Ask a Scientist
  • Sea ice

Why use the 1981 to 2010 average for sea ice?

NSIDC scientists use the 1981 to 2010 average of sea ice concentration and extent to have a consistent basis for comparison to today’s fluctuating conditions. The 30 years observed in this baseline provide enough data to even out short-term variability.

NSIDC could use the newer baseline period of 1991 to 2020 adopted by the National Oceanic and Atmospheric Administration (NOAA) and other parties. But NSIDC scientists have decided to continue using the 1981 to 2010 baseline because it is better suited to sea ice conditions.

Why use a baseline at all?

To show how current sea ice conditions compare to past conditions, scientists could simply compare current conditions with those of the same day or month from a previous year. While such comparisons sometimes make sense, for instance, “Here is what Arctic sea ice looked like when we graduated from college compared to now,” the same approach can also be used to cherry-pick data. Year-to-year variability is significant for sea ice, even against a backdrop of long-term decline. So, depending on the years chosen, a comparison can look striking or negligible.

Sea ice extent comparison, September 1979, 1980, 2005
In September 2005, Arctic sea ice reached what was, at the time, the lowest extent on record. But depending on whether the September extent was compared to 1980 or 1979, the new record low looked more dramatic or less dramatic. The median extent for 1981 to 2010 (magenta line) provides a more reliable yardstick of how current conditions compare to the long term. — Credit: NSIDC Sea Ice Index

By instead comparing daily or monthly extents to decades-long averages for that time of year, scientists use data more representative of observed conditions. But when choosing the longer period for comparison, scientists have multiple options, some better than others.

Why use a fixed-period baseline?

NSIDC could recalculate the baseline period each year, to incorporate the entire satellite record through the most recent full year. But using a sliding or growing baseline would prevent a meaningful comparison of different years. For example, imagine if NSIDC used a baseline of 1979 through 2010 for evaluating the year 2011, and then used a baseline of 1979 through 2011 for evaluating the year 2012. With that approach, 2011 would be compared to a 32-year baseline, and 2012 would be compared to a 33-year baseline. To borrow the common phrase, NSIDC would be comparing apples to oranges.

A fixed-period baseline allows any two or more years to be compared to the same long-term average and, by extension, compared to each other in terms of how much they differed from that average.

Why 30 years?

A fixed-period baseline has advantages over a changing baseline, but the fixed period could be any length of time—anywhere from 5 to 50 to 100 years. A five-year baseline would likely be too short to reflect the range of conditions likely to occur whereas a 50-year baseline would reflect a wide range of conditions nicely. But the longer a baseline period is, the more observations are needed to fill it. The continuous satellite record of sea ice data only goes back to 1978. Right now, 50 years of those observations are not yet available.

The standard meteorological reference period is 30 years. Both NOAA and the World Meteorological Organization (WMO) use a 30-year baseline for weather and climate, operating on the principle that 30 years of data provide a reliable reference period for predicting weather and making climate-related decisions. For instance, when planning a trip, you want to dress appropriately for the conditions at your destination. A 30-year average for that place at that time of year cannot give you a precise forecast of the temperature and precipitation you will experience, but it can tell you what conditions are typical. Thirty-year averages can also help municipalities plan for operations such as construction, road maintenance, or fuel deliveries to remote locations.

Prior to July 2013, NSIDC used the baseline period of 1979 to 2000. Starting July 1, 2013, NSIDC’s Arctic Sea Ice News & Analysis and Sea Ice Index both moved to the baseline period of 1981 to 2010. NSIDC made the move because 30 years of satellite data had finally become available, and because it coincided with NOAA’s baseline period.

Since NSIDC’s adoption of the 1981 to 2010 baseline, NOAA has moved to the 1991 to 2020 baseline. So has the Arctic Report Card. Why has NSIDC has chosen to stick with the 1981 to 2010 baseline?

Why not use the newer 1991 to 2020 baseline?

NOAA and the WMO update baseline periods every 10 years. NOAA and WMO provide forecasts of weather (the short-term state of the atmosphere), and meaningful weather forecasts depend on timely data about climate (the weather of a particular place averaged over time). Using more recent climatology enables the WMO and NOAA to make more accurate forecasts. As NOAA states, “US Climate Normals are designed—and best-suited for—better understanding what is happening today.”

NSIDC’s baseline for satellite observations of sea ice serves a different purpose, however. NSIDC observations of sea ice concentration and extent show how sea ice is changing over time and by how much. NSIDC chose the 1981 to 2010 baseline period because it included the first three complete decades with an unbroken record of satellite data.

Over the course of the entire satellite record, Antarctic sea ice has shown no significant long-term trend but has instead shown tremendous year-to-year variability. In the years since 2000, however, Arctic sea ice has experienced multiple record-low extents. These record lows have dragged down the long-term Arctic average. In other words, since the turn of the twenty-first century, Arctic sea ice behavior has been less and less “normal,” compared to what was historically observed. The 1981 to 2010 is the best “normal” that sea ice has to offer for comparison.

Since Arctic sea ice has shown such significant negative trends since the year 2000, the 1981 to 2010 period included multiple years of unusually low extent. Moving to the 1991 to 2020 baseline would only exacerbate that problem. For this reason, NSIDC is sticking to the 1981 to 2010 baseline at present.

You can choose your own baseline

If you would like to compare sea ice extent to the 1991 to 2020 baseline—or to any other baseline, for that matter—you can use NSIDC’s Sea Ice Analysis tool to demarcate the baseline of your choice. Choose the option for the Extent Anomaly Graph to choose your own baseline.

Screenshots of 1981-2010 and 1991-2020 baselines for September
NSIDC’s Sea Ice Analysis tool lets users pick the long-term average period for the desired baseline. The top image uses the 1981 to 2010 baseline; the bottom image uses the 1991 to 2020 baseline. A close look at these images shows the later baseline period reduces the percentage of loss, though both baselines show a clear downward trend. — Credit: NSIDC Sea Ice Analysis

References

NSIDC. Sea Ice Index Baseline Period Change. https://nsidc.org/sites/default/files/sea-ice-index-baseline-period-change.pdf

NSIDC. Arctic Sea Ice News & Analysis: Frequently Asked Questions. https://nsidc.org/arcticseaicenews/faq/#1979average

NOAA. 2021. NOAA Delivers New US Climate Normals. https://www.ncei.noaa.gov/news/noaa-delivers-new-us-climate-normals

NSIDC. Arctic weather and climate. https://nsidc.org/learn/parts-cryosphere/arctic-weather-and-climate