There have been a number of studies which show that we can expect more extreme rainfall with a global warming (e.g. Donat et al., 2016). Hence, there is a need to increase our resilience to more rainfall in the future.

We can say something about how the rainfall statistics will be affected by a global warming, even when the weather itself is unpredictable beyond a few days.

Statistics is remarkably predictable for a large number of events where each of them is completely random (welcome to thermodynamics and quantum physics).

The normal distribution has often been used to describe the statistical character of daily temperature, but it is completely unsuitable for 24-hr precipitation. Instead, the gamma distribution has been a popular choice for describing rainfall.

I wonder, however, if there is an even better way to quantify rainfall statistics.

I have played around with the gamma distribution in an attempt to model daily rainfall statistics and its dependency on a set of physical factors. Without much success.

However, then I noticed that most daily rain gauge appeared to be almost exponentially distributed if I only included the rainy days (e.g. setting the threshold for a wet day at 1 mm).

When I plotted the histogram for rainfall on wet days with a log-y axis, I would mostly get a straight line of dots (see a typical example below).

Historgam of 24-hr precipitation measured at Bjørnholt in a forest near Oslo. There will always be some clutter at the upper end of plots like these because there are so few data points representing these extreme values.

The nice thing with the exponential distribution (which is a particular case of the gamma function) is that it only requires one parameter to specify the mathematical curve: it’s the inverse of the mean value .

I then used Bayes’ theorem to account for dry and wet days, where the probability for rainfall was taken to be the wet-day frequency .

The advantage of this approach is that I now had two parameters which were easy to estimate: the wet-day mean precipitation (or mean rainfall intensity) and the wet-day frequency .

Furthermore, it turned out that is often closely connected to the wind direction, and can easily be predicted based on circulation patterns or sea-level pressure anomalies.

It was harder to find a systematic influence on , as it is likely affected by several factors, including the air moisture (which depends on temperature) and cloud top heights.

The total precipitation is the product of , where is the number of days.

In other words,  and tell me many things I needed to know about the rainfall statistics (there are other aspects too, such as the mean duration of dry/wet spells, the spatial extent, and whether it comes as rain, sleet, snow or hail).

The equation for estimating the probability for a rain event with amounts exceeding can be written as (using 1-CDF for the exponential distribution):

(1)  

I have called it the “rain equation”, both because the name has not been taken and because it can provide many answers concerning rainfall.

It can address questions about the likelihood of heavy rainfall and whether it is due to an increase in the number of rainy days (e.g. due to changes in circulation) or because the rains have become more intense.

It is also on par with the normal distribution – in both cases, they are not meant to provide accurate probabilities for extreme events far out in the tails.

However, they are both capable of quantifying the probability of more moderate values, which can be illustrated in the figure below:

Figure 1. A comparison between probabilities estimated with the rain equation and the observed fraction of events with more than 30 mm rain in Groningen in the Netherlands. Here refers to the Heaviside function, which is a mathematical way of expressing that I only counted the number of events with more than 30 mm/day each year in the observervations (the plot was made with the R-package esd and the command test.rainequation(loc='GRONINGEN-1',threshold=20)).

The rain equation captures long-term changes as well as inter-annual variations. In this example, I used the annual wet-day mean precipitation and frequency estimated from the observations themselves to show its potential.

It can also be assessed against observations in a more systematic way, as in Figure 2:

Figure 2. A scatter plot of probabilities and corresponding fractions of events from long rain gauge records in Europe, based on the wet-day mean precipitation and frequency from the observations (the plot was made with the R-package esd and the command scatterplot.rainequation()).

A correlation of 0.98 is quite impressive, however, the rainfall is not perfectly exponentially distributed (Benestad et al., 2012). It nevertheless provides a means to address climate change connected to a change in either or .

We have used the rain equation in an attempt to downscale seasonal and decadal forecasts for precipitation (Benestad and Mezghani, 2015).

One thing that puzzles me, however, is that I cannot see this equation being used very much, despite the fact that it is so simple, seems so obvious, and can demonstrate impressive capabilities.

I would have thought it is an old formula. Perhaps one that has gotten out of fashion, but is documented in old papers that are not yet digitized and easy to google. Perhaps with a different name. Or have I missed something?

References

1. M.G. Donat, A.L. Lowry, L.V. Alexander, P.A. O’Gorman, and N. Maher, "More extreme precipitation in the world’s dry and wet regions", Nature Climate Change, vol. 6, pp. 508-513, 2016. http://dx.doi.org/10.1038/nclimate2941
2. R.E. Benestad, D. Nychka, and L.O. Mearns, "Spatially and temporally consistent prediction of heavy precipitation from mean values", Nature Climate Change, 2012. http://dx.doi.org/10.1038/nclimate1497
3. R.E. Benestad, and A. Mezghani, "On downscaling probabilities for heavy 24-hour precipitation events at seasonal-to-decadal scales", Tellus A: Dynamic Meteorology and Oceanography, vol. 67, pp. 25954, 2015. http://dx.doi.org/10.3402/tellusa.v67.25954

Some concerns about continued monitoring of sea ice by remote sensing were raised this week in Nature News an article in the (UK) Observer: Donald Trump accused of obstructing satellite research into climate change. The last headline is not really correct, but the underlying issues are real.

What is this about? Since the late seventies, there have been almost continuous observations of polar sea ice by passive microwave sensing on multiple polar-orbiting satellites. This is the preferred technique since microwaves from the surface can penetrate clouds (which are abundant in the polar regions) and can be detected during the day and night – again, important for the wintertime at the poles.

The current workhorse satellites for this measurement are the (aging) DMSP F-series (managed by the UASF). There are two currently operational for sea ice retrievals, F-16/18, which are 14 and 8 years old respectively. Another, F-17 is still in orbit, but may not be usable on its own for sea ice (. The design lifetime was nominally 5 years. A replacement satellite, F-19 failed completely in October, but in fact had not been useful for sea ice since February 2016. The last satellite in the series (F-20) was built two decades ago and kept in storage, but was decommissioned finally in November 2016 after a decision in Congress to no longer fund it in the FY16 budget. This was after the election, but before the inauguration of the Trump administration.

[Note: A comprehensive (though not always up-to-date) resource on satellite capabilities (OSCAR) is available from WMO if you want to navigate this for yourself.]

While many instruments can be used to detect sea ice, the continuity required for long-term climate monitoring makes it vital that the different products are cross-calibrated and have similar characteristics to be useful. The closest instruments to the those on the DMSP satellites are the radiometers (MWRI) on the Chinese Feng Yun-3 series of satellites. Unfortunately, again because of Congress, NASA collaborations with China are restricted and since the sea ice work at NSIDC is funded by NASA, that might prevent this source of data being used in the US (though presumably non-US colleagues would not have this problem).

Another possibility is the Japanese satellite GCOM-W1 which has a more advanced AMSR2 instrument (in space since 2012, and has also passed it’s design lifetime), but the merge of this data with the DMSP satellites is still a work in progress. This is being used for the Bremen University sea ice maps though.

Differences in views from passive microwave instruments (SSMI vs. AMSR2) via Arctic Roos.

Unfortunately, the next scheduled passive microwave sensor to be launched is not until 2022 on the European Space Agency’s 2nd Generation MetOp satellite, and will need a year’s overlap with an existing satellite to be optimally calibrated. Thus the likelihood of a gap in the record developing before then is very high.

Other measures of sea ice will be possible – IceSat2 is launching next year will have an active laser altimeter to measure sea ice height, satellites with visible or infrared capabilities are able to see ice when it’s not dark or cloudy, but cross calibration into a homogeneous record will be hard.

To be clear, people have been warning about this looming lack of capability for a while – in April and May 2016, as well as more recently. Unfortunately, new satellites and new instruments take a long while to develop, build and launch, and possibly we’ve been taking them for granted.

That probably needs to stop.

Note. Thanks to Walt Meier at NSIDC for chatting about these issues with me.

This month’s open thread. Lawsuits about scientific disputes, the new Climate Science Special Report from the National Climate Assessment, and (imminently) the WMO State of the Climate statement for 2017.

2017 is set to be one of warmest years on record. Gavin has been making regular forecasts of where 2017 will end up, and it is now set to be #2 or #3 in the list of hottest years:

With update thru September, ~80% chance of 2017 being 2nd warmest yr in the GISTEMP analysis (~20% for 3rd warmest). pic.twitter.com/k3CEM9rGHY

— Gavin Schmidt (@ClimateOfGavin) October 17, 2017

In either case it will be the warmest year on record that was not boosted by El Niño. I’ve been asked several times whether that is surprising. After all, the El Niño event, which pushed up the 2016 temperature, is well behind us. El Niño conditions prevailed in the tropical Pacific from October 2014 throughout 2015 and in the first half of 2016, giving way to a cold La Niña event in the latter half of 2016. (Note that global temperature lags El Niño variations by several months so this La Niña should have cooled 2017.)

The hot El Niño year of 1998 is comparable to 2016, since both years followed the two hitherto strongest El Niño events. And 1998 was followed by a cool 1999, only ranked #7 in the list of hottest years until then. So here is a comparison of 1998 versus 2016. Let us first look at the full time series of GISTEMP global temperature data, see Fig. 1.

Fig. 1 GISTEMP global temperature data, in 12-months running average (anomalies relative to the first 30 years). The data are available monthly and averaging over 12 months removes a considerable amount of month-to-month ‘noise’. Showing only calendar-year averages would lose some information – e.g. it would only fully show peaks in temperature if by chance the maxima aligned with the calendar year.

The El Niño peaks in 1998 and 2016 are clearly seen. It has been shown in several studies (e.g. Foster and Rahmstorf 2011) that El Niño is one of the main causes – perhaps the main cause – of short-term variability in global temperature. The following graph overlays those 2 El Niño peaks by shifting the 2016 peak back in time by 18 years and down by 0.4 °C.

Fig. 2 The two El Niño peaks in global temperature from Fig. 1, zoomed in and overlayed by shifting the 2016 peak back in time by 14 years and down by 0.4 °C. The darker red curve is the 2016 peak, as in Fig. 1.

The two peaks align very well. The first conclusion is that global temperature evolution over the last few years is very similar to that around 1998 – except the Earth is now 0.4 °C hotter. That’s 0.4 °C warming over 18 years, corresponding to 0.22 °C per decade – a bit more than expected from the long-term global warming trend since 1980, which is 0.17 °C per decade in the GISTEMP data. (So much for the “no warming since 1998” meme so popular with climate deniers.)

The second observation is that initially temperatures climbed down from the peak as fast as in 1998 – but then the cooling slowed down, and the last 12 months haven’t just been 0.4 °C warmer than in 98/99 but closer to 0.5 °C warmer. So it is clear that our planet is not cooling off as fast as after the 1998 El Niño peak. I wouldn’t over-interpret this – we’re looking at a really short interval here, so it is clearly no reason to diagnose a noteworthy acceleration of global warming. But there certainly is no sign of global warming slowing down. It will be interesting to watch how this continues over the next months; the ENSO forecast is for developing La Niña conditions again this coming fall/winter.

Guest Commentary by Scott Denning

The Orbiting Carbon Observatory (OCO-2) was launched in 2014 to make fine-scale measurements of the total column concentration of CO₂ in the atmosphere. As luck would have it, the initial couple of years of data from OCO-2 documented a period with the fastest rate of CO2 increase ever measured, more than 3 ppm per year (Jacobson et al, 2016;Wang et al, 2017) during a huge El Niño event that also saw global temperatures spike to record levels.

As part of a series of OCO-2 papers being published this week, a new Science paper by Junjie Liu and colleagues used NASA’s comprehensive Carbon Monitoring System to analyze millions of measurements from OCO-2 and other satellites to map the impact of the 2015-16 El Niño on sources and sinks of CO₂, providing insight into the mechanisms controlling carbon-climate feedback.

Uncertainty in Carbon-Climate Feedbacks is important

We’ve known for decades (Rayner et al, 1999) that El Niño influences the productivity of tropical forests and therefore CO₂, but we had very few direct observations of the effects because they are so remote. Field experiments on the ground and aircraft profiling of CO₂ over tropical forests have documented the impact of heat and drought on forest productivity, but they are few and far between. Vigorous convective mixing in the deep tropics also dilutes changes in near-surface CO₂ much more than at higher latitudes, so low-altitude sampling contains relatively less information about carbon sources and sinks.

A subset of Earth System Models (ESMs) project that El Niño-like conditions will progressively increase in coming decades as sea-surface temperatures in the tropical Pacific warm, implying increased drought and forest dieback in the Amazon. The drought-induced decline of carbon-dense tropical forests and their replacement by lower-carbon savannas would release enormous amounts of CO₂ to the atmosphere, amplifying global warming far beyond the effects of just the CO₂ released by burning fossil fuels. In the CMIP5 suite of ESMs summarized by the IPCC Fifth Assessment Report, models forced with identical fossil fuel emissions differed by as much as 350 ppm of CO₂ in 2100 due to differences in feedback between climate and the carbon cycle (Hoffmann et al, 2014). The radiative forcing of climate in these ESMs differed by up to 1.5 W/m², with much of the disparity being driven by interactions among warming oceans, atmospheric circulation, and tropical forests. The climate outcomes due to differences carbon-climate feedback are as different as those arising from different future emission scenarios (RCP6 compared to RCP4.5) or from differences in clouds and aerosols in atmospheric models.

The NASA Carbon Monitoring System

NASA’s Carbon Monitoring System (CMS) combines mechanistic “forward” models and empirical “inverse” models of atmospheric CO2 and other variables using a technique called “data assimilation” that is closely analogous to operational weather forecasting (Bowman et al, 2017). The forward models include emissions of CO₂ and carbon monoxide (CO) from fossil fuel burning and wildfires; air-sea gas exchange; and photosynthesis, respiration, and decomposition on land. These simulated emissions are then used as input to a model (GEOS-Chem) that uses high-resolution weather data atmospheric transport of CO₂ and CO by winds, clouds, and turbulence. The resulting 3D simulations are then sampled at the locations of OCO-2 observations to determine the error in the forward model of atmospheric variations. An “adjoint” of the atmospheric transport model is then run backward in time to quantify the contributions of errors in specified surface sources and sinks of CO₂ and CO to the mismatches between the forward models and the satellite observations.

OCO-2 has given us two revolutionary new ways to understand the effects of drought and heat on tropical forests. The instrument directly measures CO₂ over these regions thousands of times every day (Crisp et al, 2004). These column-averaged concentration retrievals respond to the net amount of CO₂ passing in and out of the atmosphere under the instrument. OCO-2 also senses the rate of photosynthesis by detecting fluorescent chlorophyll in the trees themselves (Frankenberg et al, 2011). Liu et al used observations of CO from the MOPITT instrument aboard NASA’s Terra satellite to identify CO₂ released from upwind wildfires. They used solar-induced chlorophyll fluorescence (SIF) to quantify changes in plant photosynthesis (also called gross primary production, GPP). Their results include time-resolved maps of the sources and sinks of atmospheric CO2 that are optimally consistent with both mechanistic forward models and the CO₂, CO, and SIF observed by the satellite instruments.

Fig. Extreme heat and drought impacted the carbon cycle in tropical forests differently in different regions, leading to the fastest growth rate of CO₂ in at least 10,000 years. (NASA/JPL-Caltech).

Carbon-Climate Feedback During the 2015-16 El Niño

As previously reported based on in-situ data, the rate of increase in atmospheric CO₂ during the strong El Niño in 2015-16 was about 3 ppm/yr compared with ~2 ppm/yr in recent decades. This is the fastest increase in CO₂ ever observed, and plausibly the fastest since the end of deglaciation 10,000 years ago. Yet this rapid increase in CO₂ occurred during a period when fossil fuel emissions were nearly flat (though still massively more than the biosphere and ocean can quickly absorb).

Liu et al found that 80% of the extra CO₂ in the atmosphere during this period originated in tropical forests. Relative to a more normal year (2011), they found that tropical forests lost about 2.5 billion tons of carbon (GtC) in 2015-16. (1 Gt = 1012 kg is the mass of 1 cubic km of water, and 1 GtC produces about 2.12 ppm of CO₂ in the air).

During the huge El Niño, parts of the Amazon experienced the driest conditions in at least 30 years as well as unusually warm temperatures. Changes in column-averaged CO₂ and in SIF showed that these hot, dry conditions suppressed gross primary production (GPP, photosynthesis), leading to a reduction of about 0.9 GtC/yr. Equatorial Africa also experienced extreme heat, but precipitation was near normal. Impacts on GPP were not significant, but respiration and decomposition were enhanced by about 0.8 GtC/yr. In Hot dry conditions in Indonesia during the period led to an increase in fires, including a large peat fire that burned huge amounts of stored carbon. Emissions due to these fires showed up in the observations as increases in both CO2 and CO, and were estimated at about 0.8 GtC/yr.

These results help us understand how drought and heat affect these forests, some of the most productive ecosystems on Earth. The Amazon has experienced three extreme droughts in the past 11 years, in 2005, 2010, and now 2015-16. These extreme events have occurred more frequently than they did in the previous century. Understanding how the tropical forest responds to big droughts and heat waves help us to evaluate the strength of carbon-climate feedback in ESMs, allowing us to better understand and predict climate change over coming decades. The new results show that each of the major tropical forest regions experienced different combinations of heat and drought during the recent El Niño, so their carbon cycles responded in different ways, but the net result was increased emissions in all cases. Based on these results, further warming and drying of tropical forests is expected to result in less uptake and more release of carbon on land, unfortunately amplifying the effect of fossil fuel emissions warming the climate.

References

1. J. Wang, N. Zeng, M. Wang, F. Jiang, H. Wang, and Z. Jiang, "Contrasting terrestrial carbon cycle responses to the two strongest El Niño events: 1997–98 and 2015–16 El Niños", Earth System Dynamics Discussions, pp. 1-32, 2017. http://dx.doi.org/10.5194/esd-2017-46
2. J. Liu, K.W. Bowman, D.S. Schimel, N.C. Parazoo, Z. Jiang, M. Lee, A.A. Bloom, D. Wunch, C. Frankenberg, Y. Sun, C.W. O’Dell, K.R. Gurney, D. Menemenlis, M. Gierach, D. Crisp, and A. Eldering, "Contrasting carbon cycle responses of the tropical continents to the 2015–2016 El Niño", Science, vol. 358, pp. eaam5690, 2017. http://dx.doi.org/10.1126/science.aam5690
3. P.J. Rayner, R.M. Law, and R. Dargaville, "The relationship between tropical CO2fluxes and the El Niño-Southern Oscillation", Geophysical Research Letters, vol. 26, pp. 493-496, 1999. http://dx.doi.org/10.1029/1999GL900008
4. "Climate Change 2013 - The Physical Science Basis", 2009. http://dx.doi.org/10.1017/CBO9781107415324
5. K.W. Bowman, J. Liu, A.A. Bloom, N.C. Parazoo, M. Lee, Z. Jiang, D. Menemenlis, M.M. Gierach, G.J. Collatz, K.R. Gurney, and D. Wunch, "Global and Brazilian Carbon Response to El Niño Modoki 2011-2010", Earth and Space Science, 2017. http://dx.doi.org/10.1002/2016EA000204
6. D. Crisp, R. Atlas, F. Breon, L. Brown, J. Burrows, P. Ciais, B. Connor, S. Doney, I. Fung, D. Jacob, C. Miller, D. O'Brien, S. Pawson, J. Randerson, P. Rayner, R. Salawitch, S. Sander, B. Sen, G. Stephens, P. Tans, G. Toon, P. Wennberg, S. Wofsy, Y. Yung, Z. Kuang, B. Chudasama, G. Sprague, B. Weiss, R. Pollock, D. Kenyon, and S. Schroll, "The Orbiting Carbon Observatory (OCO) mission", Advances in Space Research, vol. 34, pp. 700-709, 2004. http://dx.doi.org/10.1016/j.asr.2003.08.062
7. C. Frankenberg, C. O'Dell, J. Berry, L. Guanter, J. Joiner, P. Köhler, R. Pollock, and T.E. Taylor, "Prospects for chlorophyll fluorescence remote sensing from the Orbiting Carbon Observatory-2", Remote Sensing of Environment, vol. 147, pp. 1-12, 2014. http://dx.doi.org/10.1016/j.rse.2014.02.007

Guest commentary by Ben Sanderson

Millar et al’s recent paper in Nature Geoscience has provoked a lot of lively discussion, with the authors of the original paper releasing a statement to clarify that their paper did not suggest that “action to reduce greenhouse gas emissions is no longer urgent“, rather that 1.5ºC (above the pre-industrial) is not “geophysically impossible”.

The range of post-2014 allowable emissions for a 66% chance of not passing 1.5ºC in Millar et al of 200-240GtC implies that the planet would exceed the threshold after 2030 at current emissions levels, compared with the AR5 analysis which would imply most likely exceedance before 2020. Assuming the Millar numbers are correct changes 1.5ºC from fantasy to merely very difficult.

But is this statement overconfident? Last week’s post on Realclimate raised a couple of issues which imply that both the choice of observational dataset and the chosen pre-industrial baseline period can influence the conclusion of how much warming the Earth has experienced to date. Here, I consider three aspects of the analysis – and assess how they influence the conclusions of the study.

Figure 1: (a) shows temperature change in the CMIP5 simulations relative to observed temperature products. Grey regions show model range under historical and RCP8.5 forcing relative to a 1900-1940 baseline. Right-hand axis shows temperatures relative to 1861-1880 (offset using HadCRUT4 temperature difference). (b) shows temperature change as a function of cumulative emissions. Black solid line shows the CMIP5 historical mean, and black dashed is the RCP8.5 projection. Colored lines represent regression reconstructions as in Otto (2015) using observational temperatures from HadCRUT4 and GISTEMP, with cumulative emissions from the Global Carbon Project. Colored points show individual years from observations.

The choice of temperature data

We can illustrate how these effects might influence the Millar analysis by repeating the calculation with alternative temperature data. Their approach requires an estimate of the forced global mean temperature in a given year (excluding any natural variability), which are derived from Otto et al (2015), who employ a regression approach to reconstruct a prediction of global mean temperatures as a function of anthropogenic and natural forcing agents. In Fig. 1(a), we apply the Otto approach to data from GISTEMP as well as the HadCRUT4 product used in the original paper – again using data up to 2014. Although the HadCRUT4 forced Otto-style reconstruction suggests 2014 temperatures were less than the 25th percentile of the CMIP5 distribution, following the same procedure with GISTEMP yields 2014 temperatures of 1.08K – corresponding to the 58th percentile of the CMIP5 distribution.

This draws into question the justification for changing the baseline for the cumulative emissions analysis, given it quickly becomes apparent is that the use of a different dataset can undermine the conclusion that present day temperatures lie outside of the model distribution. Fig. 1(b) shows that the anomaly between observations and the CMIP5 mean temperature response to cumulative emissions is halved by repeating the Millar analysis with the GISTEMP product instead of HadCRUT.

The role of internal variability

There is also an important question of the degree to which internal variability can influence the attributable temperature change, given that the Millar result is contingent on knowing what the forced temperature response of the system is. We apply the approach to the CESM Large Ensemble, a 40-member initial condition ensemble of historical and future climate simulations where ensemble members differ only in their realization of natural variability. Although all models have identical forcing and model configuration, Fig 2(a) shows the range of estimated forced warming in 2014 in the CESM model by the Otto approach varies from 0.68-0.94K (almost as much as the actual 2005-2014 decadal average temperature itself in the CESM ensemble), and there is a strong correlation between inferred forced warming in 2014 and the global mean temperatures in the preceding decade, suggesting that the unforced estimate in the Otto approach can be strongly influenced by temperatures in the preceding decade.

Figure 2: (a) Temperatures of reconstructed global mean temperature in 2014 for the CESM large ensemble following the Otto (2015) regression methodology, plotted as a function of average global mean temperature in the years 2005-2014. (b) correlation between mean grid-point temperatures in 2005-2014 and reconstructed global mean temperature in 2014, ellipses show regions proposed for Pacific Climate Index. (c) CESM large ensemble reconstructed global mean temperature in 2014 as a function of Pacific Climate Index. Vertical lines show index values for observations in period 2005-2014 (solid) and historical (dashed).

In short, irrespective of what observational dataset was used – it’s likely that an estimate of forced response made in 2014 would be biased cold, which on its own would translate to an overestimate of the available budget of about 40GtC.

The low CMIP5 compatible emissions

Millar’s paper also points out that the discrepancy between the CMIP5 ensemble and the observations arises not only due to temperature, but also because cumulative emissions were greater in the real world than the mean CMIP5 model in 2014. But this only translates to a justifiable increase in the emissions budget if the real world is demonstrably off the CMIP5 cumulative emissions/temperature line.   By some estimates, cumulative emissions in 2014 might be higher than the models simply be because emissions were consistently above the RCP range between 2005-2014. In other words – by 2014 we’d used more of the carbon budget than any of the RCPs had anticipated and if we are not confident that the real world is cooler than the models at this level of cumulative emissions, this means that available emissions for 1.5 degrees should decrease proportionately.

A key point to note is that, by resetting the cumulative emissions baseline, the Millar et al available emissions budget is insensitive to the actual cumulative emissions to date. The unforced temperature estimate is used as a proxy for what cumulative emissions should be given the current level of warming. This is only justified if we are confident that we know the current unforced temperature more accurately than we know the current cumulative emissions. However, the combined evidence of the influence of natural variability on the unforced temperature estimate, the disagreement between different observational datasets on warming level, and the uncertainty introduced by an uncertain pre-industrial temperature baseline means that we can’t be confident as the Millar paper suggests on what the current level of warming is, and that the balance of evidence suggests that the Otto warming estimate may be biased cold. If this is right, the Millar available cumulative emissions budget would be biased high.

So, is it appropriate to say that 1.5ºC is geophysically possible? Perhaps plausible would be a better word. Depending on which temperature dataset we choose, the TEB for 1.5 degrees may already be exceeded. Although it would certainly be useful to know what the underlying climate attractor of the Earth system is, any estimate we produce is subject to error.

We ultimately face a question of what we trust more: our estimate of our cumulative emissions to date combined with our full knowledge of how much warming that might imply, or an estimate of how warm the system was in 2014 which is subject to error due to observational uncertainty and natural variability. Changing the baseline for warming and cumulative emissions is effectively a bias correction, a statement that models have simulated the past sufficiently poorly that they warrant bias correction which allows for emissions to date to be swept under the carpet. Alternatively, we trust the cumulative emissions number and treat the models as full proxies for reality, as was done in AR5, which would tell us that the emissions to date have already brought us to the brink of exceedance of the 1.5 degree threshold.

Methods

For Figure 1, global mean temperatures are plotted from the HadCRUT4 and GISTEMP products relative to a 1900-1940 baseline, together with global mean temperatures from 81 available simulations in the CMIP5 archive, also relative to the 1900-1940 baseline, where all available ensemble members are taken for each model. In each year from 1900-2016, the 5^th,25^th,75^th and 95^th percentiles of the CMIP5 distribution are plotted. Figure 1(b) the CMIP5 cumulative emissions and temperature data used are identical to those in AR5 for the historical and RCP8.5 trajectories. Observational data is from GISTEMP and HadCRUT4 global mean products, and annual cumulative emission data is replicated from the Global Carbon Project. Regression analyses are performed as in Otto (2015), using natural and anthropogenic forcing timeseries (historical and the RCP8.5 scenario) with a regression constructed using data from 1850-2016 (for HadCRUT4), and from 1880-2016 (for GISTEMP).

Figure 2 uses data from the CESM large ensemble, where the Otto (2015) analysis is applied to each ensemble member. Regressions are performed using data from years 1850-2014 for each member of the archive, to replicate the years used in the Otto (2015) analysis. Figure 2(a) shows the regression reconstructed temperature for 2014 plotted as a function of model temperatures in the preceding decade (2005-2014). Figure 2(b) shows the correlation pattern of 2005-2014 temperatures with the 2014 regression reconstruction in the CESM Large Ensemble. Ellipses are constructed to approximately highlight regions of high positive and negative correlation in the pattern. A central Pacific ellipse is centered on 2N,212E, while a South Pacific ellipse is centered on 34S,220E. Figure 2(c) employs a Pacific Climate Index specific to this analysis, constructed using the difference of mean annual temperatures in the years 2005-2014 in the two ellipses in Figure 2(b) (central Pacific minus south Pacific region), showing reconstructed 2014 regression temperatures as a function of the index for each member of the CESM Large Ensemble. A linear regression was computed to predict 2014 Otto (2015) forced temperatures as a function of the Pacific Climate Index. Dashed lines show the 5^th and 95^th percentile uncertainty in the regression coefficients. The same index is then calculated for the 2005-2014 period and the historical 1880-2000 period in the HadCRUT4 and GISTEMP datasets.

Back in February we highlighted an obviously wrong paper by Harde which purported to scrutinize the carbon cycle. Well, thanks to a crowd sourced effort which we helped instigate, a comprehensive scrutiny of those claims has just been published. Lead by Peter Köhler, this included scientists from multiple disciplines working together to clearly report on the mistaken assumptions in the Harde paper.

The comment is excellent, and so should be well regarded, but the fact that it is a comment means that the effort will likely be sorely underappreciated. Part of problem is the long time for the process (almost 8 months) which means that the nonsense is mostly forgotten about by the time the comments are published. We’ve discussed trying to speed up and improve the process by having a specialized journal for comments and replications but really the problem here is the low quality of peer review and editorial supervision that allows these pre-rebunked papers to appear in the first place.

GPC is not the only (nor the worst) culprit for this kind of nonsense – indeed we just noticed a bunch of astrology papers in the International Journal of Heat and Technology (by Nicola Scatetta [natch]). It does seem to demonstrate that truly you can indeed publish anything somewhere.

References

1. P. Köhler, J. Hauck, C. Völker, D.A. Wolf-Gladrow, M. Butzin, J.B. Halpern, K. Rice, and R.E. Zeebe, "Comment on “ Scrutinizing the carbon cycle and CO 2 residence time in the atmosphere ” by H. Harde", Global and Planetary Change, 2017. http://dx.doi.org/10.1016/j.gloplacha.2017.09.015

This month’s open thread. Carbon budgets, Arctic sea ice minimum, methane emissions, hurricanes, volcanic impacts on climate… Please try and stick to these or similar topics.

There has been a bit of excitement and confusion this week about a new paper in Nature Geoscience, claiming that we can still limit global warming to below 1.5 °C above preindustrial temperatures, whilst emitting another ~800 Gigatons of carbon dioxide. That’s much more than previously thought, so how come? And while that sounds like very welcome good news, is it true? Here’s the key points.

Emissions budgets – a very useful concept

First of all – what the heck is an “emissions budget” for CO2? Behind this concept is the fact that the amount of global warming that is reached before temperatures stabilise depends (to good approximation) on the cumulative emissions of CO2, i.e. the grand total that humanity has emitted. That is because any additional amount of CO2 in the atmosphere will remain there for a very long time (to the extent that our emissions this century will like prevent the next Ice Age due to begin 50 000 years from now). That is quite different from many atmospheric pollutants that we are used to, for example smog. When you put filters on dirty power stations, the smog will disappear. When you do this ten years later, you just have to stand the smog for a further ten years before it goes away. Not so with CO2 and global warming. If you keep emitting CO2 for another ten years, CO2 levels in the atmosphere will increase further for another ten years, and then stay higher for centuries to come. Limiting global warming to a given level (like 1.5 °C) will require more and more rapid (and thus costly) emissions reductions with every year of delay, and simply become unattainable at some point.

It’s like having a limited amount of cake. If we eat it all in the morning, we won’t have any left in the afternoon. The debate about the size of the emissions budget is like a debate about how much cake we have left, and how long we can keep eating cake before it’s gone. Thus, the concept of an emissions budget is very useful to get the message across that the amount of CO2 that we can still emit in total (not per year) is limited if we want to stabilise global temperature at a given level, so any delay in reducing emissions can be detrimental – especially if we cross tipping points in the climate system, e.g trigger the complete loss of the Greenland Ice Sheet. Understanding this fact is critical, even if the exact size of the budget is not known.

But of course the question arises: how large is this budget? There is not one simple answer to this, because it depends on the choice of warming limit, on what happens with climate drivers other than CO2 (other greenhouse gases, aerosols), and (given there’s uncertainties) on the probability with which you want to stay below the chosen warming limit. Hence, depending on assumptions made, different groups of scientists will estimate different budget sizes.

Computing the budget

The standard approach to computing the remaining carbon budget is:

(1) Take a bunch of climate and carbon cycle models, start them from preindustrial conditions and find out after what amount of cumulative CO2 emissions they reach 1.5 °C (or 2 °C, or whatever limit you want).

(2) Estimate from historic fossil fuel use and deforestation data how much humanity has already emitted.

The difference between those two numbers is our remaining budget. But there are some problems with this. The first is that you’re taking the difference between two large and uncertain numbers, which is not a very robust approach. Millar et al. fixed this problem by starting the budget calculation in 2015, to directly determine the remaining budget up to 1.5 °C. This is good – in fact I suggested doing just that to my colleague Malte Meinshausen back in March. Two further problems will become apparent below, when we discuss the results of Millar et al.

So what did Millar and colleagues do?

A lot of people were asking this, since actually it was difficult to see right away why they got such a surprisingly large emissions budget for 1.5 °C. And indeed there is not one simple catch-all explanation. Several assumptions combined made the budget so big.

The temperature in 2015

To compute a budget from 2015 to “1.5 °C above preindustrial”, you first need to know at what temperature level above preindustrial 2015 was. And you have to remove short-term variability, because the Paris target applies to mean climate. Millar et al. concluded that 2015 was 0.93 °C above preindustrial. That’s a first point of criticism, because this estimate (as Millar confirmed to me by email) is entirely based on the Hadley Center temperature data, which notoriously have a huge data gap in the Arctic. (Here at RealClimate we were actually the first to discuss this problem, back in 2008.) As the Arctic has warmed far more than the global mean, this leads to an underestimate of global warming up to 2015, by 0.06 °C when compared to the Cowtan&Way data or by 0.17 °C when compared to the Berkeley Earth data, as Zeke Hausfather shows in detail over at Carbon Brief.

Figure: Difference between modeled and observed warming in 2015, with respect to the 1861-1880 average. Observational data has had short-term variability removed per the Otto et al 2015 approach used in the Millar et al 2017. Both RCP4.5 CMIP5 multimodel mean surface air temperatures (via KNMI) and blended surface air/ocean temperatures (via Cowtan et al 2015) are shown – the latter provide the proper “apples-to-apples” comparison. Chart by Carbon Brief.

As a matter of fact, as Hausfather shows in a second graph, HadCRUT4 is the outlier data set here, and given the Arctic data gap we’re pretty sure it is not the best data set. So, while the large budget of Millar et al. is based on the idea that we have 0.6 °C to go until 1.5 °C, if you believe (with good reason) that the Berkeley data are more accurate we only have 0.4 °C to go. That immediately cuts the budget of Millar et al. from 242 GtC to 152 GtC (their Table 2). [A note on units: you need to always check whether budgets are given in billion tons of carbon (GtC) or billion tons of carbon dioxide. 1 GtC = 3.7 GtCO2, so those 242 GtC are the same as 887 GtCO2.] Gavin managed to make this point in a tweet:

Headline claim from carbon budget paper that warming is 0.9ºC from pre-I is unsupported. Using globally complete estimates ~1.2ºC (in 2015) pic.twitter.com/B4iImGzeDE

— Gavin Schmidt (@ClimateOfGavin) September 20, 2017

Add to that the question of what years define the “preindustrial” baseline. Millar et al. use the period 1861-80. For example, Mike has argued that the period AD 1400-1800 would be a more appropriate preindustrial baseline (Schurer et al. 2017). That would add 0.2 °C to the anthropogenic warming that has already occurred, leaving us with just 0.2 °C and almost no budget to go until 1.5 °C. So in summary, the assumption by Millar et al. that we still have 0.6 °C to go up to 1.5 °C is at the extreme high end of how you might estimate that remaining temperature leeway, and that is one key reason why their budget is large. The second main reason follows.

To exceed or to avoid…

Here is another problem with the budget calculation: the model scenarios used for this actually exceed 1.5 °C warming. And the 1.5 °C budget is taken as the amount emitted by the time when the 1.5 °C line is crossed. Now if you stop emitting immediately at this point, of course global temperature will rise further. From sheer thermal inertia of the oceans, but also because if you close down all coal power stations etc., aerosol pollution in the atmosphere, which has a sizeable cooling effect, will go way down, while CO2 stays high. So with this kind of scenario you will not limit global warming to 1.5 °C. This is called a “threshold exceedance budget” or TEB – Glen Peters has a nice explainer on that (see his Fig. 3). All the headline budget numbers of Millar et al., shown in their Tables 1 and 2, are TEBs. What we need to know, though, is “threshold avoidance budgets”, or TAB, if we want to stay below 1.5 °C.

Millar et al also used a second method to compute budgets, shown in their Figure 3. However, as Millar told me in an email, these “simple model budgets are neither TEBs nor TABs (the 66 percentile line clearly exceeds 1.5 °C in Figure 3a), they are instead net budgets between the start of 2015 and the end of 2099.” What they are is budgets that cause temperature to exceed 1.5 °C in mid-century, but then global temperature goes back down to 1.5 °C in the year 2100!

In summary, both approaches used by Millar compute budgets that do not actually keep global warming to 1.5 °C.

How some media (usual suspects in fact) misreported

We’ve seen a bizarre (well, if you know the climate denialist scene, not so bizarre) misreporting about Millar et al., focusing on the claim that climate models have supposedly overestimated global warming. Carbon Brief and Climate Feedback both have good pieces up debunking this claim, so I won’t delve into it much. Let me just mention one key aspect that has been misunderstood. Millar et al. wrote the confusing sentence: “in the mean CMIP5 response cumulative emissions do not reach 545GtC until after 2020, by which time the CMIP5 ensemble-mean human-induced warming is over 0.3 °C warmer than the central estimate for human-induced warming to 2015”. As has been noted by others, this is comparing model temperatures after 2020 to an observation-based temperature in 2015, and of course the latter is lower – partly because it is based on HadCRUT4 data as discussed above, but equally so  because of comparing different points in time. This is because it refers to the point when 545 GtC is reached. But the standard CMIP5 climate models used here are not actually driven by emissions at all, but by atmospheric CO2 concentrations. For the historic period, these are taken from observed data. So the fact that 545 GtC are reached too late doesn’t even refer to the usual climate model scenarios. It refers to estimates of emissions by carbon cycle models, which are run in an attempt to derive the emissions that would have led to the observed time evolution of CO2 concentration.

Does it all matter?

We still live in a world on a path to 3 or 4 °C global warming, waiting to finally turn the tide of rising emissions. At this point, debating whether we have 0.2 °C more or less to go until we reach 1.5 °C is an academic discussion at best, a distraction at worst. The big issue is that we need to see falling emissions globally very very soon if we even want to stay well below 2 °C. That was agreed as the weaker goal in Paris in a consensus by 195 nations. It is high time that everyone backs this up with actions, not just words.

Technical p.s. A couple of less important technical points. The estimate of 0.93 °C above 1861-80 used by Millar et al. is an estimate of the human-caused warming. I don’t know whether the Paris agreement specifies to limit human-caused warming, or just warming, to 1.5 °C – but in practice it does not matter, since the human-caused warming component is almost exactly 100 % of the observed warming. Using the same procedure as Millar yields 0.94 °C for total observed climate warming by 2015, according to Hausfather.

However, updating the statistical model used to derive the 0.93 °C anthropogenic warming to include data up to 2016 gives an anthropogenic warming of 0.96 °C in 2015.

Weblink

Statement by Millar and coauthors pushing back against media misreporting. Quote: “We find that, to likely meet the Paris goal, emission reductions would need to begin immediately and reach zero in less than 40 years’ time.”

The 2017 annual assembly of the European Meteorological Society (EMS) had a new set-up with a plenary keynote each morning. I though some of these keynotes were very interesting. There was a talk by Florence Rabier from the European Centre for Medium-range Weather Forecasts (ECMWF), who presented the story of ensemble forecasting. Keith Seitter, the executive director of the American Meteorological Society (AMS), talked about the engagement with the society on the Wednesday.

The Helix at DCU was the main venue of #EMS2017

It is impossible to attend all talks of such a conference with several parallel sessions. Besides, I gave most of my attention to the session on synoptic climatology which I co-convened together with Radan Huth.

One particularly interesting talk was a presentation by Rodrigo Caballero on standing waves, the jet stream and mid-latitude storms. He discussed the connection between cold extremes over North America, a perturbed upper-level jet stream, and eddy activity over the North Atlantic.

The talk covered several interesting topics: the idea that North Atlantic Oscillation (NAO) is a breaking Rossby wave; episodes of double wave-breaking on the northern and southern side of the jet resulting intense extreme winds; storm clustering; and planetary waves with wave number 5 and zero phase speed but eastward group velocity.

Another brilliant talk was given by Stephen Blenkinsop on intense rain with short duration. He presented observations which indicate that there has been a global intensification of short-duration (time scale: minutes-hours) rainfall events. A new global sub-daily precipitation data set HadISD was presented, which is an output from the EU-project INTENSE.

I have become more selective when it comes to attending talks with age, in order to avoid conference fatigue and information overload. There probably were many other excellent presentations at EMS 2017 which I didn’t catch.

I also think that too many talks tend to be overly cryptic and too many presenters try to cram too much information into way too many slides (e.g 25 slides in a 12 minute talk, many of which loaded with a large number of tiny maps).

It is important to make it easy for the audience, keeping in mind that people often try to digest about 20 of these talks per day.

There was an excellent talk on science communication by C. Alex Young from NASA, and it was too bad it wasn’t streamed. He had some good advice to science presenters that may help them get their message across.