation, and melt anomalies, together estimated at 14% of the value (Medley et al. 2022).
Sea Ice Extent
Data Source: | Linear Trend Analysis
This chart is interactive. Rollover each data point to see details about the chart data. Toggle on or off graphic elements by clicking on them. Click the three lines to the top left of the graph to download, print or share the chart.
Rates of Change
- 38,000
(March) Kilometers squared (km2/yr) per year
- 76,100
(September) Kilometers squared (km2/yr) per year
Key Takeaway
A record low extent in March 2025 punctuates a continuing decline in both seasonal minimum and maximum extents.
Methods and Data
Sea ice extent values are from the NSIDC Sea Ice Index (Fetterer et al. 2025), based on passive microwave derived sea ice concentrations from the NASA Team algorithm (Cavalieri et al. 1996), though other high-quality products exist (e.g. Lavergne et al. 2019). For 2025, the algorithm was adapted for data from the JAXA Advanced Microwave Scanning Radiometer 2 (AMSR2) to create a consistent NASA Team product (Stewart et al. 2025).
Sea Surface Temperature
Data Source: Satellite Observations (NOAA OISST V2.1)
This chart is interactive. Rollover each data point to see details about the chart data. Toggle on or off graphic elements by clicking on them. Click the three lines to the top left of the graph to download, print or share the chart.
Latest (August 2025)
- 0.82°C
+0.82°C above the 1991–2020 August mean.
Key Takeaway
– In the marginal seas of the Arctic Ocean’s Atlantic sector, August 2025 mean sea surface temperatures (SSTs) were as much as ~7° C warmer than 1991–2020 August mean values.
– Anomalously cool August 2025 SSTs (~1–2° C cooler) were observed in parts of the marginal seas of the Arctic Ocean’s Pacific sector.
-August mean SSTs show warming trends for 1982–2025 in almost all Arctic Ocean regions that are ice-free in August, with mean SST increases of ~0.3° C per decade in the region north of 65° N.
-Warming Arctic SSTs alter local ecosystems and accelerate sea-ice loss, with wide-ranging global climate and societal consequences.
Methods and Data
The SST data are from the 0.25° × 0.25° NOAA Optimum Interpolation Sea Surface Temperature (OISST) Version 2.1 product. The datafile “sst.mon.mean.nc” (comprising monthly means from the daily data) was retrieved from https://downloads.psl.noaa.gov/Datasets/noaa.oisst.v2.highres/ (accessed 3 September 2025).
Ocean Primary Productivity
Data Source: MODIS Aqua
This chart is interactive. Rollover each data point to see details about the chart data. Toggle on or off graphic elements by clicking on them. Click the three lines to the top left of the graph to download, print or share the chart.
Rate of Change
- 21.6
+21.6 Tg C/yr Annual rate of change (2003–2025)
Key Takeaway
Satellite estimates of ocean primary productivity (i.e., the rate at which marine algae transform dissolved inorganic carbon into organic material) show higher values for 2025 (relative to the 2003–22 mean) for eight of nine regions assessed across the Arctic. All regions, except for the Amerasian Arctic (the combined Chukchi Sea, Beaufort Sea, and Canadian Archipelago), continue to exhibit positive trends in ocean primary productivity during 2003–25, with the largest percent changes in the Eurasian Arctic, Barents Sea, and Hudson Bay.
Methods and Data
Measurements of the algal pigment chlorophyll (specifically, chlorophyll-a) serve as a proxy for algal biomass present in the ocean as well as overall plant health. The complete, updated MODIS-Aqua satellite record of chlorophyll-a concentrations within northern polar waters for the years 2003–25 serves as a time-series against which individual years can be compared. Satellite-based chlorophyll-a data across the pan-Arctic region were derived using the MODIS-Aqua Reprocessing 2022.0.2 (November 2024) chlor_a algorithm. For this report, we show mean monthly chlorophyll-a concentrations calculated as a percentage of the 2003–22 average. This same reference period (2003–22) has been utilized for the last three consecutive Arctic Ocean Primary Productivity Arctic Report Card essays (i.e., Frey et al. 2023b; Frey et al. 2024; this essay) ever since the MODIS-Aqua satellite record accrued 20 years of data. Satellite-based sea ice concentrations were derived from the Special Sensor Microwave/Imager (SSM/I) and Special Sensor Microwave Imager/Sounder (SSMIS) passive microwave instruments, calculated using the Goddard Bootstrap (SB2) algorithm (Comiso et al. 2017). Primary productivity data were derived using chlorophyll-a concentrations from MODIS-Aqua data (Reprocessing 2022.0.2, chlor_a algorithm), the NOAA 1/4° daily Optimum Interpolation Sea Surface Temperature dataset (or daily OISST), incident solar irradiance, mixed layer depths, and additional parameters. Primary productivity values were calculated based on the Vertically Generalized Production Model (VGPM) algorithm described by Behrenfeld and Falkowski (1997) as applied by Frey et al. (2023a). We included only pixels with less than 10% sea ice concentration, balancing ice contamination concerns against capturing the productive sea ice-edge. We define annual productivity as productivity over the March-September period. The 2025 annual primary productivity percent of average (compared to 2003–22) was calculated the same way as for chlorophyll-a, as described above. Spatial trends of primary productivity (Fig. 3b) were calculated using a Theil-Sen median trend estimator, and regional (and total Arctic) linear trends/percent change (Table 1, Figs. 4 and 5) were calculated through ordinary least squares regression. The statistical significance of all trends (p<0.05) was determined using the Mann-Kendall trend test. The MODIS-Aqua Reprocessing 2022.0.2 that took place in November 2024 includes revised data from 2022–present in response to satellite orbital shifts and resulting declines in data accuracy. As such, values and trends shown in our time series analyses this year (e.g., Table 1, Figs. 1, 3, 4, and 5) are updated from previous Arctic Report Card essays based on these newly revised data for 2022 onwards.
Importantly, our estimates exclude sea ice algae and under-ice phytoplankton blooms, which can be significant (Ardyna et al. 2020). Furthermore, it is known that satellite observations can underestimate production under stratified conditions when a deep chlorophyll maximum is present. The variable distribution of sediments and CDOM (owing to riverine delivery, coastal erosion, and sea ice dynamics) can also affect the accuracy of satellite-based estimations of chlorophyll-a and primary productivity in Arctic waters (Lewis and Arrigo 2020; Zoffoli et al. 2025). As such, in-situ observations continue to provide important overall context for changes to and drivers of primary productivity across Arctic marine ecosystems. This is particularly relevant given the vulnerabilities of satellite systems to finite temporal extents where in-situ observations can help to accurately dovetail data from different satellite platforms.
Tundra Greenness
Data Source: | Ordinary Least Squares Regression
This chart is interactive. Rollover each data point to see details about the chart data. Toggle on or off graphic elements by clicking on them. Click the three lines to the top left of the graph to download, print or share the chart.
Rate of Change
- .013
NDVI units/decade (MODIS, 2000-2024) and +0.016 NDVI units/decade (GIMMS, 1982-2023)
Key Takeaway
- Satellites provide foundational information regarding the productivity, health, and dynamics of Earth’s ecosystems over time.
- The “greening of the Arctic,” first reported in the late 1990s as an increase in the productivity and abundance of tundra vegetation due to rapid warming and sea-ice decline, is an ongoing phenomenon evident in all available long-term satellite records.
- Tundra greening signifies far-reaching changes to Arctic landscapes, wildlife habitats, biodiversity, permafrost conditions, and the livelihood of Arctic people, with implications for global climate and the carbon cycle.
Methods and Data
- The circumpolar mean MaxNDVI is computed for each year from 4 source datasets (see below) and trends are calculated as the change per decade using ordinary least squares regression. Note that we display these trends spatially in the circumpolar map figures, but do not quantitatively present the circumpolar average temporal trends of the regression lines in the brief essay.
- GIMMS 3g+
- MODIS
- Landsat Collection 2
- VIIRS