Arguments
Climate models are accurately predicting ocean and global warming
Posted on 27 July 2016 by John Abraham
For those of us who are concerned about global warming, two of the most critical questions we ask are, “how fast is the Earth warming?” and “how much will it warm in the future?”.
The first question can be answered in a number of ways. For instance, we can actually measure the rate of energy increase in the Earth’s system (primarily through measuring changing ocean temperatures). Alternatively, we can measure changes in the net inflow of heat at the top of the atmosphere using satellites. We can also measure the rate of sea-level rise to get an estimate of the warming rate.
Since much of sea-level rise is caused by thermal expansion of water, knowledge of the water-level rise allows us to deduce the warming rate. We can also use climate models (which are sophisticated computer calculations of the Earth’s climate) or our knowledge from Earth’s past (paleoclimatology).
Many studies use combinations of these study methods to attain estimates and typically the estimates are that the planet is warming at a rate of perhaps 0.5 to 1 Watt per square meter of Earth’s surface area. However, there is some discrepancy among the actual numbers.
So assuming we know how much heat is being accumulated by the Earth, how can we predict what the future climate will be? The main tool for this is climate models (although there are other independent ways we can study the future). With climate models, we can play “what-if scenarios” and input either current conditions or hypothetical conditions and watch the Earth’s climate evolve within the simulation.
Two incorrect but nevertheless consistent denial arguments are that the Earth isn’t warming and that climate models are inaccurate. A new study, published by Kevin Trenberth, Lijing Cheng, and others (I was also an author) answers these questions.
The study was just published in the journal Ocean Sciences; a draft of it is available here. In this study, we did a few new things. First, we presented a new estimate of ocean heating throughout its full depth (most studies only consider the top portion of the ocean). Second, we used a new technique to learn about ocean temperature changes in areas where there are very few measurements. Finally, we used a large group of computer models to predict warming rates, and we found excellent agreement between the predictions and the measurements.
According to the measurements, the Earth has gained 0.46 Watts per square meter between 1970 and 2005. Since, 1992 the rate is higher (0.75 Watts per square meter) and therefore shows an acceleration of the warming. To put this in perspective, this is the equivalent of 5,400,000,000,000 (or 5,400 billion) 60-watt light bulbs running continuously day and night. In my view, these numbers are the most accurate measurements of the rate at which the Earth is warming.
What about the next question – how did the models do? Amazingly well. From 1970 through 2005, the models on average showed a warming of 0.41 Watts per square meter and from 1992-2005 the models gave 0.77 Watts per meter squared. This means that since 1992, the models have been within 3 % of the measurements. In my mind, this agreement is the strongest vindication of the models ever found, and in fact, in our study we suggest that matches between climate models and ocean warming should be a major test of the models.
Despite these excellent results, scientists want to do better. During a conversation with Dr. Trenberth, he told me:
"Earth has gained 0.46 Watts per square meter between 1970 and 2005. Since, 1992 the rate is higher (0.75 Watts per square meter) and therefore shows an acceleration of the warming."
So would this mean Earth had a net loss of 0.29 Watts per square meter between 1970 and 1992? If so, I wonder which years experienced a significant loss. I would assume sometime in the 1970s.
@DAK4Blizzard:
As I calculate it, if in the 13 years between 1992 and 2005 the warming is 0.75W/m2 that would be a total warming of 13*0.75*(Sy*AE). For the 35 years between 1970 and 2005 you would have 35*0.46*(Sy*AE). For the 22 years between 1970 and 1992 you would have (35*0.46-13*0.75)*(Sy*AE) or an average warming of (35*0.46-13*0.75)/22=0.29W/m2.
So the earth was warming even then, though at a lower rate.
In the above Sy is the number of seconds in a year and AE the area of the earth (the cancel out).
Only somewhat related to the original post, but are there any prospects in the near/medium term of making a direct measurement of the energy imbalance of the earth? In principle there is this nice clean interface at the top of the atmosphere where there is only shortwave radiation coming in from the sun, and only reflected shortwave and emitted thermal radiation from the earth. The imbalance should be about 0.75/340=0.2% so the precision of the two measurements must be better than that. Is this impossible? Or is it just that ignorance is strength (profit)?
DAK4Blizzard @1, ATTP has a post on the study which is more informative. In particular it shows a graph of observations of full depth ocean heat content compared to the CMIP5 models:
As you can see, OHC increased on average, over the period from 1970-1992, although with some periods of decline within that interval.
ATTP also produces a table from the paper showing the total increase in OHC over the intervals:
The change in OHC from 1970-1991 is, evidently, 28.3-13.5 = 14.8 (X 10^22) Joules. We thus have an average gain over the intervals of:
1970-2005 -— 0.49 W/m^2
1970-1991 -— 0.42 W/m^2
1992-2005 -— 0.6 W/m^2
To confuse things further ATTP quotes the paper as indicating changes in OHC of 0.68 x 10^22 Joules/Year from 1970-2005 and 1.22 x10^22 Joules/Year for 1992-2005, which resolve to 0.42 W/m^2 and 0.76 W/m^2 respectively. The later is 0.01 W/m^2 higher than the value given in the article above, but the former is 0.04 W/m^2 smaller. The annual values from the paper likely differ from the values quoted above because, firstly, I assumed a year of 365.25 days, whereas the actual average year lenght will have varied slightly depending on just how many leap years fall within a given interval. (Particularly a factor given that 2000 was not a leap year). Secondly, the annual values will differ because they will be based on the trendline, whose endpoint need not coincide with the terminal points of the actual values (and will not for the 1970-2005 interval, given the large inflection).
I do not know why the article above calculated a different W/m^2 value from the annual Joules/Year values cited in the paper. It may be because they took into account the actual number of leap years involved.