By Michon Scott
Is Earth warming? Yes. Exactly how much is Earth warming? That question is harder to answer.
The difficulty in answering that question is partly due to different parts of the planet warming at different rates. The difficulty is also caused by a dearth of historical observations in many regions. Historically, humans have only kept routine temperature records in places where we have lived and worked. Earth’s global ocean, where humans have mostly been passing through, covers roughly 70 percent of our planet. Satellite sensors designed to measure ocean temperatures, and float arrays designed to take in situ measurements have filled ocean data gaps—but only for as long as those tools have been available. Identifying how much Earth has warmed over longer periods involves climate models and well-informed estimates.
A new study focuses on improving global temperature data sets in light of uneven warming across the globe. To fill gaps in historical climate records, the study relies in part on sea ice data from the National Oceanic and Atmospheric Administration program at the National Snow and Ice Data Center (NOAA@NSIDC).
Why warming is uneven
Earth does not heat up at the same pace everywhere because our planet does not have a uniform surface. Nor does every part of Earth receive even amounts of sunlight, experience all the same natural cycles, or feel the same effects of human-caused greenhouse gas emissions.
Earth’s land surfaces range from mountain peaks to desert basins, from lush forests to concrete jungles. Even broader contrasts occur between land and ocean. Earth’s ocean has a high heat capacity; it must absorb tremendous amounts of energy before it warms, in contrast to most land surfaces. The ocean also cools more slowly than landmasses. Land areas can frequently experience broad temperature ranges on a monthly or even daily basis. Ocean temperatures vary, too, especially by region, but ocean temperatures vary on a smaller scale than land temperatures.
Adding to differences in Earth surfaces are differences in the amount sunlight they receive. Locations along the equator receive about 12 hours of sunlight and darkness each day while polar days and polar nights last months. The greater differences in sunlight duration drive greater seasonality nearer the poles. Meanwhile, sunlight reaching the equator is direct, whereas sunlight reaching the polar regions is low angled and dispersed over a bigger region, even in summer.
Natural cycles add another layer of complexity. These cycles, such as El Niño-Southern Oscillation, often have far-reaching effects, but the effects differ by location. El Niño tends to boost global temperatures, and La Niña tends to depress them, but both phases of the climate cycle have regional exceptions. Either phase of the cycle brings above- or below-average temperatures to different parts of the continental United States, temperatures that might seem out of sync with long-term trends.
So, day-to-day temperatures result from complex interactions between land or water surface, sunlight, natural cycles, and human-caused greenhouse gas emissions. Many of the same factors affecting daily temperatures influence temperature trends over time.
One of the clearest examples of uneven warming over time occurs in the Arctic, which is warming roughly three times as fast as the rest of the globe. Scientists refer to this faster-than-global-average warming as Arctic amplification. Multiple factors contribute to the amplification, such as the transfer of warm air masses from lower latitudes, and the region’s declining ability to reflect incoming sunlight as snow and sea ice decline. Sea ice decline is a self-reinforcing process.
In the new study on Earth’s non-uniform warming, an understanding of sea ice plays a key role in the Arctic, around Antarctica, and across the globe.
NOAA@NSIDC data leveraged in new study
The presence or absence of sea ice has a significant impact on local temperatures. Sea ice acts as a gatekeeper between the ocean and the atmosphere, insulating ocean waters from frigid air overhead, and regulating transfers of salinity, moisture, and heat.
The polar regions present researchers with a challenge. Sea ice and temperature changes are clear in the continuous satellite record dating back to late October 1978, and they are clear in buoy observations dating back to the start of the twenty-first century. Further back in time, however, the polar regions were sparsely observed. Cold War superpowers took observations, but those observations did not encompass the entire Arctic Ocean. To fill in the gaps, the 2024 study turned to two products archived and distributed by NOAA@NSIDC.
One product the study used is the NOAA/NSIDC Climate Data Record of Passive Microwave Sea Ice Concentration, Version 4. This data set is an official NOAA climate data record, which means it meets NOAA’s criteria for scientific robustness and trustworthiness. The data set provides a consistent time series, both daily and monthly values, of sea ice concentrations dating back to October 25, 1978. Observations span the polar regions and most of the mid-latitudes.
The other product the study used is the Gridded Monthly Sea Ice Extent and Concentration, 1850 Onward, Version 2. As its name implies, this data set provides observations dating back to the mid-nineteenth century. The data extend through 2017. Pre-satellite observations come from compilations by naval oceanographers, national ice service analyses, and historical ship observations. Observations span the Northern Hemisphere as far south as 30 degrees North.
Bruce Calvert, the 2024 study author, pointed out the value of sea ice records extending so far back in time. He explained that a failure to adequately account for sea ice changes—and the consequent changes in the broad differences between ice-covered and ice-free polar oceans—could underestimate rates of Arctic warming.
Findings and areas for refinement
In addition to the new paper, Calvert compiled a global instrumental temperature data set. He designed the data set to account for non-uniform warming by allowing warming trends to vary by location and season, and by incorporating differences between ice-covered and ice-free ocean.
As sunlight-reflecting sea ice gives way to sunlight-soaking ocean water, temperatures tend to climb rapidly. Areas losing sea ice cover count among the planet’s fast-warming places. A key motivation for the non-uniform warming study was Calvert’s hypothesis that climate models might underestimate long-term warming in the Antarctic and, especially, the Arctic.
The study compared temperatures from the late nineteenth century (1850 to 1900) to those of 2023. Synthesizing global climate models and historical data sets, including NOAA@NSIDC’s sea ice data, Calvert found that changes in sea ice concentrations over time had a significant impact on temperature trends.
Calvert combined sea ice concentration changes with other adjustments to accommodate location and season. He estimated the global mean surface temperature change between the late nineteenth century and 2023 at 1.548 degrees Celsius (2.786 degrees Fahrenheit), with likely values ranging from 1.449 degrees Celsius to 1.635 degrees Celsius (2.608 degrees Fahrenheit to 2.943 degrees Fahrenheit). Accounting for sea ice increased the estimated global surface temperature change between the late nineteenth century and 2023 by 0.079 degrees Celsius (0.142 degrees Fahrenheit).
The study showed that changes in sea ice can have global implications. It also demonstrated the value of sea ice data archived by NOAA@NSIDC.
Access data through NOAA@NSIDC
NSIDC archives a variety of data sets on behalf of NOAA. In addition to archiving data, the NOAA@NSIDC provides data users with multiple methods of support: documentation, help articles, data tools, training, and on-demand support. Learn more about NOAA@NSIDC services.
NOAA@NSIDC data highlighted in this article include:
Meier, W.N., Fetterer, F., Windnagel, A.K. & Stewart, J.S. (2021). NOAA/NSIDC Climate Data Record of Passive Microwave Sea Ice Concentration. (G02202, Version 4). [Data Set]. Boulder, Colorado USA. National Snow and Ice Data Center. https://doi.org/10.7265/efmz-2t65. [describe subset used if applicable]. Date Accessed 01-27-2025.
Walsh, J.E., Chapman, W.L., Fetterer, F. & Stewart, J.S. (2019). Gridded Monthly Sea Ice Extent and Concentration, 1850 Onward. (G10010, Version 2). [Data Set]. Boulder, Colorado USA. National Snow and Ice Data Center. https://doi.org/10.7265/jj4s-tq79. [describe subset used if applicable]. Date Accessed 01-27-2025.
Reference
Calvert, Bruce T.T. 2024. Improving global temperature datasets to better account for non-uniform warming. Quarterly Journal of the Royal Meteorological Society. doi:10.1002/qj.4791.