The Atmospheres of Two Exoplanets

By: Peter Cook

One of the big discoveries of recent decades has been the finding of thousands of exoplanets, and it now seems that most stars have planets.  Remarkably, detailed measurements have now been made of some of these despite the huge distances involved.  The first two exoplanets to have their atmospheres studied in detail, and their spectra directly observed, are HD209458b (also called Osiris) at a distance of 159 light years and HD189733b (having no other name at present) at a distance of 64.5 light years.  Both are large planets, similar in size to Jupiter and mostly composed of hydrogen and helium.  They orbit very close to their stars so are referred to as being “Hot Jupiters” (many of the exoplanets found so far are Hot Jupiters, though this is probably a selection effect as such planets are the easiest to find).

Osiris was first found using the motion of its star (as both go around a mutual centre of gravity) which revealed its mass, and then again when it transited its star (revealing its size).  Osiris is only 0.047 astronomical units (AU) from its star (one AU is the average distance between the Earth and Sun) and orbits in only 3.5 days.  While roughly similar in mass and radius to Jupiter it has a very low density (0.37 times that of water), and since its star is slightly more massive and brighter than the Sun Osiris has probably been inflated by violent heating of its outer atmosphere.  By studying changes in the total infrared emission from the planet and star, as Osiris hides part of the star during transit and is then eclipsed by the star, the temperature of the atmosphere was revealed to be around 1400 K by the Spitzer Space Telescope.  The spectra obtained seems to show a dry atmosphere containing sodium atoms and silicate particles, while high-precision observations by the European Southern Observatory’s (ESO) Very Large Telescope and the CRyogenic high-resolution InfraRed Echelle Spectrograph (CRIRES) measured gasses streaming from the day to night side at a speed of around 7000 km/hour (Osiris must be tidally locked and so keep the same side facing its star).  Further measurements imply that Osiris is very dark in colour, reflecting less than 10% of visible light, and has patchy high-altitude hot clouds composed of metal oxides.  Osiris is surrounded by an envelope of hydrogen, carbon and oxygen, more than twice the size of the planet, with a significant tail pointing away from the star.  This was discovered by the Hubble Space Telescope Imaging Spectrometer by measuring star light passing thorugh it and shows mass lost from the planet and blown away by its star.  Osiris has probably lost a few percent of its mass over its lifetime.

HD189733b was also found using both the transit and radial velocity methods, only 0.031 AU from its star and orbits in 2.2 days (though is not quite as hot as Osiris since its star is less massive and fainter than the Sun).  Observations by the Spitzer Space Telescope over 33 hours were used to create a temperature map (see figure).  These are in the range 973-1212 K with the peak offset 30 degrees east of the sub-stellar point (again the planet must be tidally locked) showing that absorbed energy is distributed fairly evenly through the atmosphere and hence wind speeds would be of order 8700 km/hour.  Again, the atmosphere seems to be fairly dry; sodium has been detected (3 times the concentration as at Osiris) and carbon dioxide.  Visible light polarimetry shows that HD189733b is blue in colour (unlike the very dark Osiris), probably due to Rayleigh scattering of blue light in the atmosphere and the absorption of red light by molecules.  An X-ray spectrum of the planet’s transit (H1 Lyman-Alpha) shows an extended exosphere of hydrogen and hence a slow evaporation.  Some further measurements indicated the presence of water vapour and methane in the atmosphere (though these should quickly react together in the high temperatures to form carbon monoxide), and silicate particles (in which case, HD189733b could have rain consisting of molten glass!).

This information is from Wikipedia, and the Spitzer Space Telescope map was also found online.

Posted in Climate, Space | Leave a comment

What is “net zero” for methane?

By: Bill Collins

Recent research is suggesting that the way methane is accounted for in climate targets overemphasises its contribution to climate change at the end of the century. This might mean that countries or sectors (e.g. agriculture) with large methane emissions might have to impose overly stringent CO2 cuts to compensate (but is that a bad thing?), while countries that are able to reduce methane emissions might get away with insufficient CO2 cuts.

The new way of looking at methane reveals that to stabilise temperatures it is the rate of methane emission that is important rather than the total methane emitted (Collins et al. 2020, Allen et al. 2018, Cain et al. 2019). Therefore, a useful way of comparing methane and CO2 is to compare a change in the rate of methane emissions (tonnes per year) with a change in cumulative CO2 emissions (tonnes). This would have considerable implications as to what “net zero” means for each country.

The big step change in understanding from the IPCC 5th Assessment Report (IPCC 2013) was the concept of an almost linear relationship between the amount of carbon dioxide emitted and resulting temperature. From this it follows that to stabilise temperatures the net amount of carbon dioxide emitted will have to decline to zero. This was recognised in the Paris Agreement in which Article 4 refers to achieving “a balance between anthropogenic emissions by sources and removals by sinks of greenhouse gases in the second half of this century”. Some countries (including the UK) have therefore adopted a target of “net zero” greenhouses gas emissions by 2050.

Net zero is clear for CO2, but what about for methane? Methane has a natural removal sink through chemical oxidation in the atmosphere. Should this process be considered part of the “removals by sinks” specified in the Paris Agreement or should only direct capture of methane be counted? Methane behaves very differently to CO2 such that if we maintain global methane emissions at the current level then the concentrations of methane will stabilise and not contribute further to global warming. Reduction in methane emissions will even make a cooling contribution to climate. This contrasts with CO2 where constant emissions lead to a constant warming rate and reduced emissions still lead to a reduced warming rate until they decline to zero.

So how do we include methane in the net zero target? A common way to compare greenhouse gases is through a metric called the 100-year Global Warming Potential (GWP100) that compares the energy imbalance (radiative forcing) caused by emissions of a tonne of each gas averaged over the following 100 years. The GWP100 suggests 1 tonne of methane emission is equivalent to 34 tonnes of CO2 (Collins et al. 2020). Because methane has a short lifetime in the atmosphere (around 12 years) its climate impacts are much smaller by the end of the century (say 2070 or 2095) – only equivalent to 14 or 8 tonnes of CO2 respectively (Collins et al. 2020). This means that GWP100 overestimates the temperature effects of methane on the sort of time frames when we might reach peak warming.

Achieving “net zero in 2050”: schematic comparison of assumptions assuming a constant decline in methane emissions to 2100. CO2 equivalent emissions for methane (dotted red) are based on a rate/cumulative metric of 3000 years. Total CO2 equivalent (dotted black) are the sum of the solid green and dotted red.

Left: Scaling the methane emissions by 30 (GWP100), non-zero methane in 2050 means negative CO2 emissions are needed to compensate. The CO2 equivalent emissions become net zero earlier than 2050 and stay negative leading to an early peak in temperature and a decline afterwards.

Right: using the rate vs cumulative metric (GWP* or combined-GTP – see text), decreasing methane emissions in 2050 means CO2 emissions can be slightly positive. The CO2 equivalent emissions become net zero in 2050 and remain zero leading to a stabilisation of temperature at 2050.

The solution is to put all this together. Temperatures depend on the rate of methane emissions, but the accumulation of CO2 emissions. Hence a useful metric compares changes in these two quantities rather than GWP100 which compares changes in cumulative emissions of both. These rate vs. cumulative metrics are called GWP* in Allen et al. (2018) and combined-GTP in Collins et al. (2020), and a mix of GWP100 and GWP* has been called “CO2-warming equivalent” in Cain et al. (2019). These metrics have values of around 3000 years – depending exactly on how they are calculated. This importance of these new metrics is that a zero change (i.e. constant) emission rate of methane is equivalent to a net zero change in cumulative CO2, and a reduction in methane emission rate of 1 tonne per year is equivalent to a one-off negative emission of around 3000 tonnes of CO2.

Using the standard GWP100 metric for methane means CO2 emissions have to be more strongly negative to compensate, whereas for the GWP* or combined-GTP metric the CO2 emissions can be slightly positive (or more likely less negative to compensate for other gases). The second method leads to the desired stabilisation of climate, but the first method has lower temperatures. Is it better to overvalue methane if it makes us take more severe action on CO2, or to use a metric more closely tied to the temperature stabilisation that may lead to complacency?

References:

Allen, M. R., J. S.Fuglestvedt, K. P. Shine, A. Reisinger, R. T. Pierrehumbert, and P. M. Forster, 2016: New use of global warming potentials to compare cumulative and short-lived climate pollutants. Nat. Climate Change, 6, 773–776. https://doi.org/10.1038/nclimate2998

Cain, M., J. Lynch, M. Allen, J. S. Fuglestvedt, D. J. Frame, and A. H. Macey, 2019: Improved calculation of warming-equivalent emissions for short-lived climate pollutants, npj Climate Atmos. Sci., 2, https://doi.org/10.1038/s41612-019-0086-4

Collins, W. J., D. J. Frame, J. S. Fuglestvedt, and K. P. Shine, 2020: Stable climate metrics for emissions of short and long-lived species – combining steps and pulses. Environ. Res. Lett., 15, 2, 024018 https://doi.org/10.1088/1748-9326/ab6039

Posted in Climate, Climate change, Climate modelling | Leave a comment

Housebuilding ban on floodplains isn’t enough – flood-prone communities should take back control

By: Hannah Cloke

February 2020 has brought more than its fair share of bad weather to the north of England, the Midlands and Wales. Shrewsbury, Bewdley and Telford swam in the Severn, while the Ouse invaded York. For some, the adage that it’s grim up north rang true.

The recent flooding is a reminder that all parts of the UK are vulnerable to natural hazards, and the costs aren’t just economic. Flood water can enter a building in minutes, but the impact on communities can last years. Flooded homes and businesses take months to clean up and dry out, and the long-term impact on the health and relationships of those affected is often overlooked.

Climate change has made some types of floods more likely, but past government policy has ensured that the ensuing crises are worse than they might otherwise have been. While there’s an urgent need for new homes, 1.8 million people currently live in areas at significant risk of flooding, and homes are still being built on floodplains.

The chief executive of the Environment Agency, Sir James Bevan, argued that if we must build in the floodplain, homes should be built with garages on the ground floor and living space upstairs, to ensure floods cause minimal damage. This might sound radical, but some of the oldest buildings along Britain’s rivers and coastline, such as millhouses and warehouses, have stone floors and steps up to first floor entrances. They were designed hundreds of years ago to ensure their occupants could ride out intermittent flooding.

Figure 1: Old millhouses – like this building in the US – are often raised above ground level, in anticipation of flooding. StudioKismet/Shutterstock

Climate change may force countries, such as the UK, to adopt radical practices from parts of the world that flood more frequently, such as houses that are designed to float when floods come, rising on stilts as the water rises. It’s an idea that is familiar to those living in stilt houses beside the Amazon river, and it’s also found favour in the Netherlands.

For those who already live on the floodplain, there are less dramatic but important adaptations that can be made. Internally reinforced, mechanically sealable flood doors can be installed to keep water out. Carpets and wooden floorboards that soak up dirty flood waters can be swapped for waterproof concrete and stone-slab floors. Electrical sockets can be raised and non-return valves can be fitted to toilets to stop sewage filling homes when it floods.

Communities that can weather floods

Making Britain more resilient to floods is not just a task for individuals. The government now has an opportunity to prove its long-term commitment to the north of England in particular, by creating detailed plans to increase resilience to floods. Backing the HS2 railway and moving the House of Lords out of London might grab headlines, but making sure the region is resilient in the face of future flooding requires less glamorous investment.

It’s equally important that money isn’t just thrown at the areas that were flooded last – or which might have voted Conservative in the last election. Every flooded community, no matter where they are in the country, deserves support, and managing flood risk is about more than just installing large flood defences or water-proofing homes. A long-term approach requires policies that link the necessary changes in land use, agriculture, housing and development.

Part of this could give communities the power to take control of their own destiny by creating regional flood forums, giving residents’ groups a say in how communities prepare for and respond to floods. Too many people are left with the impression that the government, or other distant authorities, can be trusted with sole responsibility for preventing flooding, and are then left feeling powerless and abandoned when floods happen.

 
Figure 2: Communities in Wales reel from the impact of Storm Dennis. Could involving residents in flood management make them more resilient to future floods? EPA-EFE/DIMITRIS LEGAKIS

It’s essential that everyone who lives in or moves to an area at risk of flooding knows that the risk of flooding is real. It’s difficult to imagine that your home could be flooded if you have never seen water lapping at your front door or rising up your kitchen cabinets. This is especially true for those living in properties that don’t immediately border running water.

But the historical record tells us that even before modern climate change, big floods have always happened. And floods are like buses, sometimes nothing happens for ages and then several come, one after the other. Flood risk maps exist, but the message is not getting through to those who need them.

Perhaps flood risk should come to be seen like earthquake risk zones in other countries. It would mean public signs clearly positioned to show that you are in a flood zone and that there is danger to life, health and property. It would ensure that strong building regulations are adhered to and school lessons are specifically devoted to knowing how to keep people safe.

We certainly need new ideas for connecting people with the realities of living with climate change. Ensuring that everyone can find out about their flood risk, shape decision-making on land management, and adapt their homes accordingly could empower people to prepare for the floods that will inevitably come.

This article was originally published in the Conversation.

Posted in Climate, Climate change, Flooding | Leave a comment

Meiyu—Baiu—Changma Rains

By: Amulya Chevuturi

The arrival of the summer monsoon rains over southern China is called Meiyu, literally translated as “plum rains”. These are also called Baiu in Japan and Changma in Korea. As the monsoon progresses, these rain belts first occur over Taiwan, southern China and Okinawa region from early-May to mid-June, then continue towards the Yangtze River valley and the main islands of Japan from mid-June to mid-July, and finally reach the Korean Peninsula and north-eastern China during mid-July to mid-August. This northward march of the rain belts occurs in tandem with the seasonal progression of the East Asian summer monsoon. These rain belts, which last on an average for 8 days, are named after the respective regional names of the season. 

Figure 1: Example of the Meiyu—Baiu—Changma rain belt cloud cover (demarcated with a red dotted line) over parts of southern China, Taiwan and Okinawa Islands on 14th June 2017. Source: WorldView NASA.

These rain belts are formed due to monsoon frontal systems, which are strongly influenced by the moisture flow from the monsoon winds in the lower atmosphere. These east-to-west extended frontal systems may last for 3 to 22 days and lead to local-scale variability within the large-scale monsoon season over East Asia. As these frontal systems move north, their characteristics transition from those of tropical disturbances to mid-latitude ones. The rainfall from these systems contributes up to 45% of the total monsoon rainfall over some regions of East Asia. A snapshot of cloud cover during one such rain belt over East Asia, on 14th June 2017, is shown in Figure 1. This weather system led to devastating amounts of rainfall, which lasted almost up to 10 days. It left many displaced people, had damaged infrastructure and ruined crops, due to subsequent floods and landslides.

Figure 2: (a) Mean position of the monsoon front band during May (red), June (purple) and July (blue) and (b) mean June rainfall associated with the mean June frontal band (purple) in ERA5 dataset from 1979-2014.

Researchers in our department, inspired by previous research, have defined an objective method to identify these frontal systems by detecting regions with strong temperature and moisture gradients by using the north-to-south gradient of the equivalent potential temperature. Through this method we can identify the narrow band of the monsoon frontal systems and detect any rainfall within approximately 600 kilometres of the band, as the Meiyu—Baiu—Changma rain belt. Using this method, we have identified the monthly average position of the monsoon front and accompanying rainfall for May, June and July from 1979-2014, using the newest reanalysis dataset from European Centre for Medium-Range Weather Forecasts (ECMWF), named ERA5, shown in Figure 2. We clearly observe the northward progression of the frontal bands from May—June—July, and more rainfall south of the frontal band position. We also found larger contribution to total monsoon rainfall during May than in July, from the monsoon fronts, due to a greater number of frontal systems early in the season.

When we analysed the representation of these monsoon fronts in the UK Met Office model, we found that although the model can generally represent the average position of these frontal systems, it overestimates the associated rainfall. This rainfall error in the model stems from the overall overestimation of the summer monsoon rainfall over East Asia. Further, the monsoon fronts are influenced by the El Niño Southern Oscillation (ENSO) conditions (based on the equatorial Pacific sea surface temperature) in the proceeding winter. Warmer temperatures (El Niño phase), increase the occurrence of these frontal systems, whereas an opposite response is seen due to colder sea surface temperatures (La Niña phase). Although the model correctly depicts this remote teleconnection between ENSO phases and the monsoon fronts, the strength of the remote relationship is much stronger in the model.

The monsoon rainfall from such systems affect the lives of over a billion people. Our continued work for better understanding of these monsoon frontal systems and their representation in the models, will ultimately improve the prediction of rainfall during the monsoon season. This will allow for improved disaster management, agricultural planning and water resource allocation, and thus benefitting the regional economy.

References

Charlier, P., 2017: Thousands Evacuated, Army on Standby as Heavy Rain Pounds Taiwan. Taiwan English News, accessed 04 March 2020, https://taiwanenglishnews.com/thousands-evacuated-army-on-standby-as-heavy-rain-pounds-taiwan/.

Ding, Y.H, and J.C. Chan, 2005: The East Asian summer monsoon: an overview, Meteor. Atmos. Phys., 89, 117-142. DOI: https://doi.org/10.1007/s00703-005-0125-z.

Jun-Mei, L., J. Jian-Hua, and T. Shi-Yan, 2013: Re-discussion on East Asian Meiyu rainy season, Atmos. Oceanic. Sci. Lett., 6(5), 279-283. DOI: https://doi.org/10.3878/j.issn.1674-2834.13.0024.

Li, Y., Y. Deng, S. Yang, and H. Zhang, 2018: Multi-scale temporospatial variability of the East Asian Meiyu-Baiu fronts: characterization with a suite of new objective indices, Climate Dynamics, 51, 1659–1670. DOI: https://doi.org/10.1007/s00382-017-3975-4.

Sampe, T., and S.P. Xie, 2010: Large-scale dynamics of the Meiyu-Baiu rainband: Environmental forcing by the westerly jet, J. Climate, 23, 113-134, DOI: https://doi.org/10.1175/2009JCLI3128.1.

Posted in ENSO, Monsoons, Rainfall | Leave a comment

Water Vapour Absorption and its Role in the Earth’s Energy Budget

By: Jon Elsey

I’m Jon, a postdoc working with Prof. Keith Shine in the Atmospheric Radiation, Composition and Climate (ACRC) group. My work is very much in the “R” of ACRC, and specifically on the role of water vapour on atmospheric radiation.

When we think of “radiation”, you might think of the Hulk, or conjure up images of nuclear meltdown, or some other catastrophe. What I actually mean here by radiation is light emitted by the Sun (visible light, ultraviolet and high frequency infrared) which we call shortwave radiation, or by the Earth and its atmosphere (mostly lower-frequency infrared) which we call longwave radiation.

 Figure 1: The longwave and shortwave atmospheric absorption processes are circled. All units in W m-2. (from Stephens et al. (2012), Nature Geoscience).

This radiation is part of the Earth’s energy budget shown in Figure 1; shortwave radiation comes in from the Sun, which is then absorbed by the atmosphere and surface, with some being reflected back to space. The absorbed radiation heats up the surface and atmosphere, which then emit radiation in the longwave. The radiation emitted up from the surface in the longwave is then absorbed again by the atmosphere, trapping this energy within the Earth system. This is the well-known greenhouse effect, the main contributors to which are water vapour (H2O) and CO2.

The energy emitted by the Sun and Earth is in the form of a spectrum, made up of the visible, infrared and ultraviolet light I mentioned earlier. This radiation is absorbed by atmospheric gases at specific frequencies. For example, CO2 and water vapour absorb radiation very differently – CO2 absorbs very strongly in only very specific parts of the spectrum, whereas H2O absorbs radiation very strongly across most of the shortwave and longwave, with opaque regions (the so-called absorption bands) interspersed with much more transparent regions (the atmospheric windows).

Figure 2: Plot of the optical depth from CO2 (blue), H2O (orange) and H2O continuum (green) for the entire depth of a real atmosphere (measured in Cornwall, UK). The red dotted line shows the demarcation between the longwave and shortwave. Note that the y-axis is on a logarithmic scale – the 105 line is 1000x stronger than the 102 line, etc.

Figure 2 shows the absorption in a real atmosphere measured during my PhD, in units of optical depth, which is just a measure of how opaque the atmosphere is. So, looking at Figure 2, if water vapour absorbs more radiation than CO2, then why are we worried so much about CO2? The issue here is to do with what we call climate forcings and climate feedbacks. A forcing can be thought of as due to something outside of the climate system causing a change in its properties, while a feedback is the result of the climate system responding to a forcing.

In this context, CO2 contributes to the forcing (we emit CO2, which causes the atmosphere to heat up) and water vapour is a feedback (higher temperature leads to more water vapour in the atmosphere from evaporation of the oceans). This means that as CO2 heats up the Earth, it also causes an increase in the water vapour absorption, trapping much more radiation than the CO2 would alone, amplifying the overall effect on the climate by a factor of 2-3.

My own work focuses on what we call atmospheric spectroscopy, the study of how water vapour (and other molecules!) absorb radiation. Specifically, I study the continuum absorption, the green line in Figure 2. Like the spectral lines of water vapour, this continuum absorption occurs both in the shortwave and longwave, but is distinct from the lines in that it is smoothly-varying (in frequency) with no distinguishable features in the atmospheric windows. It is in these windows that the continuum is most significant.

From a climate science point of view, it is the effect on the energy budget that we are most interested in. Previous studies indicate that the shortwave continuum which I study could be significant in this context. However, measuring the continuum in the lab is really rather difficult, since there are no obvious features to look at, leading to difficulty in extracting the real absorption from any kind of instrumental noise.

There are a few groups working on studying this shortwave continuum, but with poor agreement between their measurements. My work is a collaboration between Reading and the Rutherford Appleton Laboratory, using state-of-the-art spectroscopic techniques. These measurements are still ongoing, but will hopefully go some way to solving this puzzle and contributing to our understanding of how water vapour absorption affects the climate.

References:

Myhre, G., D. Shindell, F.-M. Bréon, W. Collins, J. Fuglestvedt, J. Huang, D. Koch, J.-F. Lamarque, D. Lee, B. Mendoza, T. Nakajima, A. Robock, G. Stephens, T. Takemura and H. Zhang, 2013: Anthropogenic and Natural Radiative Forc-ing. 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 [Stocker, T.F., D. Qin, G.-K. Plattner, M. Tignor, S.K. Allen, J. Boschung, A. Nauels, Y. Xia, V. Bex and P.M. Midgley (eds.)]. Cambridge University Press, Cambridge, United Kingdom and New York, NY, USA

Radel, G., Shine, K. P. and Ptashnik, I. V. (2015) Global radiative and climate effect of the water vapour continuum at visible and near-infrared wavelengths. Quarterly Journal of the Royal Meteorological Society, 141 (688). pp. 727-738. ISSN 1477-870X. https://doi.org/10.1002/qj.2385

Stephens, G., Li, J., Wild, M. et al. An update on Earth’s energy balance in light of the latest global observations. Nature Geoscience, 5, 691–696 (2012). https://doi.org/10.1038/ngeo1580

Shine, K.P., Campargue, A., Mondelain, D., McPheat, R.A., Ptashnik, I.V. and Weidmann, D., The water vapour continuum in near-infrared windows–current understanding and prospects for its inclusion in spectroscopic databases. Journal of Molecular Spectroscopy, 327, pp.193-208. (2016). https://doi.org/10.1016/j.jms.2016.04.011

Posted in Climate, Climate change, Radiation | Leave a comment

Building a predictive framework for studying causality in complex systems

By: Nachiketa Chakraborty

I’m Nachiketa Chakraborty, a postdoctoral researcher working on the ERC project CUNDA (Causality under Non-linear Data Assimilation) led by Peter Jan van Leeuwen. My central goal is to come up with a Bayesian framework for studying causal relations within complex systems that are higher dimensional and have non-linear, coupled processes. My background is in astrophysics, with a special interest in using mathematical techniques (like time-series analyses) to study the origin of variability in emissions from astrophysical sources involving black holes at their center [1,2,3]. It is fascinating to note that these methods are in fact highly relevant in building this causal framework for the processes in the Earth sciences, such as for the inter-ocean exchange problem in Indian, the South Atlantic and the Southern Ocean – a critical ingredient of the climate system [4].

Study of cause and effect is central to every science. It is what allows us to identify processes that are instrumental in giving rise to the phenomena we observe. For instance, it is the force of gravity that causes objects in air to fall back towards the Earth. In geosciences, this relationship between causes and effects plays out in more complex ways. For one, this is because there are multiple competing causes for every effect. Furthermore, this relationship is highly nonlinear; in other words, amplifying or increasing the strength of the causal process does not simply increase or decrease the observed effect proportionally. Finally, the different processes are interdependent. Therefore, it is not possible to decouple different effects and study them independently building up our understanding from simpler models. 

All these aspects come into play in the longstanding inter-ocean exchange problem in oceanography. We wish to know if exchange between the Indian, the South Atlantic and Southern Ocean is caused predominantly by local effects (namely, the highly turbulent eddying south of the African continent), or is in fact due to global scale dynamics (such as large scale modes of the Southern Ocean). What is known currently is that processes ranging from the very local to very global are plausible drivers. Disentangling these requires development of sophisticated methods building from existing methods [5].

As a person from a physics background using mathematical techniques to analyse data, causality represents a perfect area to blend these two. In both mathematics and physics, causality is a key concept. Within physics, central to the concept of causality is the existence of a plausible mechanism that causes the effects we observe and time ordering. This is quantified through equations which can be used to predict observations.

Mathematically, causality as a concept has many components. First, we need a set of conditions defining the presence of a causal connection or cause and effect. The most popular way of using data to do this is by establishing statistical dependencies between the variables observed. This brings us to the next key mathematical aspect of causality which is the mathematical measure quantifying causation. Shannon’s information theory measures are the most popular and best studied of the measures quantifying causality. This is the strategy that is used to establish causality in systems or areas where we do not have known plausible mechanisms (such as econometrics) or it is too complex to have one (as in biological systems). Finally, in order to properly quantify causality within processes in systems, we need to provide the uncertainty on the estimates. In order to do that, we wish to adopt a Bayesian approach to causality. Combining the physical and the mathematical, data-driven approach, we hope to build towards solutions for complex systems.

References:

[1] Ait Benkhali, F., W. Hofmann, F. M.Rieger, and N. Chakraborty, 2019: Evaluating Quasi-Periodic Variations in the γ-ray Lightcurves of Fermi-LAT Blazars, Astron. Astrophys., DOI: https://doi.org/10.1051/0004-6361/201935117, available at https://arxiv.org/abs/1901.10246   

[2] Morris, P., N. Chakraborty, and G. Cotter, 2017: Deviations from normal distributions in artificial and real time series: A false positive prescription, Mon. Not. R. Astron. Soc., 2, 2117–2129, https://doi.org/10.1093/mnras/stz2259

[3] Chakraborty, N., 2020: Investigating Multiwavelength Lognormality with Simulations: Case of Mrk 421, Galaxies 2020, 8(1), 7, https://doi.org/10.3390/galaxies8010007

[4]  de Ruijter, W. P. M., A. Biastoch S. S. Drijfhout J. R. E. Lutjeharms R. P. Matano T. Pichevin P. J. van Leeuwen W. Weijer, 1999: Indian‐Atlantic interocean exchange: Dynamics, estimation and impact, J. Geophys. Res.104, C9, 20885– 20910, doi:10.1029/1998JC900099.

[5] Amblard, P.-O., and O. Michel, 2012: The Relation between Granger Causality and Directed Information Theory: A Review. Entropy15, 113–143, doi:10.3390/e15010113. http://dx.doi.org/10.3390/e15010113.

Posted in Climate, data assimilation, Oceans | Leave a comment

Rainfall decline over Eastern Africa linked to shorter wet seasons

By: Caroline Wainwright

2019 was a year of extremes for East Africa. The long rains (March-May) started late and low rainfall during the season exacerbated the drought conditions following the dry short rains (October-December) in 2018. Conversely, the short rains in 2019 were extremely wet, with reports of floods and landslides from across the region, and over 2.8 million affected by the heavy rainfall.

High dependence upon climate for subsistence means that East Africa is particularly vulnerable to climate extremes and climate variability. Recently, the HyCRISTAL project has been working to develop new understanding of East African climate variability and change and its impacts, and working with regional decision makers to support the use of climate information in long-term decision making in the region.

Since the mid-1980s rainfall during the long rains (the dominant of the two East African wet seasons) has declined, bringing major socio-economic consequences. However, future climate projections from climate models show long rains rainfall increasing in the future. This inconsistency has been termed the “East Africa Climate Change Paradox” and has limited the use of climate change projections in decision making in the region.  In our recent study, completed as part of the HyCRISTAL project, we examine the causes and implications of this apparent paradox.

Figure 1: a) Long Rains rainfall anomaly over Eastern Africa 1985–2018 from 3 satellite-based rainfall products (CHIRPS, TAMSAT and GPCP). The bars represent the anomaly for each year (CHIRPS); the lines are smoothed using a 3 year moving average. b) Mean seasonal cycle over Eastern Africa from CHIRPS over the 3 periods of study. The schematic on the right shows the position of the stronger Arabian Heat Low and stronger Somali.

Analysis of satellite-based rainfall datasets shows that the decline in rainfall during the long rains was associated with a later start to the season (later onset) and earlier end of the season (earlier cessation, Figure 1). We found that since the late 2000s there has been some recovery of the long rains, although with high year to year variability.

Concurrently with the rainfall decline over Eastern Africa is a decrease in the surface pressure over Arabia in May. This indicates a strengthening of the Arabian Heat Low. We previously discovered a similar strengthening of the Sahara Heat Low was linked with later wet seasons over West Africa (Dunning et al., 2018). The Arabian Heat Low is a dominant feature over the Arabian Peninsula in spring and summer, that interacts with local weather phenomena including the Somali Jet and the Indian Monsoon. The southerly winds, forming part of the Somali Jet, are also found to strengthen in May. Furthermore, over the period of the decline, May sea surface temperatures increased more over the very north Arabian Sea than further south in the Indian Ocean.

The increase in wind speed, strengthening of the Arabian Heat Low (Figure 1), and change in sea surface temperature gradient across the Arabian Sea act to draw the rain-band northwards faster and further, away from Eastern Africa, consistent with the decline in rainfall and earlier end to the long rains.

Considering the later onset, warming sea surface temperatures south of Madagascar are associated with reduced March rainfall and later onsets, as warmer temperatures to the south delay the northward progression of the rain-band.

The next challenge is to identify the role of natural variability and the role of anthropogenic carbon emissions on this mechanism, to ascertain whether the recent decline is part of natural variability or part of a climate change signal. Other studies have linked natural variability in Indian Ocean sea surface temperatures with multi-decadal climate variability over East Africa (Tierney et al., 2013). Understanding the response of the mechanism identified here to rising greenhouse gas concentrations will contribute more information on the current uncertain future rainfall changes over East Africa.

Read more and download the paper “‘Eastern African Paradox’ rainfall decline due to shorter not less intense Long Rains” here.

The key findings from the HyCristal project (so far!) on climate change are summarised here.

References:

Dunning, C.M., Black, E. and Allan, R.P. (2018) Later Wet Seasons with More Intense Rainfall over Africa under Future Climate Change, Journal of Climate, 31, 9719–9738, doi: https://doi.org/10.1175/JCLI-D-18-0102.1

Finney, D., Marsham, J., Rowell, D., Way, C., Evans, B., Cornforth, R., … & Anyah, R. (2019). Scientific Understanding of East African climate change from the HyCRISTAL project., https://doi.org/10.5518/100/19

Tierney, J., Smerdon, J., Anchukaitis, K. et al. (2013). Multidecadal variability in East African hydroclimate controlled by the Indian Ocean. Nature 493, 389–392 https://doi.org/10.1038/nature11785

Wainwright, C. M., Marsham, J. H., Keane, R. J., Rowell, D. P., Finney, D. L., Black, E., & Allan, R. P. (2019). ‘Eastern African Paradox’ rainfall decline due to shorter not less intense Long Rains. npj Climate and Atmospheric Science2, 34 https://doi.org/10.1038/s41612-019-0091-7

Online material:

Laura Harrison, Juliet Way-Henthorne, and Chris Funk (2019) Climate Hazards Center Early Estimates and the East Africa March-to-May 2019 Drought

OCHA, EASTERN AFRICA REGION Regional Floods Snapshot (2019) 

 

 

Posted in Africa, Climate, Climate change, Rainfall | Leave a comment

From Indonesia to the British Isles: using El Niño and weather patterns in the tropics to help predict North Atlantic and European weather

By: Robert Lee

The winter weather in the UK and Europe can be split into different patterns based on the large-scale flow in the atmosphere. A commonly used method is to use a type of machine learning algorithm – a clustering algorithm – to partition the weather into distinct categories, known as ‘regimes’. Weather regimes can persist for a number of days or even weeks, before transitioning to another regime. Using large-scale weather regimes such as these are useful when considering the ‘subseasonal’ forecasting period, approximately 10–20 days ahead since they give an approximate indication about the average weather covering many countries for many days, without focusing on precise differences on a local and hour-by-hour scale, which cannot be forecasted well due to the inherent chaos.

For the combined North Atlantic and European region, it turns out that the optimal number of large-scale weather regimes is four [1]. These regimes are often called the (1) NAO− (a.k.a. Greenland Blocking), (2) NAO+ (a.k.a. Zonal), (3) Atlantic Ridge, and (4) Scandinavian Blocking. As a side note, the NAO− and NAO+ weather regimes are similar, but not identical, to the North Atlantic Oscillation index. Figure 1 illustrates the differing winter weather conditions under each regime. For example, NAO− brings colder and calmer conditions to the UK and North Sea region, while NAO+ brings milder, windier and wetter conditions.

Figure 1: Anomalies of temperature, wind, and rainfall (precipitation) during the extended winter season during each of the four weather regimes.

Subseasonal forecasts of weather regimes can help authorities and businesses to plan ahead in sectors such as agriculture, energy, health, aviation and transportation, water, and retail. In the energy sector, for example, if NAO− weather is forecasted, then the energy industry can plan for a ‘shortfall’ in electricity [2]. A shortfall is caused when the demand exceeds the supply, due to increased electricity used for heating in the colder weather. A (growing) proportion of the electricity supply comes from wind generation: during NAO− days the winds are much calmer. Similarly in the health sector [3], a heads-up about an upcoming period of NAO− can aid preparations for an increase in average and peak (extreme 5-year return interval) hospital admissions and mortality in the UK, associated with the colder conditions.

Remote atmospheric links from other regions, known as ‘teleconnections’, provide a way of getting an advanced warning for upcoming weather, thereby improving the potential for predictions on subseasonal timescales. For over 30 years there has been a known link between weather in the tropics and weather in the North Atlantic and European region that exists on these subseasonal timescales [4]. These links are primarily driven by a weather phenomenon in the tropics known as the Madden-Julian Oscillation (MJO), which is an eastward moving ‘pulse’ of suppressed and enhanced cloud and rainfall near the equator which typically recurs every 30 to 60 days during the boreal winter (although the MJO is not active about one third of the time). An important study in 2008 [5] showed how these teleconnections from the MJO in the tropics influence the weather regimes in the North Atlantic and European region. The study also showed how packets of energy, known as ‘Rossby waves’, can travel eastwards from the Pacific to the North Atlantic in the jet streams of the atmosphere, revealing some of the physics involved in these teleconnections.  

At the inception of our study [6], the El Niño–Southern Oscillation (ENSO) was hypothesised to also play a role in altering these teleconnections. ENSO characterises the periodic variation of sea surface temperatures and accompanying atmospheric circulation over the tropical Pacific Ocean. The warm phase is known as ‘El Niño’, and the cool phase as ‘La Niña’, reaching peak intensity during boreal winter. The Pacific sea surface temperature is commonly split into three categories: (1) El Niño, (2) neutral, and (3) La Niña, as shown in Figure 2.

Figure 2: Sea surface temperature anomalies in the tropical Pacific during the extended North Hemisphere winter season, under El Niño, neutral, and La Niña conditions.

ENSO conditions change on a longer (slower) timescale than the MJO, and so from a subseasonal perspective, ENSO conditions remain approximately constant. We find that ENSO modifies the MJO [6], making these regions of wet and dry weather narrower or broader in longitude (Figure 3).

Figure 3: The classic animation of the MJO (top row), here shown as an Outgoing Longwave Radiation (OLR) anomaly (to represent the clouds); and during El Niño, neutral, and La Niña conditions (lower three rows). The phase of the MJO is labelled at the top.

This ENSO modification of the MJO then has a consequence on the type of teleconnection that is triggered and the path that it takes. Our study [6] shows that during El Niño winters the teleconnection from MJO phases 1–3 makes the NAO+ regime occur twice as often as the full climatology, with the signal travelling along the jet streams (Figure 4). Whereas during La Niña this MJO phases 1–3 teleconnection is absent and there is no increase in NAO+ regime occurrence. During La Niña years we also find the teleconnection from the MJO phases 6–8 makes the NAO– regime occur up to 2.5 times as often as the full climatology – this signal travels via the stratosphere, warming it and slowing the stratospheric polar vortex (Figure 4), with the total pathway taking around 20 days. There is a strong subseasonal link between the stratospheric polar vortex and the weather regimes throughout all winters [7], however, it is during La Niña years when there is the strongest subseasonal link between the MJO and the stratosphere [6].

Figure 5: Simplified schematic of teleconnection from MJO phases 3 and 7, to the North Atlantic and European region during El Niño and La Niña conditions, respectively. Schematic adapted from: [6].

This statistical method, using the ENSO and MJO conditions, indicates which weather regimes we might expect 10-20 days in advance. By combining this with dynamical models, which can forecast the MJO up to 10 days ahead, we may improve the weather outlook for up to 30 days ahead. These results have important implications, including for the achievable skill in predicting these weather regimes on subseasonal timescales. In addition, this dependence of the teleconnections on the ENSO state should be well represented in weather and climate models. Finally, long term changes (both past and future) in MJO amplitude and in ENSO strength, may alter the proportions of time in each the weather regimes experienced overall.

This work [6 & 7] is part of the InterDec project, connecting scientists from 6 countries, researching inter-regional linkages on subseasonal-to-decadal timescales, and has been funded by NERC via the 2015 joint call from the Belmont Forum and JPI Climate.

References:
  1. Michelangeli, P.-A., R. Vautard, and B. Legras, 1995: Weather Regimes: Recurrence and Quasi Stationarity. Atmos. Sci., 52, 1237–1256, https://doi.org/10.1175/1520-0469(1995)052<1237:WRRAQS>2.0.CO;2
  2. Wiel, K. Van Der, H. C. Bloom, R. W. Lee, L. P. Stoop, R. Blackport, J. A. Screen, and F. M. Selten, 2019: The influence of weather regimes on European renewable energy production and demand. Res. Lett., 14, 094010, https://doi.org/10.1088/1748-9326/ab38d3
  3. Charlton-Perez, A. J., R. W. Aldridge, C. M. Grams, and R. Lee, 2019: Winter pressures on the UK health system dominated by the Greenland Blocking weather regime. Weather Clim. Extrem., 25, 100218, https://doi.org/10.1016/j.wace.2019.100218
  4. Ferranti, L., T. N. Palmer, F. Molteni, and E. Klinker, 1990: Tropical-Extratropical Interaction Associated with the 30–60 Day Oscillation and Its Impact on Medium and Extended Range Prediction. Atmos. Sci., 47, 2177–2199, https://doi.org/10.1175/1520-0469(1990)047<2177:TEIAWT>2.0.CO;2
  5. Cassou, C., 2008: Intraseasonal interaction between the Madden–Julian Oscillation and the North Atlantic Oscillation. Nature, 455, 523–527, https://doi.org/10.1038/nature07286
  6. Lee, R. W., S. J. Woolnough, A. J. Charlton‐Perez, and F. Vitart, 2019: ENSO modulation of MJO teleconnections to the North Atlantic and Europe. Res. Lett., 46, 13535– 13545, https://doi.org/10.1029/2019GL084683
  7. Charlton-Perez, A. J., L. Ferranti, and R. W. Lee, 2018: The influence of the Stratospheric state on North Atlantic Weather Regimes. J. R. Meteorol. Soc., 144, 1140– 1151, https://doi.org/10.1002/qj.3280

 

Posted in Climate, ENSO, Madden-Julian Oscillation (MJO), Predictability, subseasonal forecasting | Leave a comment

Desert Dust in the Atmosphere: Giant Particles, Giant Consequences?

By: Claire Ryder

As I write, storm Gloria decays over the Mediterranean Sea, while large amounts of desert dust whipped up by strong winds over the Sahara desert have been whirled in to action by Gloria and remain in the atmosphere. This dust is now being blown northwards across Europe and may travel as far as the UK. Dust reaching the UK is usually so dispersed that it’s often only noticeable by producing pretty red sunsets and dirty red rain on our cars, sometimes referred to as ‘blood rain.’ Did you notice any the weekend before last?

Figure 1: Dust forecast for 12:00 on 23rd January 2020 from the Barcelona Dust Forecast Centre (https://dust.aemet.es/forecast) showing Saharan dust being transported across Europe by storm Gloria

Most Saharan dust in the atmosphere is transported westwards across the Atlantic Ocean. Unlike dust reaching the UK, the Atlantic dust plume is present throughout the year, at varying latitudes, altitudes and intensities. It is important for climate models to be able to accurately predict dust in our atmosphere. This is because dust particles interact with sunlight (which causes an overall cooling effect) as well as thermal radiation (heat) from the Earth (which causes an overall warming effect). The balance of these two effects is the decider on whether dust in our atmosphere warms or cools the planet, and usually, scientists estimate that the cooling effect wins.

Figure 2: Satellite image for 1st August 2013 of dust stretching all the way across the Atlantic Ocean (from Suomi NPP VIIRS instrument)

However, the balance of cooling versus warming from dust is strongly sensitive to how big the dust particles actually are. Larger particles shift the impact of dust from more cooling towards more warming, as a result of how they interact with radiation.

Previously, scientists didn’t think that ‘giant’ particles, larger than about 20 microns – half the width of a human hair – could travel very far, or impact climate very much, so giant dust was excluded from climate models. However, new research shows that this is not the case (Ryder et al., 2019). Scientists have been flying aircraft over deserts and dusty oceans, taking new measurements of dust which can properly measure the giant particles.

Figure 3: The UK FAAM research aircraft flies through dusty air taking measurements near the Cape Verde Islands in the tropical Eastern Atlantic during the AER-D fieldwork

The results clearly demonstrate that even after more than a week in the air, the largest dust particles remain doggedly present. We would expect these particles to be falling to the ground much more quickly than this, certainly within a few days, so some additional mechanism which we don’t yet understand must be keeping them airborne.

Figure 4: Dust size measurements from aircraft flights sampling dust at different points during transport over and away from the Sahara (from Ryder et al., 2019), showing that dust particle size drops rapidly in new dust storms, but doesn’t change much during longer range transport.

Our new research estimated the warming/cooling impact of the excluded giant dust particles. Over the Sahara, climate models miss 18-26% of the interactions between dust and radiation. This means that climate models are underestimating the warming impacts of dust on the Earth-atmosphere system. This may have important effects in the future – for example, if our world becomes dustier due to desertification, dust may cause additional warming which is being missed by climate models.

References

Ryder, C. L., Highwood, E. J., Walser, A., Seibert, P., Philipp, A., and Weinzierl, B.: Coarse and giant particles are ubiquitous in Saharan dust export regions and are radiatively significant over the Sahara, Atmos. Chem. Phys., 19, 15353–15376, https://doi.org/10.5194/acp-19-15353-2019, 2019.

Posted in Aerosols, Climate, Climate modelling | Leave a comment

How can we provide accurate and useful weather forecasts to Tropical African countries?

By: Carlo Cafaro 

Probabilistic weather forecasts contain useful information that could save lives by issuing early and accurate warnings. This is especially true in Africa, where the livelihoods of people rely significantly on agricultural activities. These, in turn, depend on weather conditions – particularly rainfall. Every year, droughts and floods pose a serious threat by altering the levels of food security.

In the last months of 2019, for instance, unusually heavy rains battered East Africa, causing landslides and flooding with over 2.8 million people affected, according to the UN and at least 137 people dead in Kenya and Uganda.

Most of the annual rainfall of tropical East Africa comes from deep convective systems.  This is true also for West Africa, where 90% of the annual rainfall in the Sahelian is produced by a small number (12% of the total) of organized mesoscale convective systems (see [2] for details).

For example, convective storms usually hit Lake Victoria overnight. The lake is crucial for the local fisheries and for providing a means of transport between the local communities. However, according to the World Bank, every year up to 5000 thousand fishermen die by drowning in the lake. Sudden strong winds caused by nocturnal thunderstorms can capsize their boats, which often are overloaded or are not properly safe to use.

Figure 1: The UK and African partners of the SWIFT project. Courtesy: P. Hill

This is why there is urgent need of accurate and useful weather forecasts, from hourly to seasonal timescales. This is the overarching aim of the GCRF African SWIFT project, which I am working on. It kicked off in November 2017 and brings together UK and African scientists, operational forecasters and users to improve the weather forecasting capability of the African partners (see Figure 1 for a summary of the different partners of the project).

Numerical weather prediction (NWP) global deterministic models have too coarse a grid spacing to realistically represent convection. Hence they use parametrization techniques, which are found to be not so skillful.

This is because they tend to produce too much light rain and miss the most intense events. Another well known problem is that they cannot reproduce well the diurnal cycle of precipitation, with the peak of rainfall too early with respect to observations.

There are thus two options (not mutually exclusive) to try to improve the skill of rainfall predictions: run an ensemble of forecasts, hence producing probabilistic weather forecasts and/or decrease the size of the grid spacing down to kilometer-scale, thus allowing the convection to be explicitly represented and not parametrized (the so-called convection-permitting models, CP).

When considering an ensemble of coarse grid-box global models, the results are not so encouraging [4].

The Met Office has thus been running, since 2011, a finer deterministic grid size model over East Africa (4.4 km grid spacing). Although CP models have been shown to be more skillful than the respective coarser grid models (e.g. [1, 5]) for the Lake Victoria region, the skill is still fairly poor and so there is still room for improvement.

Hence the motivation to combine the two options previously discussed, to generate convection-permitting ensemble prediction systems (CP-EPSs). These are usually run on limited areas, with boundary and initial conditions provided by the global ensembles.

These models are fairly novel (~10 years) and have been run only in extra-tropical regions by several forecasting and research centres.

Figure 2: SWIFT news

In April 2019, the Met Office, as part of the SWIFT project, ran CP-EPSs for the tropical Africa domain for the first time. These were tested in real time during the forecast testbed hosted by the Kenya Meteorological Department (KMD) in Nairobi between 23rd April and 6 May 2019 (in Figure 2 a group picture of some of the participants).

Scientists and operational forecasters gathered in the same room, working in shifts for 24 hours. We were using nowcasting and CP ensembles forecasts to predict convective storms over West and East Africa up to 72 hours ahead. One of the morning tasks was the ‘subjective’ evaluation of these forecasts.

Participating in this real-time experiment allowed me to see the task of verifying CP-EPSs, part of my PhD work as well, from a completely different perspective.

This is because speaking directly with the local forecasters and the users in Kenya helped me shape my activities, making me think in a more user-friendly way.

First of all, it is important to be able to extract the information from the “big data” produced by the CP-EPSs. Then to communicate and verify them.

For example, over Lake Victoria, one piece of forecast information could be the probability of rainfall exceeding a threshold over a certain accumulation period. The shorter this period is, the more useful the information. Another option could be to detect the location and timing of the peak of rainfall.

Figure 3: Convective-scale probabilistic forecast map of rainfall exceeding 10 mm accumulated on 03/05/2019 between 00 and 06 UTC. Courtesy of Stuart Webster.

An appealing graphic, which can be understood by a local forecaster could then be used to communicate this information.

Figure 3 shows the map produced during the testbed of the probabilistic forecast of rainfall accumulation exceeding 10 mm over 6 hours.

Also, a map displaying the location of initiation of a thunderstorm, along with the trajectory paths of its propagation, is another possibility.

The last step, forecast verification, consists of two aspects, as stated by WMO: assessing the technical skill of a forecast (for example by producing the reliability diagram as explained  in this blog post by Jochen Broecker) and to ascertain whether the forecasts issued help the user to take decisions. This was stated in other terms also by [3] in terms of quality and value of the forecasts.

Even computing the forecast skill can be done in a user-oriented way: WMO issued a challenge to find the best way to do that [2].

In light of this, an important activity carried by the KMD was the collection of feedback of forecasts users.

This feedback will certainly help all of us involved in this project to reach our goals: make a positive impact into the lives of the African people by providing skillful weather forecasts.

References:

[1] Chamberlain, J.M., Bain, C.L., Boyd, D.F.A., McCourt, K., Butcher, T. and Palmer, S., 2014, Forecasting storms over Lake Victoria using a high resolution model. Met. Apps, 21: 419-430. doi:10.1002/met.1403

[2] Ebert, E., Brown, B. G., Göber, M., Haiden, T., Mittermaier, M., Nurmi, P., … Schuster, D.,2018. The WMO challenge to develop and demonstrate the best new user-oriented forecast verification metric. Meteorologische Zeitschrift, 27, 435-440. doi:10.1127/metz/2018/0892

[3] Lebel, T., Diedhiou, A., and Laurent, H. ( 2003), Seasonal cycle and interannual variability of the Sahelian rainfall at hydrological scales, J. Geophys. Res., 108, 8389, doi:10.1029/2001JD001580.

[4] Murphy, A.H., 1993: What Is a Good Forecast? An Essay on the Nature of Goodness in Weather Forecasting. Wea. Forecasting, 8, 281–293, doi:10.1175/1520-0434(1993)008<0281:WIAGFA>2.0.CO;2

[5] Vogel, P., P. Knippertz, A.H. Fink, A. Schlueter, and T. Gneiting, 2018: Skill of Global Raw and Postprocessed Ensemble Predictions of Rainfall over Northern Tropical Africa. Wea. Forecasting, 33, 369–388, doi:10.1175/WAF-D-17-0127.1

[6] Woodhams, B.J., C.E. Birch, J.H. Marsham, C.L. Bain, N.M. Roberts, and D.F. Boyd, 2018: What Is the Added Value of a Convection-Permitting Model for Forecasting Extreme Rainfall over Tropical East Africa?. Mon. Wea. Rev., 146, 2757–2780, doi:10.1175/MWR-D-17-0396.1

Posted in Africa, Climate, Convection, Weather forecasting | Leave a comment