Geoscience Reference
In-Depth Information
Fig. 2: Trends ( o C/100 yr) in area-averaged annual-mean Indian Ocean SST (north of
30 o S) are shown as a function of starting year, with all trends ending in 2010. The
black curve shows trends from observations (HadCRUT4), where observational
uncertainty is depicted as a range showing the 5 th to 95 th percentile ranges of trends
obtained using the 100-member HadCRUT4 ensemble. The red curve shows ensemble
mean trends from the All Forcing runs of seven CMIP5 models. The dark blue curve
shows ensemble mean trends for Natural Forcing Only runs extending to 2010. The
different models are weighted equally for the multi-model ensemble mean trends,
regardless of the number of ensemble members they contributed. The pink shading
shows the 5 th to 95 th percentile range of the distribution of trends obtained by combining
random samples from each of the seven CMIP5 model control runs with the
corresponding model's ensemble-mean forced trend (All Forcing runs) to create a total
multi-model distribution of trends that reflects uncertainty in both the forced response
(i.e., the spread among the different model ensemble means) and the influence of internal
climate variability (as estimated from the model control runs). The blue-shaded region
shows the same, but for Natural Forcing Only runs. Violet shading indicates where the
pink- and blue-shaded regions overlap. The seven-CMIP5-model subset used here
includes: CanESM2, CNRM-CM5, CSIRO-Mk3-6-0, FGOALS-g2, HadGEM2-ES,
IPSL-CM5A-LR, and NorESM1-M.
assessment of the observational uncertainty, which is not trivial, but is much
smaller than the uncertainty due to the spread of model results (forcing response
and internal variability). When the observed trend (black-shaded region) lies
outside of the blue shaded region, the trend is termed “detectable” (i.e.,
significantly different from natural forcing runs). Trends are detectable for
Search WWH ::




Custom Search