The Science of Doom

Impacts – II – GHG Emissions Projections: SRES and RCP

Impacts – IV – Temperature Projections and Probabilities

Impacts – V – Climate change is already causing worsening storms, floods and droughts

Impacts – X – Sea Level Rise 5 – Bangladesh

Impacts XI – Rainfall 1

Impacts – XII – Rainfall 2

Impacts – XIII – Rainfall 3

References

Hurricanes and climate: the US CLIVAR working group on hurricanes, American Meteorological Society, Kevin Walsh et al (2015) – free paper

Whither Hurricane Activity? Gabriel A Vecchi, Kyle L Swanson & Brian J. Soden, Science (2008) – free paper

Simulation of the Recent Multidecadal Increase of Atlantic Hurricane Activity Using an 18-km-Grid Regional Model, Thomas Knutson et al, American Meteorological Society, (2007) – free paper

Modeled Impact of Anthropogenic Warming on the Frequency of Intense Atlantic Hurricanes, Morris A Bender et al, Science (2010) – free paper

Notes

Note 1: The scenario is A1B, which is similar to RCP6 – that is, an approximate doubling of CO2 by the end of the century. The simulations came from the CMIP3 suite of model results.

Impacts – XIII – Rainfall 3

Posted in Atmospheric Physics, Climate Models, Impacts on August 21, 2017| 57 Comments »

In XII – Rainfall 2 we saw the results of many models on rainfall as GHGs increase. They project wetter tropics, drier subtropics and wetter higher latitude regions. We also saw an expectation that rainfall will increase globally, with something like 2-3% per ºC of warming.

Here is a (too small) graph from Allen & Ingram (2002) showing the model response of rainfall under temperature changes from GHG increases. The dashed line marked “C-C” is the famous (in climate physics) Clausius–Clapeyron relation which, at current temperatures, shows a 7% change in water vapor per ºC of warming. The red triangles are the precipitation changes from model simulations showing about half of that.

From Allen & Ingram (2002)

Figure 1

Here is another graph from the same paper showing global mean temperature change (top) and rainfall over land (bottom):

From Allen & Ingram (2002)

Figure 2

The temperature has increased over the last 50 years, and models and observations show that the precipitation has.. oh, it’s not changed. What is going on?

First, the authors explain some important background:

The distribution of moisture in the troposphere (the part of the atmosphere that is strongly coupled to the surface) is complex, but there is one clear and strong control: moisture condenses out of supersaturated air.

This constraint broadly accounts for the humidity of tropospheric air parcels above the boundary layer, because almost all such parcels will have reached saturation at some point in their recent history. Physically, therefore, it has long seemed plausible that the distribution of relative humidity would remain roughly constant under climate change, in which case the Clausius-Clapeyron relation implies that specific humidity would increase roughly exponentially with temperature.

This reasoning is strongest at higher latitudes where air is usually closer to saturation, and where relative humidity is indeed roughly constant through the substantial temperature changes of the seasonal cycle. For lower latitudes it has been argued that the real-world response might be different. But relative humidity seems to change little at low latitudes under a global warming scenario, even in models of very high vertical resolution, suggesting this may be a robust ’emergent constraint’ on which models have already converged.

They continue:

If tropospheric moisture loading is controlled by the constraints of (approximately) unchanged relative humidity and the Clausius-Clapeyron relation, should we expect a corresponding exponential increase in global precipitation and the overall intensity of the hydrological cycle as global temperatures rise?

This is certainly not what is observed in models.

To clarify, the point in the last sentence is that models do show an increase in precipitation, but not at the same rate as the expected increase in specific humidity (see note 1 for new readers).

They describe their figure 2 (our figure 1 above) and explain:

The explanation for these model results is that changes in the overall intensity of the hydrological cycle are controlled not by the availability of moisture, but by the availability of energy: specifically, the ability of the troposphere to radiate away latent heat released by precipitation.

At the simplest level, the energy budgets of the surface and troposphere can be summed up as a net radiative heating of the surface (from solar radiation, partly offset by radiative cooling) and a net radiative cooling of the troposphere to the surface and to space (R) being balanced by an upward latent heat flux (LP, where L is the latent heat of evaporation and P is global-mean precipitation): evaporation cools the surface and precipitation heats the troposphere.

[Emphasis added].

Basics Digression

Picture the atmosphere over a long period of time (like a decade), and for the whole globe. If it hasn’t heated up or cooled down we know that the energy in must equal energy out (or if it has only done so only marginally then energy in is almost equal to energy out). This is the first law of thermodynamics – energy is conserved.

What energy comes into the atmosphere?

Solar radiation is partly absorbed by the atmosphere (most is transmitted through and heats the surface of the earth)
Radiation emitted from the earth’s surface (we’ll call this terrestrial radiation) is mostly absorbed by the atmosphere (some is transmitted straight through to space)
Warm air is convected up from the surface
Heat stored in evaporated water vapor (latent heat) is convected up from the surface and the water vapor condenses out, releasing heat into the atmosphere when this happens

How does the atmosphere lose energy?

It radiates downwards to the surface
It radiates out to space

..end of digression

Changing Energy Budget

In a warmer world, if we have more evaporation we have more latent heat transfer from the surface into the troposphere. But the atmosphere has to be able to radiate this heat away. If it can’t, then the atmosphere becomes warmer, and this reduces convection. So with a warmer surface we may have a plentiful potential supply of latent heat (via water vapor) but the atmosphere needs a mechanism to radiate away this heat.

Allen & Ingram put forward a simple conceptual equation:

ΔR_c + ΔR_T = LΔP

where the change in radiative cooling ΔR, is split into two components: ΔR_c that is independent of the change in atmospheric temperature; and ΔR_T that depends only on the temperature

L = latent heat of water vapor (a constant), ΔP = change in rainfall (= change in evaporation, as evaporation is balanced by rainfall)

LΔP is about 1W/m² per 1% increase in global precipitation.

Now, if we double CO2, then before any temperature changes we decrease the outgoing longwave radiation through the tropopause (the top of the troposphere) by about 3-4W/m² and we increase atmospheric radiation to the surface by about 1W/m².

So doubling CO2, ΔR_c = -2 to -3W/m²; prior to a temperature change ΔR_T = 0; and so ΔP reduces.

The authors comment that increasing CO2 before any temperature change takes place reduces the intensity of the hydrological cycle and this effect was seen in early modeling experiments using prescribed sea surface temperatures.

Now, of course, the idea of doubling CO2 without any temperature change is just a thought experiment. But it’s an important thought experiment because it lets us isolate different factors.

The authors then consider their factor ΔR_T:

The enhanced radiative cooling due to tropospheric warming, ΔR_T, is approximately proportional to ΔT: tropospheric temperatures scale with the surface temperature change and warmer air radiates more energy, so ΔR_T = kΔT, with k=3W/(m²K)..

All this is saying is that as the surface warms, the atmosphere warms at about the same rate, and the atmosphere then emits more radiation. This is why the model results of rainfall in our figure 2 above show no trend in rainfall over 50 years, and also match the observations – the constraint on rainfall is the changing radiative balance in the troposphere.

And so they point out:

Thus, although there is clearly usable information in fig. 3 [our figure 2], it would be physically unjustified to estimate ΔP/ΔT directly from 20th century observations and assume that the same quantity will apply in the future, when the balance between climate drivers will be very different.

There is a lot of other interesting commentary in their paper, although the paper itself is now quite dated (and unfortunately behind a paywall). In essence they discuss the difficulties of modeling precipitation changes, especially for a given region, and are looking for “emergent constraints” from more fundamental physics that might help constrain forecasts.

A forecasting system that rules out some currently conceivable futures as unlikely could be far more useful for long-range planning than a small number of ultra-high-resolution forecasts that simply rule in some (very detailed futures as possibilities).

This is a very important point when considering impacts.

Conclusion

Increasing the surface temperature by 1ºC is expected to increase the humidity over the ocean by about 7%. This is simply the basic physics of saturation. However, climate models predict an increase in mean rainfall of maybe 2-3% per ºC. The fundamental reason is that the movement of latent heat from the surface to the atmosphere has to be radiated away by the atmosphere, and so the constraint is the ability of the atmosphere to do this. And so the limiting factor in increasing rainfall is not the humidity increase, it is the radiative cooling of the atmosphere.

We also see that despite 50 years of warming, mean rainfall hasn’t changed. Models also predict this. This is believed to be a transient state, for reasons explained in the article.

References

Constraints on future changes in climate and the hydrologic cycle, MR Allen & WJ Ingram, Nature (2002) – freely available [thanks, Robert]

Notes

1 Relative humidity is measured as a percentage. If the relative humidity = 100% it means the air is saturated with water vapor – it can’t hold any more water vapor. If the relative humidity = 0% it means the air is completely dry. As temperature increases the ability of air to hold water vapor increases non-linearly.

For example, at 0ºC, 1kg of air can carry around 4g of water vapor, at 10ºC that has doubled to 8g, and at 20ºC it has doubled again to 15g (I’m using approximate values).

So now imagine saturated air over the ocean at 20ºC rising up and therefore cooling (it is cooler higher up in the atmosphere). By the time the air parcel has cooled down to 0ºC (this might be anything from 2km to 5km altitude) it is still saturated but is only carrying 4g of water vapor, having condensed out 11g into water droplets.

Impacts – XII – Rainfall 2

Posted in Atmospheric Physics, Climate Models, Impacts on May 18, 2017| 21 Comments »

I probably should have started a separate series on rainfall and then woven the results back into the Impacts series. It might take a few articles working through the underlying physics and how models and observations of current and past climate compare before being able to consider impacts.

There are a number of different ways to look at rainfall models and reality:

What underlying physics provides definite constraints regardless of individual models, groups of models or parameterizations?
How well do models represent the geographical distribution of rain over a climatological period like 30 years? (e.g. figure 8 in Impacts XI – Rainfall 1)
How well do models represent the time series changes of rainfall?
How well do models represent the land vs ocean? (when we think about impacts, rainfall over land is what we care about)
How well do models represent the distribution of rainfall and the changing distribution of rainfall, from lightest to heaviest?

In this article I thought I would highlight a set of conclusions from one paper among many. It’s a good starting point. The paper is A canonical response of precipitation characteristics to global warming from CMIP5 models by Lau and his colleagues, and is freely available, and as always I recommend people read the whole paper, along with the supporting information that is also available via the link.

As an introduction, the underlying physics perhaps provides some constraints. This is strongly believed in the modeling community. The constraint is a simple one – if we warm the ocean by 1K (= 1ºC) then the amount of water vapor above the ocean surface increases by about 7%. So we expect a warmer world to have more water vapor – at least in the boundary layer (typically 1km) and over the ocean. If we have more water vapor then we expect more rainfall. But GCMs and also simple models suggest a lower value, like 2-3% per K, not 7%/K. We will come back to why in another article.

It also seems from models that with global warming, rainfall increases more in regions and times of already high rainfall and reduces in regions and times of low rainfall – the “wet get wetter and the dry get drier”. (Also a marketing mantra that introducing a catchy slogan ensures better progress of an idea). So we also expect changes in the distribution of rainfall. One reason for this is a change in the tropical circulation. All to be covered later, so onto the paper..

We analyze the outputs of 14 CMIP5 models based on a 140 year experiment with a prescribed 1% per year increase in CO2 emission. This rate of CO2 increase is comparable to that prescribed for the RCP8.5, a relatively conservative business-as-usual scenario, except the latter includes also changes in other GHG and aerosols, besides CO2.

A 27-year period at the beginning of the integration is used as the control to compute rainfall and temperature statistics, and to compare with climatology (1979–2005) of rainfall data from the Global Precipitation Climatology Project (GPCP). Two similar 27-year periods in the experiment that correspond approximately to a doubling of CO2 emissions (DCO2) and a tripling of CO2 emissions (TCO2) compared to the control are chosen respectively to compute the same statistics..

Just a note that I disagree with the claim that RCP8.5 is a “relatively conservative business as usual scenario” (see Impacts – II – GHG Emissions Projections: SRES and RCP), but that’s just an opinion, as are all views about where the world will be in population, GDP and cumulative emissions 100-150 years from now. It doesn’t detract from the rainfall analysis in the paper.

For people wondering “what is CMIP5?” – this is the model inter-comparison project for the most recent IPCC report (AR5) where many models have to address the same questions so they can be compared.

Here we see (and along with other graphs you can click to enlarge) what the models show in temperature (top left), mean global rainfall (top right), zonal rainfall anomaly by latitude (bottom left) and the control vs the tripled CO2 comparison (bottom right). The many different colors in the first three graphs are each model, while the black line is the mean of the models (“ensemble mean”). The bottom right graph helps put the changes shown in the bottom left into a perspective – with the different between the red and the blue being the difference between tripling CO2 and today:

From Lau et al 2013

Figure 1 – Click to enlarge

In the figure above, the bottom left graph shows anomalies. We see one of the characteristics of models as a result of more GHGs – wetter tropics and drier sub-tropics, along with wetter conditions at higher latitudes.

From the supplementary material, below we see a better regional breakdown of fig 1d (bottom right in the figure above). I’ll highlight the bottom left graph (c) for the African region. Over the continent, the differences between present day and tripling CO2 seem minor as far as model predictions go for mean rainfall:

From Lau et al 2013

Figure 2 – Click to enlarge

The supplementary material also has a comparison between models and observations. The first graph below is what we are looking at (the second graph we will consider afterwards). TRMM (Tropical Rainfall Measuring Mission) is satellite data and GPCP one rainfall climatology that we met in the last article – so they are both observational datasets. We see that the models over-estimate tropic rainfall, especially south of the equator:

From Lau et al 2013

Figure 3 – Click to enlarge

Rainfall Distribution from Light through to Heavy Rain

Lau and his colleagues then look at rainfall distribution in terms of light rainfall through to heavier rainfall. So, take global rainfall and divide it into frequency of occurrence, with light rainfall to the left and heavy rainfall to the right. Take a look back at the bottom graph in the figure above (figure 3, their figure S1). Note that the horizontal axis is logarithmic, with a ratio of over 1000 from left to right.

It isn’t an immediately intuitive graph. Basically there are two sets of graphs. The left “cluster” is how often that rainfall amount occurred, and the black line is GPCP observations. The “right cluster” is how much rainfall fell (as a percentage of total rainfall) for that rainfall amount and again black is observations.

So lighter rainfall, like 1mm/day and below accounts for 50% of time, but being light rainfall accounts for less than 10% of total rainfall.

To facilitate discussion regarding rainfall characteristics in this work, we define, based on the ensemble model PDF, three major rain types: light rain (LR), moderate rain (MR), and heavy rain (HR) respectively as those with monthly mean rain rate below the 20th percentile (<0.3 mm/day), between response (TCO2 minus control, black) and the inter-model 1s the 40th–70th percentile (0.9–2.4mm/day), and above the 98.5% percentile (>9mm/day). An extremely heavy rain (EHR) type defined at the 99.9th percentile (>24 mm day1) will also be referred to, as appropriate.

Here is a geographical breakdown of the total and then the rainfall in these three categories, model mean on the left and observations on the right:

From Lau et al 2013

Figure 4 – Click to enlarge

We can see that the models tend to overestimate the heavy rain and underestimate the light rain. These graphics are excellent because they help us to see the geographical distribution.

Now in the graphs below we see at the top the changes in frequency of mean precipitation (60S-60N) as a function of rain rate; and at the bottom we see the % change in rainfall per K of temperature change, again as a function of rain rate. Note that the bottom graph also has a logarithmic scale for the % change, so as you move up each grid square the value is doubled.

The different models are also helpfully indicated so the spread can be seen:

From Lau et al 2013

Figure 5 – Click to enlarge

Notice that the models are all predicting quite a high % change in rainfall per K for the heaviest rain – something around 50%. In contrast the light rainfall is expected to be up a few % per K and the medium rainfall is expected to be down a few % per K.

Globally, rainfall increases by 4.5%, with a sensitivity (dP/P/dT) of 1.4% per K

Here is a table from their supplementary material with a zonal breakdown of changes in mean rainfall (so not divided into heavy, light etc). For the non-maths people the first row, dP/P is just the % change in precipitation (“d” in front of a variable means “change in that variable”), the second row is change in temperature and the third row is the % change in rainfall per K (or ºC) of warming from GHGs:

From Lau et al 2013

Figure 6 – Click to enlarge

Here are the projected geographical distributions of the changes in mean (top left), heavy (top right), medium (bottom left) and light rain (bottom right) – using their earlier definitions – under tripling CO2:

From Lau et al 2013

Figure 7 – Click to enlarge

And as a result of these projections, the authors also show the number of dry months and the projected changes in number of dry months:

From Lau et al 2013

Figure 8 – Click to enlarge

The authors conclude:

The IPCC CMIP5 models project a robust, canonical global response of rainfall characteristics to CO2 warming, featuring an increase in heavy rain, a reduction in moderate rain, and an increase in light rain occurrence and amount globally.

For a scenario of 1% CO2 increase per year, the model ensemble mean projects at the time of approximately tripling of the CO2 emissions, the probability of occurring of extremely heavy rain (monthly mean >24mm/day) will increase globally by 100%–250%, moderate rain will decrease by 5%–10% and light rain will increase by 10%–15%.

The increase in heavy rain is most pronounced in the equatorial central Pacific and the Asian monsoon regions. Moderate rain is reduced over extensive oceanic regions in the subtropics and extratropics, but increased over the extratropical land regions of North America, and Eurasia, and extratropical Southern Oceans. Light rain is mostly found to be inversely related to moderate rain locally, and with heavy rain in the central Pacific.

The model ensemble also projects a significant global increase up to 16% more frequent in the occurrences of dry months (drought conditions), mostly over the subtropics as well as marginal convective zone in equatorial land regions, reflecting an expansion of the desert and arid zones..

..Hence, the canonical global rainfall response to CO2 warming captured in the CMIP5 model projection suggests a global scale readjustment involving changes in circulation and rainfall characteristics, including possible teleconnection of extremely heavy rain and droughts separated by far distances. This adjustment is strongly constrained geographically by climatological rainfall pattern, and most likely by the GHG warming induced sea surface temperature anomalies with unstable moister and warmer regions in the deep tropics getting more heavy rain, at the expense of nearby marginal convective zones in the tropics and stable dry zones in the subtropics.

Our results are generally consistent with so-called “the rich-getting-richer, poor-getting-poorer” paradigm for precipitation response under global warming..

Conclusion

This article has basically presented the results of one paper, which demonstrates consistency in model response of rainfall to doubling and tripling of CO2 in the atmosphere. In subsequent articles we will look at the underlying physics constraints, at time-series over recent decades and try to make some kind of assessment.

Articles in this Series

Impacts – II – GHG Emissions Projections: SRES and RCP

Impacts – IV – Temperature Projections and Probabilities

Impacts – V – Climate change is already causing worsening storms, floods and droughts

Impacts – X – Sea Level Rise 5 – Bangladesh

Impacts XI – Rainfall 1

References

A canonical response of precipitation characteristics to global warming from CMIP5 models, William K.-M. Lau, H.-T. Wu, & K.-M. Kim, GRL (2013) – free paper

Impacts XI – Rainfall 1

Posted in Atmospheric Physics, Climate Models, Impacts, Measurement on May 8, 2017| 10 Comments »

If we want to assess forecasts of floods, droughts and crop yields then we will need to know rainfall. We will also need to know temperature of course.

The forte of climate models is temperature. Rainfall is more problematic.

Before we get to model predictions about the future we need to review observations and the ability of models to reproduce them. Observations are also problematic – rainfall varies locally and over short durations. And historically we lacked effective observation systems in many locations and regions of the world, so data has to be pieced together and estimated from reanalysis.

Smith and his colleagues created a new rainfall dataset. Here is a comment from their 2012 paper:

Although many land regions have long precipitation records from gauges, there are spatial gaps in the sampling for undeveloped regions, areas with low populations, and over oceans. Since 1979 satellite data have been used to fill in those sampling gaps. Over longer periods gaps can only be filled using reconstructions or reanalyses..

Here are two views of the global precipitation data from a dataset which starts with the satellite era, that is, 1979 onwards – GPCP (Global Precipitation Climatology Project):

From Adler et al 2003

Figure 1

From Adler et al 2003

Figure 2

For historical data before satellites we only have rain gauge data. The GPCC dataset, explained in Becker et al 2013, shows the number of stations over time by region:

From Becker et al 2013

Figure 3- Click to expand

And the geographical distribution of rain gauge stations at different times:

From Becker et al 2013

Figure 4 – Click to expand

The IPCC compared the global trends over land from four different datasets over the last century and the last half-century:

From IPCC AR5 Ch. 2

Figure 5 – Click to expand

And the regional trends:

From IPCC AR5 Ch. 2

Figure 6 – Click to expand

The graphs for the annual change in rainfall, note the different scales for each region (as we would expect given the difference in average rainfall in different region):

From IPCC AR5 ch 2

Figure 7

We see that the decadal or half-decadal variation is much greater than any apparent long term trend. The trend data (as reviewed by the IPCC in figs 5 & 6) shows significant differences in the datasets but when we compare the time series it appears that the datasets match up better than indicated by the trend comparisons.

The data with the best historical coverage is 30ºN – 60ºN and the trend values for 1951-2000 (from different reconstructions) range from an annual increase of 1 to 1.5 mm/yr per decade (fig 6 / table 2.10 of IPCC report). This is against an absolute value of about 1000 mm/yr in this region (reading off the climatology in figure 2).

This is just me trying to put the trend data in perspective.

Models

Here is the IPCC AR5 chapter 9 on model comparisons to satellite-era rainfall observations. Top left is observations (basically the same dataset as figure 1 in this article over a slightly longer period with different colors) and bottom right is percentage error of model average with respect to observations:

From IPCC AR5 ch 9

Figure 8 – Click to expand

We can see that the average of all models has substantial errors on mean rainfall.

Articles in this Series

Impacts – II – GHG Emissions Projections: SRES and RCP

Impacts – IV – Temperature Projections and Probabilities

Impacts – V – Climate change is already causing worsening storms, floods and droughts

Impacts – X – Sea Level Rise 5 – Bangladesh

References

IPCC AR5 Chapter 2

Improved Reconstruction of Global Precipitation since 1900, Smith, Arken, Ren & Shen, Journal of Atmospheric and Oceanic Technology (2012)

The Version-2 Global Precipitation Climatology Project (GPCP) Monthly Precipitation Analysis (1979–Present), Adler et al, Journal of Hydrometeorology (2003)

A description of the global land-surface precipitation data products of the Global Precipitation Climatology Centre with sample applications including centennial (trend) analysis from 1901–present, A Becker, Earth Syst. Sci. Data (2013)

Impacts – X – Sea Level Rise 5 – Bangladesh

Posted in Impacts, Ocean Physics on May 4, 2017| 7 Comments »

FitzGerald et al 2008:

Sea-level rise (SLR) poses a particularly ominous threat because 10% of the world’s population (634 million people) lives in low-lying coastal regions within 10 m elevation of sea level (McGranahan et al. 2007). Much of this population resides in portions of 17 of the world’s 30 largest cities, including Mumbai, India; Shanghai, China; Jakarta, Indonesia; Bangkok, Thailand; London; and New York.

In the last article – Sinking Megacities – we saw that some of these cities are sinking due to ground water depletion. To those megacities, this is a much more serious threat than global sea level rise (probably why we see so many marches and protests about ground water depletion).

The paper continues:

..The potential loss of life in low-lying areas is even more graphically illustrated by the 1970 Bhola cyclone that traveled northward through the Bay of Bengal producing a 12-m-high wall of water that drowned a half million people in East Pakistan (now Bangladesh) (Garrison 2005).

In Bangladesh, storms and cyclones are much more of a threat than sea level rise. Here is Karim and Mimura (2008) listing the serious cyclones over the last 60 years:

From Karim and Mimura 2008

Figure 1 – Click to expand

There is an interesting World Bank Report from 2011. First on floods:

In an average year, nearly one quarter of Bangladesh is inundated, with more than three-fifths of land area at risk of floods of varying intensity (Ahmed and Mirza 2000). Every four or five years, a severe flood occurs during the monsoon season, submerging more than three-fifths of the land..

The most recent exceptional flood, which occurred in 2007, inundated 62,300 km² or 42 percent of total land area, causing 1,110 deaths and affecting 14 million people; 2.1 million ha of standing crop land were submerged, 85,000 houses completely destroyed, and 31,533 km of roads damaged. Estimated asset losses from this one event totaled US$1.1 billion (BWDB 2007).

Flooding in Bangladesh results from a complex set of factors, key among which are extremely low and flat topography, uncertain transboundary flow, heavy monsoon rainfall, and high vulnerability to tidal waves and congested drainage channels. Two-thirds of Bangladesh’s land area is less than 5 m above sea level. Each year, an average flow of 1,350 billion m³ of water from the GBM [Ganges, Brahmaputra, and Meghna] basin drains through the country.

From World Bank 2011

Figure 2

I recommend this World Bank report, very interesting, and you can see some idea of the costs of mitigating against floods. These problems are already present – floods are a regular occurrence, some mitigation has already taken place, and more mitigation continues.

I read the entire report and all I could find was that rising sea levels would exacerbate the problems already faced from storm surges: p.6:

Increase in ocean surface temperature and rising sea levels are likely to intensify cyclonic storm surges and further increase the depth and extent of storm surge induced coastal inundation.

However, the projections indicate that sea level rise is much less of a problem compared with possible increases in future storm surges and possible increases in future flooding. And compared with current storm surges and current flooding. We will look at floods and storm surges in future articles.

In the report it’s clear that floods and storms are already major problems. Sea level is harder to analyze. Trying to account for a sea level rise of 0.3m by 2050 when severe storm surges are already 5-10m is not going to make much of a difference. If we had accurate prediction of storm surges, to +/- 0.3m, then sea level rise of 0.3m should definitely be accounted for. But we don’t have anything like that kind of accuracy.

Well, they do some calculations of adaption against storm surges for projected changes up to 2050:

Under the baseline scenario, the adaptation costs total $2.46 billion. In a changing climate, the additional adaptation cost totals US$892 million.

In essence the question is “what is the storm surge for a once in a 10 year storm in 2050”? (I’m sure Bangladesh would really prefer to build protection against a once in 100 year storm). An extra $1bn for future problems, or a total of $3.5bn to cover existing and future problems, seems like money that would be very well spent, representing excellent value.

Nicholls and Cazenave (2010), in relation to the susceptible coastline of Asia and Africa, comment on adaption:

Many impact studies do not consider adaptation, and hence determine worst-case impacts. Yet, the history of the human relationship with the coast is one of an increasing capacity to adapt to adverse change. In addition, the world’s populated coasts became increasingly managed and engineered over the 20th century. The subsiding cities discussed above all remain protected to date, despite large relative SLR.

Analysis based on benefit-cost methods show that protection would be widespread as well-populated coastal areas have a high value and actual impacts would only be a small fraction of the potential impacts, even assuming high-SLR (>1 m/century) scenarios. This suggests that the common assumption of a widespread forced retreat from the shore in the face of SLR is not inevitable. In many densely populated coastal areas, communities advanced the coast seaward via land claim owing to the high value of land (e.g., Singapore).

Yet, protection often attracts new development in low lying areas, which may not be desirable, and coastal defense failures have occurred, such as New Orleans in 2005. Hence, we must choose between protection, accommodation, and planned retreat adaptation options. This choice is both technical and sociopolitical, addressing which measures are desirable, affordable, and sustainable in the long term. Adaptation remains a major uncertainty concerning the actual impacts of SLR.

In the World Bank 2011 report, in chapter 4, after their analysis on risks and costs of storm-induced inundations in 2050 resulting from projected higher cyclonic wind speeds and a projected increase in sea level of 0.27m, they comment, p.24:

As a cautionary note, it should be noted that this analysis did not address the out-migration from coastal zones that a rise in sea level and intensified cyclonic storm surges might induce.

In fact the cost data assumes population growth in the vulnerable regions.

Likewise, here is Hinkel et al (2014):

Coastal flood damages are expected to increase significantly during the 21st century as sea levels rise and socioeconomic development increases the number of people and value of assets in the coastal floodplain.

[Emphasis added].

This assumption bias creates an interpretation challenge. It would be useful to see notes to the effect: “If the population migrates away from this area due to the higher risk, instead the cost will be $X assuming a reduction of Y% in population in this region by 2050“. This extra item of data would create a useful contrast and I’m guessing that we would see impact assessments reduce by a factor of 5 or 10.

It is difficult to see realistic global sea level changes, even to the end of the century, having a big impact on Bangladesh compared with their current problems of annual flooding and frequent large storm surges. Of course, adding an extra 0.5m to the sea level doesn’t improve the situation, but it is an order of magnitude smaller than storm surges.

The adaption costs estimated by the World Bank to protect against storm surges (already required but at least a work in progress) seem moderate in value.

Lastly, I wasn’t able to find a detailed elevation map (with, say, 0.5m resolution), instead the ones I found graded the elevation with respect to sea level in fairly coarse steps. I’m sure the information exists but may be proprietary (in GIS data for example):

Figure 2 – Click to expand

I have to admit that I believed something like 25% of the Bangladesh population were around 1.0m or less above current sea level. This map says that the 0-3m area is quite small. If anyone does have a better resolution map I will post it up.

Articles in this Series

Impacts – II – GHG Emissions Projections: SRES and RCP

Impacts – IV – Temperature Projections and Probabilities

Impacts – V – Climate change is already causing worsening storms, floods and droughts

References

Coastal Impacts Due to Sea-Level Rise, Duncan M. FitzGerald et al, Annual Rev. Earth Planet. Sci. (2008)

Impacts of climate change and sea-level rise on cyclonic storm surge floods in Bangladesh, Mohammed Fazlul Karim & Nobuo Mimura, Global Environmental Change (2008) – free paper

The Cost of Adapting to Extreme Weather Events in a Changing Climate – Bangladesh, World Bank (2011) – free report

Sea-Level Rise and Its Impact on Coastal Zones, Robert J Nicholls & Anny Cazenave, Science (2010) – free paper

Coastal flood damage and adaptation costs under 21st century sea-level rise, Jochen Hinkel et al, PNAS (2014) – free paper

Impacts – IX – Sea Level 4 – Sinking Megacities

Posted in Impacts, Ocean Physics on May 2, 2017| 14 Comments »

In Impacts – VIII – Sea level 3 – USA I suggested this conclusion:

So the cost of sea level rise for 2100 in the US seems to be a close to zero cost problem.

Probably the provocative way I wrote the conclusion confused some people. I should have said that it was a very expensive problem. But that it wasn’t a problem that society should pay for, given that anyone moving to the coast since 2005 at the latest would have known that future sea level was considered to be a major problem. By 2100 the youngest people still living right on the sea front, who bought property there before 2005, would be at least 115 years old.

The idea is that “externalities” as economists call them should be paid by the creators of the problem, not the people that incur the problem. In this case, the “victims” are people who ignored the evidence and moved to the coast anyway. Are they still victims? That was my point.

Well, what about outside the US?

Some mega cities have huge problems. Here is Nicholls 2011:

Coastal areas constitute important habitats, and they contain a large and growing population, much of it located in economic centers such as London, New York, Tokyo, Shanghai, Mumbai, and Lagos. The range of coastal hazards includes climate-induced sea level rise, a long-term threat that demands broad response.

Global sea levels rose 17 cm through the twentieth century, and are likely to rise more rapidly through the twenty-first century when a rise of more than 1 m is possible.

In some locations, these changes may be exacerbated by

(1) increases in storminess due to climate change, although this scenario is less certain
(2) widespread human-induced subsidence due to ground fluid withdrawal from, and drainage of, susceptible soils, especially in deltas.

Subsidence?

Over the twentieth century, the parts of Tokyo and Osaka built on deltaic areas subsided up to 5 m and 3 m, respectively, a large part of Shanghai subsided up to 3 m, and Bangkok subsided up to 2 m.

This human-induced subsidence can be mitigated by stopping shallow, subsurface fluid withdrawals and managing water levels, but natural “background” rates of subsidence will continue, and RSLR will still exceed global trends in these areas. A combination of policies to mitigate subsidence has been instituted in the four delta cities mentioned above, combined with improved flood defenses and pumped drainage systems designed to avoid submergence and/ or frequent flooding.

In contrast, Jakarta and Metro Manila are subsiding significantly, with maximum subsidence of 4 m and 1 m to date, respectively (e.g., Rodolfo and Siringan, 2006; Ward et al., 2011), but little systematic policy response is in place in either city, and future flooding problems are anticipated.

Subsidence graphic:

From Nicholls 2011

Figure 1

To put these figures in context, sea level rise from 1900-2000 was about 0.2m and according to the latest IPCC report the forecast of sea level rise by 2100 might be around an additional 0.5m (for RCP 6.0, see earlier article). In the light of the idea that global society should pay for problems to people caused by global society, perhaps the problems of Shanghai, Bangkok and other sinking cities are not global problems?

Here is Wang et al from 2012:

Shanghai is low-lying, with an elevation of 3–4 m. A quarter of the area lies below 3 m. The city’s flood-control walls are currently more than 6 m high. However, given the trend of sea level rise and land subsidence, this is inadequate. Shanghai is frequently affected by extreme tropical storm surges. The risk of flooding from overtopping is considerable..

..From 1921 to 1965, the average cumulative subsidence of the city center was 1.76 m, with a maximum of 2.63 m. From 1966 to 1985, a monitoring network was established and subsidence was mitigated through artificial recharge. Land subsidence was stabilized at an average of 0.9 mm/year. As a result of rapid urban development and large-scale construction projects between 1986 and 1997, subsidence of the downtown area increased rapidly, at an average rate of 10.2 mm/year..

..In 2100, sea level rise and land subsidence will be far greater than before. Sea level rise is estimated to be 43 cm, while land subsidence is estimated to be 3–229 cm, and neotectonic subsidence is estimated to be 14 cm. Flooding will be severe in 2100 (Fig. 8).

[Note I changed the data in the last paragraph cited to round numbers in cm from their values quoted to 0.01cm – for example, 43cm instead of the paper’s values of 43.31 etc].

So for Shanghai at least global sea level rise is not really the problem.

Given that I don’t pay much attention to media outlets I probably missed the big Marches against Ground Water Depletion Slightly Accentuating Global Warming’s Sea Level Rise in Threatened Megacities.

As with the USA data the question of increased storm surges accentuating global sea level rise is still on the agenda (i.e., has not yet been discussed).

References

Planning for the impacts of sea level rise, RJ Nicholls, Oceanography (2011)

Evaluation of the combined risk of sea level rise, land subsidence, and storm surges on the coastal areas of Shanghai, China, Jun Wang, Wei Gao, Shiyuan Xu & Lizhong Yu, Climatic Change (2012)

Impacts – VIII – Sea level 3 – USA

Posted in Commentary, Impacts on February 19, 2017| 101 Comments »

In Parts VI and VII we looked at past and projected sea level rise. It is clear that the sea level has risen over the last hundred years, and it’s clear that with more warming sea level will rise some more. The uncertainties (given a specific global temperature increase) are more around how much more ice will melt than how much the ocean will expand (warmer water expands). Future sea level rise will clearly affect some people in the future, but very differently in different countries and regions. This article considers the US.

A month or two ago, via a link from a blog, I found a paper which revised upwards a current calculation (or average of such calculations) of damage due to sea level rise in 2100 in the US. Unfortunately I can’t find the paper, but essentially the idea was people would continue moving to the ocean in ever increasing numbers, and combined with possible 1m+ sea level rise (see Part VI & VII) the cost in the US would be around $1TR (I can’t remember the details but my memory tells me this paper concluded costs were 3x previous calculations due to this ever increasing population move to coastal areas – in any case, the exact numbers aren’t important).

Two examples that I could find (on global movement of people rather than just in the US), Nicholls 2011:

..This threatened population is growing significantly (McGranahan et al., 2007), and it will almost certainly increase in the coming decades, especially if the strong tendency for coastward migration continues..

And Anthoff et al 2010

Fifthly, building on the fourth point, FUND assumes that the pattern of coastal development persists and attracts future development. However, major disasters such as the landfall of hurricanes could trigger coastal abandonment, and hence have a profound influence on society’s future choices concerning coastal protection as the pattern of coastal occupancy might change radically.

A cycle of decline in some coastal areas is not inconceivable, especially in future worlds where capital is highly mobile and collective action is weaker. As the issue of sea-level rise is so widely known, disinvestment from coastal areas may even be triggered without disasters..

I was struck by the “trillion dollar problem” paper and the general issues highlighted in other papers. The future cost of sea level rise in the US is not just bad, it’s extremely expensive because people will keep moving to the ocean.

Why are people moving to the coast?

So here is an obvious take on the subject that doesn’t need an IAM (integrated assessment model).. Perhaps lots of people missed the IPCC TAR (third assessment report) in 2001. Perhaps anthropogenic global warming fears had not reached a lot of the population. Maybe it didn’t get a lot of media coverage. But surely no could have missed Al Gore’s movie. I mean, I missed it from choice, but how could anyone in rich countries not know about the discussion?

So anyone since 2006 (arbitrary line in the sand) who bought a house that is susceptible to sea level rise is responsible for their own loss that they incur around 2100. That is, if the worst fears about sea level rise play out, combined with more extreme storms (subject of a future article) which create larger ocean storm surges, their house won’t be worth much in 2100.

Now, barring large increases in life expectancy, anyone who bought a house in 2005 will almost certainly be dead in 2100. There will be a few unlucky centenarians.

Think of it as an estate tax. People who have expensive ocean-front houses will pass on their now worthless house to their children or grandchildren. Some people love the idea of estate taxes – in that case you have a positive. Some people hate the idea of estate taxes – in that case strike it up as a negative. And, drawing a long bow here, I suspect a positive correlation between concern about climate change and belief in the positive nature of estate taxes, so possibly it’s a win-win for many people.

Now onto infrastructure.

From time to time I’ve had to look at depreciation and official asset life for different kinds of infrastructure and I can’t remember seeing one for 100 years. 50 years maybe for civil structures. I’m definitely not an expert. That said, even if the “official depreciation” gives something a life of 50 years, much is still being used 150 years later – buildings, railways, and so on.

So some infrastructure very close to the ocean might have to be abandoned. But it will have had 100 years of useful life and that is pretty good in public accounting terms.

Why is anyone building housing, roads, power stations, public buildings, railways and airports in the US in locations that will possibly be affected by sea level rise in 2100? Maybe no one is.

So the cost of sea level rise for 2100 in the US seems to be a close to zero cost problem.

These days, if a particular area is recognized as a flood plain people are discouraged from building on it and no public infrastructure gets built there. It’s just common sense.

Some parts of New Orleans were already below sea level when Hurricane Katrina hit. Following that disaster, lots of people moved out of New Orleans to a safer suburb. Lots of people stayed. Their problems will surely get worse with a warmer climate and a higher sea level (and also if storms gets stronger – subject of a future article). But they already had a problem. Infrastructure was at or below sea level and sufficient care was not taken of their coastal defences.

A major problem that happens overnight, or over a year, is difficult to deal with. A problem 100 years from now that affects a tiny percentage of the land area of a country, even with a large percentage (relatively speaking) of population living there today, is a minor problem.

Perhaps the costs of recreating current threatened infrastructure a small distance inland are very high, and the existing infrastructure would in fact have lasted more than 100 years. In that case, people who believe Keynesian economics might find the economic stimulus to be a positive. People who don’t think Keynesian economics does anything (no multiplier effect) except increase taxes, or divert productive resources into less productive resources will find it be a negative. Once again, drawing a long bow, I see a correlation between people more concerned about climate change also being more likely to find Keynesian economics a positive. Perhaps again, there is a win-win.

In summary, given the huge length of time to prepare for it, US sea level rise seems like a minor planning inconvenience combined with an estate tax.

Articles in this Series

Impacts – II – GHG Emissions Projections: SRES and RCP

Impacts – IV – Temperature Projections and Probabilities

Impacts – V – Climate change is already causing worsening storms, floods and droughts

References

Planning for the impacts of sea level rise, RJ Nicholls, Oceanography (2011)

The economic impact of substantial sea-level rise, David Anthoff et al, Mitig Adapt Strateg Glob Change (2010)

Impacts – VI – Sea Level Rise 1

Posted in Impacts, Ocean Physics on February 16, 2017| 206 Comments »

In Part V we looked at the IPCC, an outlier organization, that claimed floods, droughts and storms had not changed in a measurable way globally in the last 50 -100 years (of course, some regions have seen increases and some have seen decreases, some decades have been bad, some decades have been good).

This puts them at a disadvantage compared with the overwhelming mass of NGOs, environmental groups, media outlets and various government departments who claim the opposite, but the contrarian in me found their research too interesting to ignore. Plus, they come with references to papers in respectable journals.

We haven’t looked at future projections of these events as yet. Whenever there are competing effects to create a result we can expect it to be difficult to calculate future effects. In contrast, one climate effect that we can be sure about is sea level. If the world warms, as it surely will with more GHGs, we can expect sea level to rise.

In my own mental list of “bad stuff to happen”, I had sea level rise as an obvious #1 or #2. But ideas and opinions need to be challenged and I had not really investigated the impacts.

The world is a big place and rising sea level will have different impacts in different places. Generally the media presentation on sea level is unrelentingly negative, probably following the impact of the impressive 2004 documentary directed by Roland Emmerich, and the dramatized adaption by Al Gore in 2006 (directed by Davis Guggenheim).

Let’s start by looking at some sea level basics.

Like everything else related to climate, getting an accurate global dataset on sea level is difficult – especially when we want consistency over decades.

The world is a big place and past climatological measurements were mostly focused on collecting local weather data for the country or region in question. Satellites started measuring climate globally in the late 1970s, but satellites for sea level and mass balance didn’t begin measurements until 10-20 years ago. So, climate scientists attempt to piece together disparate data systems, to reconcile them, and to match up the results with what climate models calculate – call it “a sea level budget”.

“The budget” means balancing two sides of the equation:

how has sea level changed year by year and decade by decade?
what contributions to sea level do we calculate from the effects of warming climate?

Components of Sea Level Rise

If we imagine sea level as the level in a large bathtub it is relatively simple conceptually. As the ocean warms the level rises for two reasons:

warmer water expands (increasing the volume of existing mass)
ice melts (adding mass)

The “material properties” of water are well known and not in doubt. With lots of measurements of ocean temperature around the globe we can be relatively sure of the expansion. Ocean temperature has increasing coverage over the last 100 years, especially since the Argo project that started a little more than 10 years ago. But if we go back 30 years we have a lot less measurements and usually only at the surface. If we go back 100 years we have less again. So there are questions and uncertainties.

Arctic ice melting has no impact on sea level because it is already floating. Water or ice that is already floating doesn’t change the sea level by melting/freezing. Ice on a continent that melts and runs into the ocean increases sea level due to increasing the mass. Some Antarctic ice shelves are in the ocean but are part of the Antarctic ice sheet that is supported by the continent of Antarctica – melt these ice sheets and they will add to ocean level.

Sea level over the last 100 years has increased by about 0.20m (about 8 inches, if we use advanced US units).

To put it into one perspective, 20,000 years ago the sea level was about 120m lower than today – this was the end of the last ice age. About 130,000 years ago the sea level was a few meters higher (no one is certain of the exact figure). This was the time of the last “interglacial” (called the Eemian interglacial).

If we melted all of Greenland’s ice sheet we would see a further 7m rise from today, and Greenland and Antarctica together would lead to a 70m rise. Over millennia (but not a century), the complete Greenland ice sheet melting is a possibility, but Antarctica is not (at around -30ºC, it is a very long way below freezing).

Complications

Why not use tide gauges to measure sea level rise? Some have been around for 100 years and a few have been around for 200 years.

There aren’t many tide gauges going back a long time, and anyway in many places the ground is moving relative to the ocean. Take Scandinavia. At the end of the last ice age Stockholm was buried under perhaps 2km of ice. No wonder Scandinavians today appear so cheerful – life is all about contrasts. As the ice melted, the load on the ground was removed and it is “springing back” into a pre-glacial position. So in many places around the globe the land is moving vertically relative to sea level.

In Nedre Gavle, Sweden, the land is moving up twice as fast as the average global sea level rise (so relative sea level is falling). In Galveston, Texas the land is moving down faster than sea level rise (more than doubling apparent sea level rise).

That is the first complication.

The second complication is due to wind and local density from salinity changes. So as an example, picture a constant sea level but Pacific winds change from W->E to E->W. The water will “pile up” in the west instead of the east, due to the force of the wind. Relative sea level will increase in the west and decrease in the east. Likewise, if the local density changes from melting ice (or ocean currents with different salinity) we can adjust the local sea level relative to the reference.

Here is AR5, chapter 3, p. 288:

Large-scale spatial patterns of sea level change are known to high precision only since 1993, when satellite altimetry became available.

These data have shown a persistent pattern of change since the early 1990s in the Pacific, with rates of rise in the Warm Pool of the western Pacific up to three times larger than those for GMSL, while rates over much of the eastern Pacific are near zero or negative.

The increasing sea level in the Warm Pool started shortly before the launch of TOPEX/Poseidon, and is caused by an intensification of the trade winds since the late 1980s that may be related to the Pacific Decadal Oscillation (PDO).

The lower rate of sea level rise since 1993 along the western coast of the United States has also been attributed to changes in the wind stress curl over the North Pacific associated with the PDO..

Measuring Systems

We can find a little about the new satellite systems in IPCC, AR5, chapter 3, p. 286:

Satellite radar altimeters in the 1970s and 1980s made the first nearly global observations of sea level, but these early measurements were highly uncertain and of short duration. The first precise record began with the launch of TOPEX/Poseidon (T/P) in 1992. This satellite and its successors (Jason-1, Jason-2) have provided continuous measurements of sea level variability at 10-day intervals between approximately ±66° latitude. Additional altimeters in different orbits (ERS-1, ERS-2, Envisat, Geosat Follow-on) have allowed for measurements up to ±82° latitude and at different temporal sampling (3 to 35 days), although these measurements are not as accurate as those from the T/P and Jason satellites.

Unlike tide gauges, altimetry measures sea level relative to a geodetic reference frame (classically a reference ellipsoid that coincides with the mean shape of the Earth, defined within a globally realized terrestrial reference frame) and thus will not be affected by VLM, although a small correction that depends on the area covered by the satellite (~0.3 mm yr–1) must be added to account for the change in location of the ocean bottom due to GIA relative to the reference frame of the satellite (Peltier, 2001; see also Section 13.1.2).

Tide gauges and satellite altimetry measure the combined effect of ocean warming and mass changes on ocean volume. Although variations in the density related to upper-ocean salinity changes cause regional changes in sea level, when globally averaged their effect on sea level rise is an order of magnitude or more smaller than thermal effects (Lowe and Gregory, 2006).

The thermal contribution to sea level can be calculated from in situ temperature measurements (Section 3.2). It has only been possible to directly measure the mass component of sea level since the launch of the Gravity Recovery and Climate Experiment (GRACE) in 2002 (Chambers et al., 2004). Before that, estimates were based either on estimates of glacier and ice sheet mass losses or using residuals between sea level measured by altimetry or tide gauges and estimates of the thermosteric component (e.g., Willis et al., 2004; Domingues et al., 2008), which allowed for the estimation of seasonal and interannual variations as well. GIA also causes a gravitational signal in GRACE data that must be removed in order to determine present-day mass changes; this correction is of the same order of magnitude as the expected trend and is still uncertain at the 30% level (Chambers et al., 2010).

The GRACE satellite lets us see how much ice has melted into the ocean. It’s not easy to calculate this otherwise.

The fourth assessment report from the IPCC in 2007 reported that sea level rise from the Antarctic ice sheet for the previous decade was between -0.3mm/yr and +0.5mm/yr. That is, without the new satellite measurements, it was very difficult to confirm whether Antarctica had been gaining or losing ice.

Historical Sea Level Rise

From AR5, chapter 3, p. 287:

From AR5, chapter 3

Figure 1 – Click to expand

The top left graph shows that various researchers are fairly close in their calculations of overall sea level rise over the past 130 years
The bottom left graph shows that over the last 40 years the impact of melting ice has been more important than the expansion of a warmer ocean (“thermosteric component” = the effect of a warmer ocean expanding)
The bottom right graph shows that over the last 7 years the measurements are consistent – satellite measurement of sea level change matches the sum of mass loss (melting ice) plus an expanding ocean (the measurements from Argo turned into sea level rise).

This gives us the mean sea level. Remember that local winds, ocean currents and changes in salinity can change this trend locally.

Many people have written about the recent accelerating trends in sea level rise. Here is AR5 again, with a graph of the 18-year trend at each point in time. We can see that different researchers reach different conclusions and that the warming period in the first half of the 20th century created sea level rise comparable to today:

From AR5, chapter 3

The conclusion in AR5:

It is virtually certain that globally averaged sea level has risen over the 20th century, with a very likely mean rate between 1900 and 2010 of 1.7 [1.5 to 1.9] mm/yr and 3.2 [2.8 and 3.6] mm/yr between 1993 and 2010.

This assessment is based on high agreement among multiple studies using different methods, and from independent observing systems (tide gauges and altimetry) since 1993.

It is likely that a rate comparable to that since 1993 occurred between 1920 and 1950, possibly due to a multi-decadal climate variation, as individual tide gauges around the world and all reconstructions of GMSL show increased rates of sea level rise during this period.

Forecast Future Sea Level Rise

AR5, chapter 13 is the place to find predictions of the future on sea level, p. 1140:

For the period 2081–2100, compared to 1986–2005, global mean sea level rise is likely (medium confidence) to be in the 5 to 95% range of projections from process-based models, which give:

0.26 to 0.55 m for RCP2.6

0.32 to 0.63 m for RCP4.5

0.33 to 0.63 m for RCP6.0

0.45 to 0.82 m for RCP8.5

For RCP8.5, the rise by 2100 is 0.52 to 0.98 m..

We have considered the evidence for higher projections and have concluded that there is currently insufficient evidence to evaluate the probability of specific levels above the assessed likely range. Based on current understanding, only the collapse of marine-based sectors of the Antarctic ice sheet, if initiated, could cause global mean sea level to rise substantially above the likely range during the 21st century.

This potential additional contribution cannot be precisely quantified but there is medium confidence that it would not exceed several tenths of a meter of sea level rise during the 21st century.

I highlighted RCP6.0 as this seems to correspond to past development pathways with little CO2 mitigation policies. No one knows the future, this is just my pick, barring major changes from the recent past.

In the next article we will consider impacts of future sea level rise in various regions.

Articles in this Series

Impacts – II – GHG Emissions Projections: SRES and RCP

Impacts – IV – Temperature Projections and Probabilities

Impacts – V – Climate change is already causing worsening storms, floods and droughts

References

Observations: Oceanic Climate Change and Sea Level. In: Climate Change 2007: The Physical Science Basis. Contribution of Working Group I to the Fourth Assessment Report of the Intergovernmental Panel on Climate Change, NL Bindoff et al (2007)

Observations: Ocean. In: Climate Change 2013: The Physical Science Basis. Contribution of Working Group I to the Fifth Assessment Report of the Intergovernmental Panel on Climate Change, M Rhein et al (2013)

Sea Level Change. In: Climate Change 2013: The Physical Science Basis. Contribution of Working Group I to the Fifth Assessment Report of the Intergovernmental Panel on Climate Change, JA Church et al (2013)

Impacts – V – Climate change is already causing worsening storms, floods and droughts

Posted in Climate History, Impacts on February 12, 2017| 52 Comments »

I generally try and avoid the media as much as possible (although the 2016 Circus did suck me in) but it’s still impossible to miss claims like the following:

Climate change is already causing worsening storms, floods and droughts

Before looking at predictions for the future I thought it was worth reviewing this claim, seeing as it is so prevalent and is presented as being the current consensus of climate science.

Droughts

SREX 2012, p. 171:

There is medium confidence that since the 1950s some regions of the world have experienced more intense and longer droughts (e.g., southern Europe, west Africa) but also opposite trends exist in other regions (e.g., central North America, northwestern Australia).

The report cites Sheffield and Wood 2008 who show graphs on a variety of drought metrics from around the world over the last 50 years – click to enlarge:

From Sheffield & Wood 2008

Figure 1 – Click to enlarge

The results above were calculated from models based on available meteorological data. According to their analysis some places have experienced more droughts, and other places less droughts. Because they are based on models we can expect that alternative researchers may produce different results.

AR5, published a year after SREX, says, chapter 2, p. 214-215:

Because drought is a complex variable and can at best be incompletely represented by commonly used drought indices, discrepancies in the interpretation of changes can result. For example, Sheffield and Wood (2008) found decreasing trends in the duration, intensity and severity of drought globally. Conversely, Dai (2011a,b) found a general global increase in drought, although with substantial regional variation and individual events dominating trend signatures in some regions (e.g., the 1970s prolonged Sahel drought and the 1930s drought in the USA and Canadian Prairies). Studies subsequent to these continue to provide somewhat different conclusions on trends in global droughts and/ or dryness since the middle of the 20th century (Sheffield et al., 2012; Dai, 2013; Donat et al., 2013c; van der Schrier et al., 2013)..

..In summary, the current assessment concludes that there is not enough evidence at present to suggest more than low confidence in a global-scale observed trend in drought or dryness (lack of rainfall) since the middle of the 20th century, owing to lack of direct observations, geographical inconsistencies in the trends, and dependencies of inferred trends on the index choice.

Based on updated studies, AR4 conclusions regarding global increasing trends in drought since the 1970s were probably overstated.

The paper by Dai is Drought under global warming: a review, A Dai, Climate Change (2011) – for some reason I am unable to access it.

A later paper in Nature, Trenberth et al 2013 (including both Sheffield and Dai as co-authors) said:

Two recent papers looked at the question of whether large-scale drought has been increasing under climate change. A study in Nature by Sheffield et al entitled ‘Little change in global drought over the past 60 years’ was published at almost the same time that ‘Increasing drought under global warming in observations and models’ by Dai appeared in Nature Climate Change (published online in August 2012). How can two research groups arrive at such seemingly contradictory conclusions?

Another later paper on droughts, Orlowski & Seneviratne 2013, likewise shows overwhelming evidence of more droughts – click to enlarge:

From Orlowsky & Seneviratne 2013

Figure 2 – Click to enlarge

Floods

SREX 2012, p. 177:

Overall, there is low confidence (due to limited evidence) that anthropogenic climate change has affected the magnitude and frequency of floods, though it has detectably influenced several components of the hydrological cycle, such as precipitation and snowmelt, that may impact flood trends. The assessment of causes behind the changes in floods is inherently complex and difficult.

AR5, Chapter 2, p. 214:

AR5 WGII assesses floods in regional detail accounting for the fact that trends in floods are strongly influenced by changes in river management (see also Section 2.5.2). Although the most evident flood trends appear to be in northern high latitudes, where observed warming trends have been largest, in some regions no evidence of a trend in extreme flooding has been found, for example, over Russia based on daily river discharge (Shiklomanov et al., 2007).

Other studies for Europe (Hannaford and Marsh, 2008; Renard et al., 2008; Petrow and Merz, 2009; Stahl et al., 2010) and Asia (Jiang et al., 2008; Delgado et al., 2010) show evidence for upward, downward or no trend in the magnitude and frequency of floods, so that there is currently no clear and widespread evidence for observed changes in flooding except for the earlier spring flow in snow-dominated regions (Seneviratne et al., 2012).

In summary, there continues to be a lack of evidence and thus low confidence regarding the sign or trend in the magnitude and/or frequency of floods on a global scale.

[Note: the text in the bottom line cited says: “..regarding the sign of trend in the magnitude..” which I assume is a typo, and so I changed of into or]

Storms

SREX, p. 159:

Detection of trends in tropical cyclone metrics such as frequency, intensity, and duration remains a significant challenge..

..Natural variability combined with uncertainties in the historical data makes it difficult to detect trends in tropical cyclone activity. There have been no significant trends observed in global tropical cyclone frequency records, including over the present 40-year period of satellite observations (e.g., Webster et al., 2005). Regional trends in tropical cyclone frequency have been identified in the North Atlantic, but the fidelity of these trends is debated (Holland and Webster, 2007; Landsea, 2007; Mann et al., 2007a). Different methods for estimating undercounts in the earlier part of the North Atlantic tropical cyclone record provide mixed conclusions (Chang and Guo, 2007; Mann et al., 2007b; Kunkel et al., 2008; Vecchi and Knutson, 2008).

Regional trends have not been detected in other oceans (Chan and Xu, 2009; Kubota and Chan, 2009; Callaghan and Power, 2011). It thus remains uncertain whether any observed increases in tropical cyclone frequency on time scales longer than about 40 years are robust, after accounting for past changes in observing capabilities (Knutson et al., 2010)..

..Time series of power dissipation, an aggregate compound of tropical cyclone frequency, duration, and intensity that measures total energy consumption by tropical cyclones, show upward trends in the North Atlantic and weaker upward trends in the western North Pacific over the past 25 years (Emanuel, 2007), but interpretation of longer-term trends in this quantity is again constrained by data quality concerns.

The variability and trend of power dissipation can be related to SST and other local factors such as tropopause temperature and vertical wind shear (Emanuel, 2007), but it is a current topic of debate whether local SST or the difference between local SST and mean tropical SST is the more physically relevant metric (Swanson, 2008).

The distinction is an important one when making projections of changes in power dissipation based on projections of SST changes, particularly in the tropical Atlantic where SST has been increasing more rapidly than in the tropics as a whole (Vecchi et al., 2008). Accumulated cyclone energy, which is an integrated metric analogous to power dissipation, has been declining globally since reaching a high point in 2005, and is presently at a 40- year low point (Maue, 2009). The present period of quiescence, as well as the period of heightened activity leading up to the high point in 2005, does not clearly represent substantial departures from past variability (Maue, 2009)..

..The present assessment regarding observed trends in tropical cyclone activity is essentially identical to the WMO assessment (Knutson et al., 2010): there is low confidence that any observed long-term (i.e., 40 years or more) increases in tropical cyclone activity are robust, after accounting for past changes in observing capabilities.

AR5, Chapter 2, p. 216:

AR4 concluded that it was likely that an increasing trend had occurred in intense tropical cyclone activity since 1970 in some regions but that there was no clear trend in the annual numbers of tropical cyclones. Subsequent assessments, including SREX and more recent literature indicate that it is difficult to draw firm conclusions with respect to the confidence levels associated with observed trends prior to the satellite era and in ocean basins outside of the North Atlantic.

Lots more tropical storms:

From AR5, wg I

Figure 3

Note that a more important metric than “how many?” is “how severe?” or a combination of both.

And for extra-tropical storms (i.e. outside the tropics), SREX p. 166:

In summary it is likely that there has been a poleward shift in the main Northern and Southern Hemisphere extratropical storm tracks during the last 50 years. There is medium confidence in an anthropogenic influence on this observed poleward shift. It has not formally been attributed.

There is low confidence in past changes in regional intensity.

And AR5, chapter 2, p. 217 & 220:

Some studies show an increase in intensity and number of extreme Atlantic cyclones (Paciorek et al., 2002; Lehmann et al., 2011) while others show opposite trends in eastern Pacific and North America (Gulev et al., 2001). Comparisons between studies are hampered because of the sensitivities in identification schemes and/ or different definitions for extreme cyclones (Ulbrich et al., 2009; Neu et al., 2012). The fidelity of research findings also rests largely with the underlying reanalyses products that are used..

..In summary, confidence in large scale changes in the intensity of extreme extratropical cyclones since 1900 is low. There is also low confidence for a clear trend in storminess proxies over the last century due to inconsistencies between studies or lack of long-term data in some parts of the world (particularly in the SH). Likewise, confidence in trends in extreme winds is low, owing to quality and consistency issues with analysed data.

Discussion

The IPCC SREX and AR5 reports were published in 2012 and 2013 respectively. There will be new research published since these reports analyzing the same data and possibly reaching different conclusions. When you have large decadal variability in poorly observed data with a small or non-existent trend then inevitably different groups will be able to reach different conclusions on these trends. And if you focus on specific regions you can demonstrate a clear and unmistakeable trend.

If you are looking for a soundbite just pick the right region.

The last 100 years have seen global warming. As this blog has made clear from the physics, more GHGs (all other things remaining equal) result in more warming. What proportion of the last 100 years is intrinsic climate variability vs the anthropogenic GHG proportion I have no idea.

The last century has seen no clear globally averaged change in floods, droughts or storms – as best as we can tell with very incomplete observing systems. Of course, some regions have definitely seen more, and some regions have definitely seen less. Whether this is different from the period from 1800-1900 or from 1700-1800 no one knows. Perhaps floods, droughts and tropical storms increased globally from 1700-1900. Perhaps they decreased. Perhaps the last 100 years have seen more variability. Perhaps not. (And in recognition of Poe’s law, I note that a few statements within the article presenting graphs did say the opposite of the graphs presented).

Articles in this Series

Impacts – II – GHG Emissions Projections: SRES and RCP

Impacts – IV – Temperature Projections and Probabilities

References

SREX = Managing the Risks of Extreme Events and Disasters to Advance Climate Change Adaptation Special Report, IPCC (2012)

Observations: Atmosphere and Surface. Chapter 2 of Working Group I to AR5, DL Hartmann et al (2013)

Global Trends and Variability in Soil Moisture and Drought Characteristics, 1950–2000, from Observation-Driven Simulations of the Terrestrial Hydrologic Cycle, Justin Sheffield & Eric Wood, Journal of Climate (2008) – free paper

Global warming and changes in drought, Kevin E Trenberth et al, Nature (2013) – free paper

Elusive drought: uncertainty in observed trends and short- and long-term CMIP5 projections, B Orlowsky & SI Seneviratne, Hydrology and Earth System Sciences (2013) – free paper

Impacts – IV – Temperature Projections and Probabilities

Posted in Climate Models, Impacts on February 7, 2017| 38 Comments »

In Impacts – II – GHG Emissions Projections: SRES and RCP we looked at projections of emissions under various scenarios with the resulting CO2 (and other GHG) concentrations and resulting radiative forcing.

Why do we need these scenarios? Because even if climate models were perfect and could accurately calculate the temperature 100 years from now, we wouldn’t know how much “anthropogenic CO2” (and other GHGs) would have been emitted by that time. The scenarios allow climate modelers to produce temperature (and other climate variable) projections on the basis of each of these scenarios.

The IPCC AR5 (fifth assessment report) from 2013 says (chapter 12, p. 1031):

Global mean temperatures will continue to rise over the 21st century if greenhouse gas (GHG) emissions continue unabated.

Under the assumptions of the concentration-driven RCPs, global mean surface temperatures for 2081–2100, relative to 1986–2005 will likely be in the 5 to 95% range of the CMIP5 models:

0.3°C to 1.7°C (RCP2.6)

1.1°C to 2.6°C (RCP4.5)

1.4°C to 3.1°C (RCP6.0)

2.6°C to 4.8°C (RCP8.5)

Global temperatures averaged over the period 2081– 2100 are projected to likely exceed 1.5°C above 1850-1900 for RCP4.5, RCP6.0 and RCP8.5 (high confidence), are likely to exceed 2°C above 1850-1900 for RCP6.0 and RCP8.5 (high confidence) and are more likely than not to exceed 2°C for RCP4.5 (medium confidence). Temperature change above 2°C under RCP2.6 is unlikely (medium confidence). Warming above 4°C by 2081–2100 is unlikely in all RCPs (high confidence) except for RCP8.5, where it is about as likely as not (medium confidence).

I commented in Part II that RCP8.5 seemed to be a scenario that didn’t match up with the last 40-50 years of development. Of course, the various scenario developers give their caveats, for example, Riahi et al 2007:

Given the large number of variables and their interdependencies, we are of the opinion that it is impossible to assign objective likelihoods or probabilities to emissions scenarios. We have also not attempted to assign any subjective likelihoods to the scenarios either. The purpose of the scenarios presented in this Special Issue is, rather, to span the range of uncertainty without an assessment of likely, preferable, or desirable future developments..

Readers should exercise their own judgment on the plausibility of above scenario ‘storylines’..

To me RCP6.0 seems a more likely future (compared with RCP8.5) in a world that doesn’t have any significant attempt to tackle CO2 emissions. That is, no major change in climate policy to today’s world, but similar economic and population development (note 1).

Here is the graph of projected temperature anomalies for the different scenarios. :

From AR5, chapter 12

Figure 1

That graph is hard to make out for 2100, here is the table of corresponding data. I highlighted RCP6.0 in 2100 – you can click to enlarge the table:

Figure 2 – Click to expand

Probabilities and Lists

The table above has a “1 std deviation” and a 5%-95% distribution. The graph (which has the same source data) has shading to indicate 5%-95% of models for each RCP scenario.

These have no relation to real probability distributions. That is, the range of 5-95% for RCP6.0 doesn’t equate to: “the probability is 90% likely that the average temperature 2080-2100 will be 1.4-3.1ºC higher than the 1986-2005 average”.

A number of climate models are used to produce simulations and the results from these “ensembles” are sometimes pressed into “probability service”. For some concept background on ensembles read Ensemble Forecasting.

Here is IPCC AR5 chapter 12:

Ensembles like CMIP5 do not represent a systematically sampled family of models but rely on self-selection by the modelling groups.

This opportunistic nature of MMEs [multi-model ensembles] has been discussed, for example, in Tebaldi and Knutti (2007) and Knutti et al. (2010a). These ensembles are therefore not designed to explore uncertainty in a coordinated manner, and the range of their results cannot be straightforwardly interpreted as an exhaustive range of plausible outcomes, even if some studies have shown how they appear to behave as well calibrated probabilistic forecasts for some large-scale quantities. Other studies have argued instead that the tail of distributions is by construction undersampled.

In general, the difficulty in producing quantitative estimates of uncertainty based on multiple model output originates in their peculiarities as a statistical sample, neither random nor systematic, with possible dependencies among the members and of spurious nature, that is, often counting among their members models with different degrees of complexities (different number of processes explicitly represented or parameterized) even within the category of general circulation models..

..In summary, there does not exist at present a single agreed on and robust formal methodology to deliver uncertainty quantification estimates of future changes in all climate variables. As a consequence, in this chapter, statements using the calibrated uncertainty language are a result of the expert judgement of the authors, combining assessed literature results with an evaluation of models demonstrated ability (or lack thereof) in simulating the relevant processes (see Chapter 9) and model consensus (or lack thereof) over future projections. In some cases when a significant relation is detected between model performance and reliability of its future projections, some models (or a particular parametric configuration) may be excluded but in general it remains an open research question to find significant connections of this kind that justify some form of weighting across the ensemble of models and produce aggregated future projections that are significantly different from straightforward one model–one vote ensemble results. Therefore, most of the analyses performed for this chapter make use of all available models in the ensembles, with equal weight given to each of them unless otherwise stated.

And from one of the papers cited in that section of chapter 12, Jackson et al 2008:

In global climate models (GCMs), unresolved physical processes are included through simplified representations referred to as parameterizations.

Parameterizations typically contain one or more adjustable phenomenological parameters. Parameter values can be estimated directly from theory or observations or by “tuning” the models by comparing model simulations to the climate record. Because of the large number of parameters in comprehensive GCMs, a thorough tuning effort that includes interactions between multiple parameters can be very computationally expensive. Models may have compensating errors, where errors in one parameterization compensate for errors in other parameterizations to produce a realistic climate simulation (Wang 2007; Golaz et al. 2007; Min et al. 2007; Murphy et al. 2007).

The risk is that, when moving to a new climate regime (e.g., increased greenhouse gases), the errors may no longer compensate. This leads to uncertainty in climate change predictions. The known range of uncertainty of many parameters allows a wide variance of the resulting simulated climate (Murphy et al. 2004; Stainforth et al. 2005; M. Collins et al. 2006). The persistent scatter in the sensitivities of models from different modeling groups, despite the effort represented by the approximately four generations of modeling improvements, suggests that uncertainty in climate prediction may depend on underconstrained details and that we should not expect convergence anytime soon.

Stainforth et al 2005 (referenced in the quote above) tried much larger ensembles of coarser resolution climate models, and was discussed in the comments of Models, On – and Off – the Catwalk – Part Four – Tuning & the Magic Behind the Scenes. Rowlands et al 2012 is similar in approach and was discussed in Natural Variability and Chaos – Five – Why Should Observations match Models?

The way I read the IPCC reports and various papers is that clearly the projections are not a probability distribution. Then the data gets inevitably gets used as a de facto probability distribution.

Conclusion

“All models are wrong but some are useful” as George Box said, actually in a quite unrelated field (i.e., not climate). But it’s a good saying.

Many people who describe themselves as “lukewarmers” believe that climate sensitivity as characterized by the IPCC is too high and the real climate has a lower sensitivity. I have no idea.

Models may be wrong, but I don’t have an alternative model to provide. And therefore, given that they represent climate better than any current alternative, their results are useful.

We can’t currently create a real probability distribution from a set of temperature prediction results (assuming a given emissions scenario).

How useful is it to know that under a scenario like RCP6.0 the average global temperature increase in 2100 has been simulated as variously 1ºC, 2ºC, 3ºC, 4ºC? (note, I haven’t checked the CMIP5 simulations to get each value). And the tropics will vary less, land more? As we dig into more details we will attempt to look at how reliable regional and seasonal temperature anomalies might be compared with the overall number. Likewise rainfall and other important climate values.

I do find it useful to keep the idea of a set of possible numbers with no probability assigned. Then at some stage we can say something like, “if this RCP scenario turns out to be correct and the global average surface temperature actually increases by 3ºC by 2100, we know the following are reasonable assumptions … but we currently can’t make any predictions about these other values..”

References

Long-term Climate Change: Projections, Commitments and Irreversibility, M Collins et al (2013) – In: Climate Change 2013: The Physical Science Basis. Contribution of Working Group I to the Fifth Assessment Report of the Intergovernmental Panel on Climate Change

Scenarios of long-term socio-economic and environmental development under climate stabilization, Keywan Riahi et al, Technological Forecasting & Social Change (2007) – free paper

Error Reduction and Convergence in Climate Prediction, Charles S Jackson et al, Journal of Climate (2008) – free paper

Notes

Note 1: As explored a little in the last article, RCP6.0 does include some changes to climate policy but it seems they are not major. I believe a very useful scenario for exploring impact assessments would be the population and development path of RCP6.0 (let’s call it RCP6.0A) without any climate policies.

For reasons of”scenario parsimony” this interesting pathway avoids attention.