Global Science Report is a weekly feature from the Center for the Study of Science, where we highlight one or two important new items in the scientific literature or the popular media. For broader and more technical perspectives, consult our monthly “Current Wisdom.”
When people think about the weather, two variables are first to come to mind—temperature and precipitation. Unless it’s sunny when it’s not supposed to be (or vice-versa), near-term temperature forecasts tend to be pretty good. What messes up your day is when it rains when it’s not supposed to, and what really screws things up is when there is a significant unforecast snow, or a lot more (or less) than there was supposed to be. If whatever oracle you consistently consult, like The Weather Channel or Channel 9, consistently blows the precipitation forecast, you’ll soon be looking elsewhere for your forecast, and if changing forecasters doesn’t help, you’re going to sour on the whole weather forecasting business
Climate forecasts made by climate models running under scenarios of increasing human emissions of greenhouse gases are blowing both their temperature and precipitation prognostications. They tend to predict far more warming to be taking place than is actually occurring, and when it comes to precipitation, the projections are all over the place—a characteristic dislexically summed up in the Second Assessment Report from the U.N.’s Intergovernmental Panel on Climate Change (IPCC)
Warmer temperatures will lead to a more vigorous hydrological cycle; this translates into prospects for more severe droughts and/or floods in some places and less severe droughts and/or floods in other places.
So, according to the IPCC, whatever happens to precipitation will have been correctly forecast!
In some areas of the U.S., it is actually possible to pin down specific climate model expectations for precipitation changes. Unfortunately (for the models), the actual observations show little if any correspondence to the magnitude, or even direction, of the modeled changes.
We pointed out this disconcerting tendency in our comments made this spring during the public comment period for the current draft version of the government’s latest National Climate Assessment, and again in a presentation made earlier this summer at the Science Policy Conference of the American Geophysical Union (AGU).
Now, there’s a new paper that has just been accepted in AGU’s scientific journal Geophysical Research Letters that pretty much confirms what we presented to the AGU in June.
The paper is authored by Indrani Pal of the University of Colorado-Denver, and colleagues. The researchers set out to study whether changes in the frequency and/or timing of precipitation across the U.S. has changed over the period 1930-2009. They generally found that there was an overall trend towards more precipitation events (i.e., the frequency of precipitation was increasing) with the co-finding that there was a decrease in the timing between precipitation events (i.e., a decline in the length of dry spells).
The researchers describe their findings, and the possible reasons behind them (citations removed):
It is also important to note that this empirical assessment of historical trends in wet- and dry-season precipitation frequency and extreme dry spell length during the period 1930-2009 is first and foremost an observationally constrained detection study. Attribution to a particular cause will require further study, particularly at the regional scale where land use change, irrigation, urbanization, changing sea-surface temperatures and regional aerosols could all serve as external drivers for the changes observed here.
But while they are hesitant to speculate what may have led to the observed precipitation trends, the researchers are less hesitant to speculate as to how climate model forecasts of such precipitation changes are faring (citations removed, emphasis added):
At the outset, however, it is interesting to note that results based upon numerical climate models—which have a well-known positive bias in the frequency of precipitation—suggest that as radiative forcing associated with increased concentrations of heat-trapping gases such as CO2 continues to increase during the coming century, the frequency of precipitation days will decrease across the mid-latitudes, resulting in fewer but more intense precipitating days [they forgot to add “in some places and/or others”!].
Given that this expected behavior over the coming century is unambiguously at odds with the observational-based results over the past century, we suggest that a more detailed evaluation of the model systems be performed based upon their long-term simulations of these historical 20th century trends using more hydrologically relevant wet and dry season periods, particularly now that the observed trends for this period are found to be robust down to the station level. In addition, because the station data analyzed here are not directly comparable to model gridded data, it will also be important to evaluate concurrently whether the observed results change when aggregating the station precipitation data to the scale of the model grids.
“[T]he expected behavior over the coming century is unambiguously at odds with the observational-based results over the past century”—in other words, the observed trends in the frequency of precipitation across the U.S. established over the previous eighty years need to do a 180 in order to come into accordance with climate model predictions. Reading between the lines of the Pal et al. paper, we don’t pick up any indications that the authors see any signs that this U-turn is in the process of being executed.
What Pal et. al didn’t mention is that getting the temperature forecast right (the models don’t) while blowing the precipitation forecasts means that the temperature hit was just blind and dumb luck. Surface moisture is exceedingly important when it comes to temperature. Wet surfaces warm more slowly than dry ones, so if the surface is forecast to be getting drier than it actually is (which is what the models predict is going on), then warming is likely to be overforecast (which is also what is happening).
Consequently, the situation remains that climate models are failing to accurately capture known changes in central climate characteristics. Poor model performance in these areas does not provide much confidence in using these models to base policy decisions regarding greenhouse gas emissions and climate change.
This is a point we continually stress. For example, here are the conclusions from our AGU presentation:
It is impossible to present reliable future projections from a collection of climate models which generally cannot simulate observed change. As a consequence, we recommend that unless/until the collection of climate models can be demonstrated to accurately capture observed characteristics of known climate changes, policymakers should avoid basing any decisions upon projections made from them. Further, those policies which have already be established using projections from these climate models should be revisited.
Assessments which suffer from the inclusion of unreliable climate model projections include those produced by the Intergovernmental Panel on Climate Change and the U.S. Global Climate Change Research Program (including the draft of their most recent National Climate Assessment). Policies which are based upon such assessments include those established by the U.S. Environmental Protection Agency pertaining to the regulation of greenhouse gas emissions under the Clean Air Act.
Knappenberger, P.C., and P.J. Michaels, 2013. Policy Implications of Climate Models on the Verge of Failure. American Geophysical Union Science Policy Conference, Washington DC, June 24-26, 2013, Paper CC-15.
Pal, I., et al., 2013. Shifting seasonality and frequency of precipitation in wet and dry seasons across the U.S. Geophysical Research Letters, doi: 10.1002/grl.50760