Multidecadal Variability and Trends in Upper Ocean Temperature

The evolution of thermal changes in the upper ocean is important to quantify for climate studies. However, sparse historical sampling of subsurface temperatures, coupled with regional decadal variability, complicates estimating the magnitude of multidecadal variability and temperature trends. The research presented here examines these issues through a suite of observational analyses based on temperature data in the World Ocean Database series.;Maps of regional 50-year and 20-year gridded subsurface temperature trend magnitudes are presented which exhibit large-scale regional patterns. Most regional trends do not pass a 90% CL test. There are coherent regions of significant cooling and warming (up to +/-1--3°C), and most regions undergo a change in 20-year trend sign depending on the period analyzed; this includes gridded SST data. Short-term trends may not be representative of long-term trends regionally. These trend results are robust to variations in grid size and data processing. Different analytical methods are shown to produce different regional results for trends that are biased towards more warming or cooling when globally averaged. The differences usually occur in the poorly sampled regions of the ocean. Strong regional multidecadal trend patterns are still present in such different analyses. Further research indicates that regional multidecadal variability is a robust result even when an important instrument bias correction is applied to the data. Large regional changes also contribute substantially to the global heat trend magnitude. Closer examination on sub-basin scales finds instances of large heat storage changes which occur over short time periods, often 1 to 3 years. These changes can last up to 10 or more years, and have a much larger impact on global multidecadal trends than do interannual changes that do not persist. The character of these large thermal shifts is explored, and estimates of their contribution to global ocean heat content are presented.;The intensity of regional multidecadal variability is an important component of global ocean temperature trends, and deserves further study. It is hoped that continuing observational operations such as the Argo program will improve estimates of multidecadal ocean variability in the coming decades.