April 2009 compared to past Aprils
Compared to previous Aprils, April 2009 is near the middle of the distribution (10th lowest of 31 years). The linear trend indicates that for the month of April, ice extent is declining by 2.8% per decade, an average of 42,400 square kilometers (16,400 square miles) of ice per year.
Figure 3. Monthly April ice extent for 1979 to 2009 shows a decline of 2.8% per decade.
—Credit: National Snow and Ice Data Center
http://nsidc.org/images/arcticseaicenews/20090504_Figure3.png
Image is huge so I didn't hot link
2008 had more sea ice than 2007; why?
For details concerning why summer of 2008 shaped up differently than previous summers, please see our analysis archives and ongoing analysis updates.
A more general question might be, if sea ice is declining, how can it be that a single day or month decades ago could actually have had less ice than the same day or month in recent years? For more accurate results, scientists avoid comparing a historical single day or month (for example, May 1980) with a recent single day or month (for example, May 2008). Comparing longer trends and averages is more appropriate because natural variability, or natural shifts in the climate system, cause changes from one day or month to the next. Scientists remove the influence of this noise in a data record by gathering many points of data over a longer time period to understand the statistical significance of trends. This is true not just in studying sea ice, but also in many areas of scientific study.
As an analogy, consider statistics from sports. One game during a winning season when the home football team lost badly wouldn’t be indicative of their season as a whole. And comparing that one bad game years ago with a really good game this year, when the team managed to win 28-0 during a terrible losing season, wouldn’t be a fair comparison, either. However, plotting all of the games on a line graph would give an accurate indication of how the team did that year. And taking the scores and plotting them over several decades, would indicate whether the team has a significant trend over its history.