Weather, Climate and Energy

By David Brayshaw

 A secure, reliable and relatively affordable electricity supply is an essential part of modern life in many parts of the world.  When charging an iPad, for example, one expects that power will be available at the flick of a switch and rarely considers the complex chain of events involved in producing and conveying the electricity to the device.

The production of energy – including electricity – accounts for a large part of national and global greenhouse gas emissions.  Decarbonizing the energy sector is therefore a key ingredient in meeting climate change targets both in the UK and worldwide.  Electricity – widely seen as the easiest form of energy to decarbonize – has seen a rapid shift towards renewables, particularly wind and solar photovoltaic (PV).

In a national-scale electricity system it is difficult to store large volumes of power efficiently and as such a near-instantaneous balance between supply and demand must be maintained.  Thuswhile the growing use of weather-sensitive renewable generation is clearly beneficial in terms of reducing carbon emissions, it presents new challenges for electricity system management as electricity production cannot be controlled to the same extent as that from, for example a traditional gas-fired power station. 

Figure 1: Simulated changes in weather sensitivity of the electricity system in Great Britain as increasing amounts of wind power capacity is installed.  The red line shows the strength of the negative relationship between temperature and annual total demand-net-wind (primarily associated with cold periods driving high demand), which decreases as installed wind capacity increases.  The blue line, in contrast, shows the increasing strength of the relationship between wind-speed and demand-net-wind as installed wind capacity increases (the sign is also negative because more wind implies a lower demand-net-wind).  Recent installed wind capacities for the GB electricity system are marked in grey lines, suggesting that the system’s dominant weather sensitivity changed from being temperature to wind sometime between 2009 and 2015.  The analysis is based on a 36-year modelled “reconstruction” of the power system’s behaviour derived from weather records.  Adapted from Bloomfield et al. (2018)1

This can be illustrated with an example.  It has long been known that electricity demand is strongly sensitive to temperature: as temperatures fall in winter, electricity demand rises.  Recent research, however, suggests that in the UK the impact of year-to-year variations in temperature via demand are now typically smaller than year-to-year variations in wind power supply (as shown in the figure 1).  That is, in terms of the amount of electricity that must be supplied by coal, gas and nuclear power stations, the impact of a calm winter now exceeds a cold winter1 while winter-time weather patterns associated with both cold and still conditions lead to a double whammy (both high demand and low supply2,3,4).  This suggest an important but subtle shift in how the weather risk should be viewed.  To assess how much non-renewable generation capacity is required, it makes more sense to ask what the maximum demand-net-renewables is (i.e., the demand remaining once the contribution from renewables is removed).  Such a calculation is easier than quantifying the minimum renewable generation available during times of peak demand (the so called “capacity credit” of renewables).

The need for high-quality tools for assessing and understanding the impact of weather and a changing climate on the energy system has never been greater.  The Energy-Meteorology group5, working closely with its industrial and academic partners, is leading the development of many such tools: from the first use of multi-decadal “energy system data reconstructions”6, to the development of new European climate services for energy (weeks-to-months ahead7,8,9), and the exploitation of state-of-the-art high-resolution climate data for informing the design of climate-resilient power systems in the coming decade10,11.


  1. Bloomfield, H. C., Brayshaw, D. J., Shaffrey, L. C., Coker, P. J., Thornton H. E., 2018: The changing sensitivity of power systems to meteorological drivers: a case study of Great Britain. Res. Lett., 13 (5), 054028,
  2. Thornton, H. E., Scaife, A. A., Hoskins, B. J., Brayshaw, D. J., 2017: The relationship between wind power, electricity demand and winter weather patterns in Great Britain. Res. Lett., 12 (6), 064017,
  3. Ely, C. R, Brayshaw, D. J., Methven, J., Cox, J., Pearce, O., 2013: Implications of the North Atlantic Oscillation for a UK–Norway renewable power system. Energy Policy, 62, 1420-1427,
  4. Brayshaw, D. J., Dent, C., Zacharay, S., 2012: Wind generation’s contribution to supporting peak electricity demand: meteorological insights. Institution of Mechanical Engineers, Part O: Journal of Risk and Reliability, 226 (1), 44-50,
  5. University of Reading, 2018: Research Themes – Energy Meteorology. Accessed 6 December 2018,
  6. Cannon, D. J., Brayshaw, D. J., Methven, J., Coker, P. J., Lenaghan, D., 2015: Using reanalysis data to quantify extreme wind power generation statistics: a 33 year case study in Great Britain. Renewable Energy, 75, 767-778,
  7. Thornton, H. E., Scaife, A., Hoskins, B. J., Brayshaw, D. J., Smith, D., Dunstone, N., Stringer, N., Bett, P. E., 2018: Skilful seasonal prediction of winter gas demand. Environmental Research Letters. Res. Lett., (in press).
  8. C3S, 2018: European Climatic Energy Mixes. Accessed 6 December 2018,
  9. Earth System Services, 2018: S2S4E Climate Services for Clean Energy. Accessed 6 December 2018,
  10. Santos-Alamillos, F. J, Brayshaw, D. J., Methven, J., Thomaidis, N. S., Ruiz-Arias, J. A., Pozo-Vázquez, D., 2017: Exploring the meteorological potential for planning a high performance European Electricity Super-grid: optimal power capacity distribution among countries. Res. Lett., 12 (11), 114030,
  11. PRIMAVERA, 2018: Primavera: User Interface Platform. Accessed 6 December 2018,


Posted in Climate, Energy meteorology, Renewable energy | Leave a comment

Steps To Make Sea Ice Projections More Robust

By David Schroeder

Climate model projections are our best approach to make predictions about future sea ice in both hemispheres for the whole 21st century. Given the large spread between individual model projections and distinct discrepancies between model results and satellite-based observations for the last decades, the question arises how accurate these projections are. The reliability of such simulations depends on how well key processes are captured in the climate model. CMIP5 (Coupled Model Intercomparison Project Phase 5) models generally simulate the observed sea ice decline better than the older CMIP3 models indicating that, for example, the improved representation of sea ice processes is important (Stroeve et al., 2012).

In Danny Feltham’s sea ice group we have added and improved several sea ice processes in the Los Alamos sea ice model CICE which is used in many climate models (e.g. the current U.K. Met Office model HadGEM3). The most important contributions are a physical melt pond model describing the evolution of melt ponds (Flocco et al., 2012; Schroeder et al., 2014), an elastic anisotropic plastic rheology accounting for the subcontinuum anisotropy of the sea ice cover (Tsamados et al., 2013) and a spatially and temporally variable form drag parameterization for the momentum exchange between ice and atmosphere as well as ice and ocean (Tsamados et al., 2014).

These new developments increase the realism of the sea ice model. However, incorporating these developments in a climate model does not result in better climate simulations as a matter of course. Many physical and empirical parameters are not very well constrained in a sea ice model requiring a tuning process applying observational data. In the past, most of the tuning has been based on sea ice concentration data derived from passive microwave imagery. This is a problem because realistic sea ice thickness is crucial for simulating variability and trends of the sea ice cover. Using radar altimetry, sea ice thickness estimates are available from CryoSat-2 (CS2) (e.g. from the Centre for Polar Observation and Modelling) during the ice growth seasons since 2010. However, the discrimination of ice and open water, the discrimination of ice type, retracking radar waveforms to obtain height estimates, constructing of sea surface height beneath the ice, and estimating the depth of the snow cover result in a grid cell ice thickness uncertainty of about 25 %.

Figure 1:  Mean effective sea ice thickness over read region shown in inlet above. Verification of CICE-default and CICE-best with Cryosat CS2 CPOM. Results from the ocean sea-ice model PIOMAS are added for comparison. Red area is defined as region where impact of winter growth on thickness change dominates other factors and where CS2 data are most accurate.

In a current study we selected an optimal region for comparing sea ice thickness between simulations with the sea ice model CICE and CS2 data (see red region in Figure 1) by taking into account the strengths and weaknesses of both approaches excluding locations where the number of CS2 observations is limited and the sea ice is thin (Schroeder et al., 2018). Figure 1 shows that a default CICE simulation (blue) strongly underestimates ice thickness, but a new CICE setup (red) agrees well. The most important impact comes from implementing a new process: accounting for the loss of drifting snow. Our optimal model configuration does not only improve the simulated sea ice thickness, but also summer sea ice concentration, melt pond fraction, and length of the melt season. In spite of its uncertainties we successfully applied CS2 sea ice thickness data to improve sea ice model physics.

Reducing uncertainty in sea ice projections requires better representation of key processes, not only of the sea ice component, but also in the atmosphere, the ocean and regarding the coupling. Polar cloud formation and the interaction with the sea ice surface has been identified as a key process which is currently not represented satisfactorily.


Flocco, D., D. Schroeder, Feltham D. L., Hunke E. C., 2012: Impact of melt ponds on Arctic sea ice simulations from 1990 to 2007. J. Geophys. Res., 117, C09032,

Schroeder, D., Feltham D. L., Flocco D., Tsamados M., 2014: September Arctic sea-ice minimum predicted by spring melt-pond fraction. Nat. Climate Change, 4, 353-357,

Schröder, D., Feltham D. L., Tsamados M., Ridout A., Tilling R., 2018: New insight from CryoSat-2 sea ice thickness for sea ice modelling. The Cryosphere Discuss. (in review),

Stroeve, J. C., Kattsov V., Barrett A., Serreze M., Pavlova T., Holland M., Meier W. N., 2012: Trends in Arctic sea ice extent from CMIP5, CMIP3 and observations. Geophys. Res. Lett., 39, L16502,

Tsamados, M., Feltham D. L., Schroeder D., Flocco D., Farrell S. L., Kurtz N., Laxon S. W., Bacon S., 2014: Impact of Variable Atmospheric and Oceanic Form Drag on Simulations of Arctic Sea Ice. J. Phys. Oceanog., 44, 1329-1353,

Tsamados, M., Feltham D. L., Wilchinsky A., 2013: Impact of a new anisotropic rheology on simulations of Arctic sea ice. J. Geophys. Res.: Oceans, 118, 91-107,


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

Seasonal Forecasting and the 2018 European Heatwave

By Len Shaffrey

The summer of 2018 has been one of the warmest on record in the UK and Europe. Warm temperatures over the summer led to impacts on agriculture, water resources and human health. One interesting question is how predictable was the 2018 European summer heatwave?

The skill of wintertime seasonal forecasts has dramatically improved in the past few years, especially for forecasts of the wintertime North Atlantic Oscillation (e.g. Scaife et al. 2014, Baker et al. 2018). However, there is much less skill for summertime seasonal forecasts for Europe. Interestingly, there was a substantial degree of convergence in the seasonal forecasts for summer 2018, with most forecast systems predicting warmer than average European temperatures. Figure 1 shows the multi-model mean seasonal forecasts of 2018 July-August-September(JAS) 2m temperature anomalies (forecasts were initialised around the 1 June 2018). The seasonal forecasts were predicting temperature anomalies of approx 1C over Eastern Europe. This led the Met Office to make the unusual step of issuing a statement on this summer’s forecast, which was picked up by the media.

Figure 1. Multi-model mean seasonal forecast of 2018 JAS 2m temperature anomalies from the Copernicus C3S seasonal forecast service.

It’s not yet clear why the summer of 2018 was more predictable than usual. This is an active research area which is being addressed in NERC projects such as SUMMERTIME and IMPETUS. Recent studies have highlighted how springtime North Atlantic seas surface temperature drive summertime European atmospheric circulation (Osso et al. 2018), the relationship between springtime Tropical Atlantic rainfall anomalies and summertime circulation over the North Atlantic (Wulff et al. 2017) and suggested that some seasonal forecasting systems can capture interannual variations in European summertime rainfall (Dunstone et al. 2018). Overall, these studies have suggested that there might be much more seasonal predictability for summertime European climate than previously thought. The studies also raise the possibility that deeper understanding of these processes may lead to substantially improved seasonal forecasts of summertime European climate.


Baker, L. H., Shaffrey, L. C., Sutton, R. T., Weisheimer, A. and Scaife, A. A, 2018: An intercomparison of skill and overconfidence/underconfidence of the wintertime North Atlantic Oscillation in multimodel seasonal forecasts. Geophysical Research Letters, 45 (15), 7808-7817.

Dunstone, N., Smith, D., Scaife, A., Hermanson, L., Fereday, D., O’Reilly, C., et al. 2018:. Skilful seasonal predictions of summer European rainfall. Geophysical Research Letters, 45, 3246–3254,

Osso, A., Sutton, R., Shaffrey, L. and Dong, B. 2018: Observational evidence of European summer weather patterns predictable from spring. Proceedings of the National Academy of Sciences of the United States of America, 115, 59-63. 

Scaife, A. A., et al. 2014: Skillful long‐range prediction of European and North American winters. Geophys. Res. Lett.,41, 2514–2519. doi: 10.1002/2014GL059637

Wulff, C. O., Greatbatch, R. J., Domeisen, D. I. V., Gollan, G., & Hansen, F. 2017:. Tropical forcing of the Summer East Atlantic pattern. Geophysical Research Letters, 44, 11,166–11,173.

Posted in Atlantic, Atmospheric circulation, Climate, Climate change, Climate modelling, Environmental hazards, Historical climatology, Hydrology, Numerical modelling, Seasonal forecasting, Waves | Leave a comment

Is it a normal season this year for tropical cyclones in the Western North Pacific?

By Xiangbo Feng

 The Western North Pacific (WNP) is the most active area for tropical cyclones (TCs).  The number of TCs occurred in the WNP so far (end of October) this year is 26 – just the average number of annual TCs over 1980-2017 (Figure 1 upper). As TC season in the north Hemisphere usually finishes up in October, we thus might think we are having another normal or slightly above-normal TC season for TC occurrence in the WNP.

Figure 1: Top: Western North Pacific (WNP) (0-60°N, 100-180°E) tropical cyclone (TC) count during the period 1980-2018. The 2018 TC record is updated to 31st October. Black, blue and green bars show annual, early summer (June-August, JJA) and late summer (September-October, SO) TC counts, respectively, while red bar is for the number of TCs that occur just in the northwestern sector (20-40°N, 120-140°E) of the WNP. Dashed lines are the long-term averages over 1980-2017. Bottom: Monthly Nino3.4 index anomaly during 1980-2018. Nino3.4 is the average sea surface temperature anomaly in the region of 5°N-5°S, 170°-120°W. Data sources: TC data are the Best Track observational TC record, retrieved on the 5th October 2018, from the Japan Meteorological Agency. Nino3.4 is obtained from NOAA PSD , retrieved on the 5th November 2018.

But, statistics from observations show a few interesting points that make this year not normal for TC activities(Only the TCs of tropical storm intensity or greater are considered here, i.e. maximum sustained wind speed ≥ O34 knots). 

  • First, it’s an active early summer (June-August, JJA) for TCs in the WNP (Figure 1 upper). In JJA, the long-term average number of named TCs over this area during 1980-2017 is 11, whilst the past early summer saw 17 tropical storms being named, 6 above the average, making it the second largest (it’s just 1 TC less than in 1994) on record since 1980.
  • Second, it then turned into a quiescent late summer (September-October) for TC occurrence. Late summer is usually a busy period when strong TCs (e.g. super typhoons) occur. The average number of TCs occurred during late summer is 8. There were only 6 TCs declared in the past late summer. More interestingly, the past October was very quiet in the WNP, featured with only one TC (Typhoon Yutu, formed on 21st October) declared. So far, it seems that some TCs, which were supposed to occur in a later time, were formed in an earlier time, making it a very busy early summer.

Figure 2: TC occurrence anomaly (with long-term average removed) in JJA for each year in the Western North Pacific. Bottom right shows the average of TC occurrence in JJA over 1980-2018.

  • Third, at regional scale (Figure 2), we see the past early summer is actually an extremely busy period for TCs especially in the northwestern parts (20-40°N, 120-140°E). The climatology of the number of TCs passing through this region in early summer is 6 (also see Figure 1 upper). But, there were 12 TCs coming through this region in JJA this year, double the average number. In relation to the normal path of TCs, e.g. through the Philippine Sea and Taiwan (Figure 2 bottom right), this year’s TCs tend to shift northward heading to the north of China and the west of Japan. Similar path years are 1994, 2002 and 2004, but they have smaller positive anomalies.

Seasonal and sub-seasonal TC forecasts

Figure 3: Correlations between JMA observed TC occurrence anomaly and Nino3.4 index, for annual values (top), early summer (June-August, JJA) values (bottom left) and late summer (September-October, SO) values (bottom right), over 1951-2018. Dot areas are where the correlations pass the 0.1 significance test (90% confidence level).

It has been well documented that the El Niño–Southern Oscillation (ENSO) plays an important role in the inter-annual variability of TC activity (especially for the TC lifetime) via modulating the large-scale atmospheric environments (e.g. wind shear and relative humidity). As the effect on number of TCs, for example (Figure 1), in 1998, when a strong El Nino event was developed, fewer TCs are formed in the WNP over the whole year; in the 2015-2016 El Nino event, TCs are more likely to recurve to north rather than straightforwardly head to west. The teleconnections of TC occurrence with ENSO phases can be simply described by the correlation coefficient map (Figure 3 top).

Some meteorological services and agencies have included ENSO as one main source of predictability for the WNP TC activities in their seasonal forecast outlooks. Among these services are the GCACIC forecast at the City University of Hong Kong and the TSR forecast at the University College London. Before TC season started this year, they had forecasted that this year would be a normal or slightly above-normal year for the WNP TC occurrence. This is mainly due to the cold ENSO-neutral condition (Figure 1 bottom). In the meantime, the Met Office’s dynamical seasonal forecast model GloSea5 also predicts a normal or slightly below-normal season for the WNP TC occurrence (contact me if you want to see this result). It seems like that those seasonal forecasts did a reasonably good job for predicting the total number of TCs for the whole season.

However, none of those forecasts, as far as I know, had predicted either an active early summer or a quiescent TC late summer for TCs as observed, even with a short forecast lead time. This means that the present seasonal forecast models (both statistical and dynamical) have less ability to predict the TC activities at sub-seasonal timescale (a time range between two weeks to two months). One of reasons for lack of such predictability might be due to the TC activities, which are ENSO-dependent in late summer, but are ENSO-independent in early summer (Figure 3 bottom). Even though the ENSO states are reasonably estimated, it does not necessarily mean that the most part of TC occurrence variability at sub-seasonal timescale can be well predicted.

The apparent sub-seasonal variations of TC activities, at least this year in the WNP, indicate the necessity of developing sub-seasonal TC forecast approaches on top of the normal seasonal forecast. Including some factors at sub-seasonal timescales, such as the Madden–Julian Oscillation and the equatorial waves, may help us to build up such forecast skills. This indicates an area where more methodological work needs to be carried out in the future. Recently, some efforts have been made to address this challenge, such as the WWRP/WCRP Sub-seasonal to Seasonal Prediction Project (S2S)

Here in Reading, the FASCINATE project, one of the Weather and Climate Science for Service Partnership (WCSSP) Southeast Asia projects funded by the Newton Fund (Nick Klingaman is the PI), is carrying out a detailed analysis of TC activities in the WNP. One aim of our project is to identify the sources of predictability for the WNP TCs, from synoptic to seasonal timescales, and evaluate the present forecast systems, such as the Met Office’s Numerical Weather Prediction and GloSea5, on the representations of such prediction sources.

Posted in Atmospheric circulation, Climate, Climate modelling, earth observation, ENSO, Equatorial waves, Historical climatology, Madden-Julian Oscillation (MJO), Numerical modelling, Seasonal forecasting, Tropical cyclones, Waves, Weather forecasting, Western North Pacific, Wind | Leave a comment

Characteristics and enhanced quality control of drifting buoy observations of sea surface temperature from the International Comprehensive Ocean Atmosphere Data Set (ICOADS)

By Simone Morak-Bozzo

Over the last two decades drifting buoys have become the most prevalent in situ measurement method for sea surface temperature(SST). Drifting buoy data are particularly popular because of their high spatial and temporal coverage. Their freedom of movement allows them to provide data offside shipping routes and moorings and the high temporal resolution gives insight not only into seasonal but also diurnal variations of SSTs, as shown in Morak-Bozzo et al, 2015.

Drifting buoy observations need to be verified before they can be used for quantitative studies involving SST data and cleaned from artefacts linked to erroneous observations or faulty instruments. The main archive of drifting buoy observations is the International Comprehensive Ocean Atmosphere Data Set (ICOADS). In our current project we investigate the life story of approximately 26000 buoys from 1979 to 2014, from the time of their deployment to their “death”. We also observe and document how the network of drifters evolves over time and we assess the quality of the data.

Figure 1: For the subperiod 1982-2012 and a total of 22678 buoys, the position of deployment is shown on the left and the position of their last signal on the right, gridded on a 2-by-2 degree grid.

Figure 2: This figure is showing the distribution of the buoy lifetime for all the buoys in the ICOADS archive. We can see that less than 40% of the buoys make it past the first year after deployment and less than 20% survives more than two years.

Figure 3: The maps are showing the cumulative number of drifting buoy observations in each 2-by-2 degree box for seven 5-year periods (1980-1984, 1985-1989, etc.). In the top right corner of each subfigure we show the sum of all observations in that 5-year period. The majority of buoys reports at least every 3 hours, if not hourly.

An important step to determine the reliability of the data is the creation of an automated protocol for quality control (QC) to complement the limited QC product provided by ICOADS. The set of algorithms we developed can detect e.g. stranded and misplaced buoys or non-plausible SST reports.

Compared to an independent data set of SSTs, such as Reynolds AVHRR-only (Reynolds et al. 2007, Banzon et al. 2016) the ensemble of QC flags shows a clear improvement of the quality of the drifting buoy data set.

The resulting data and associated uncertainty will be contributing to a new SST reconstruction dating back to 1850, under the project HOSTACE (Historical Ocean Surface Temperatures: Accuracy, Characterisation and Evaluation).


Morak-­Bozzo, S., Merchant, C. J., Kent, E. C., Berry, D. I. and Carella, G. ,2016:Climatological diurnal variability in sea surface temperature characterized from drifting buoy data. Geoscience Data Journal, 3 (1),  20­-28, doi:

Banzon, V., Smith, T. M., Chin, T. M., Liu, C., and Hankins, W., 2016: A long-term record of blended satellite and in situ sea-surface temperature for climate monitoring, modeling and environmental studies. Earth Syst. Sci. Data8, 165–176, doi:10.5194/essd-8-165-2016.

Reynolds, R. W., T. M. Smith, C. Liu, D. B. Chelton, K. S. Casey, and M. G. Schlax, 2007: Daily high-resolution-blended analyses for sea surface temperature. Journal of Climate20, 5473–5496, doi:10.1175/JCLI-D-14-00293.1.

Posted in Climate, Climate change, Data collection, Data processing, earth observation, Measurements and instrumentation | Leave a comment

Smoke, science, and sharks

By Ross Herbert

In the August of 2017 the Cloud-Aerosol-Radiation Interactions and Forcing – Year 2017 (CLARIFY) measurement campaign took place on a tiny island in the middle of the southeast Atlantic Ocean where we were surrounded by whales, sharks, and most importantly, stratocumulus clouds. During August and September dense layers of strongly absorbing smoke from biomass burning in central Africa were transported over these semi-permanent clouds. These events provided us with a unique opportunity to understand the interactions between the cloud, smoke layers, and radiation; processes which remain key uncertainties within our understanding of global aerosol radiative forcing, cloud feedbacks, and, ultimately, climate change. In this blog I will discuss my experience of participating in the measurement campaign and also outline the high-resolution cloud modelling work that I am currently doing.

The island (with Bear Grylls)

The Ascension Island is situated at 8°S 14°W in the middle of the Atlantic Ocean, 1600km from the coast of Africa and 2250km from Brazil. The island, with a diameter ~ 10km, is a volcanic island natively populated by nesting turtles, colourful land crabs, and (once) huge bird colonies, and non-natively by rats (that appeared with the boats), cats (that were let loose on the island to control the rats but decided to eat birds instead and are now outlawed), and donkeys (that are tolerated). In the mid-nineteenth century, following encouragement from Charles Darwin, the royal navy began bold plans to increase precipitation on the island by ‘greening’ the upland reaches of the largest mountain with anything that would grow. The result: bamboo, bananas, wild ginger, and guava, and the birth of the aptly named ‘Green Mountain’; however, most of the island remains volcanic and devoid of vegetation.

Figure 1:On top of Sister’s peak looking towards Green Mountain

The measurement campaign

My role on the month-long campaign was a mission scientist. Along with several others we worked as a team alongside the pilots to plan the sorties and then join the crew onboard the FAAM BAe-146 research aircraft. For the sorties we would use forecasts and satellite observations to plan flights depending on what our objectives were; this might include incoming plumes of smoke, co-ordinated satellite overpasses, inter-comparison flights, and interesting cloud or convective features.

Figure 2: The FAAM BAe-146 (right) and NASA P3 (left) on Wideawake Airfield

The pilots required detailed flight plans of bearings, altitudes, distances, and times, so the planning would often take several hours. On flight days we would be up at 6am to check the most recent observations and make any last-minute changes to the sortie plan before heading to the airfield. During flights, the mission scientists used real-time data and the expertise of the instrument operators to fine-tune the sortie so that we were focusing on the correct feature. Smoke layers and stratocumulus-to-cumulus transitions are very poorly forecast so we had to make rapid decisions based on what we observed during flight. This information would be relayed to the mission scientist who sat in the cockpit and would ultimately make the final decision before informing the pilots.

This slideshow requires JavaScript.

The campaign was a huge success; we flew 28 sorties totalling 99 hours and collected data in every cloud-aerosol-radiation regime we hoped for and more. As well as the standard atmospheric state measurements (temperature, pressure, relative humidity etc..) we made detailed measurements of the clouds, the particles, and the gases within the atmosphere and smoke layers. This included number, size, and composition, and also radiative properties such as ability to scatter and absorb radiation. The measurements will help towards improving our understanding of how the smoke layers are formed, transported, evolve, and also how they affect radiative fluxes and interact with the stratocumulus clouds. We can also use the measurements to improve our representation of smoke layers and related processes in the computer models that help us predict the climate and forecast the weather. FYI the aircraft data will be available on the Centre for Environmental Data Analysis (CEDA)archive (ask if you want more specifics).

End of campaign. Time to go back to the office

So enough of the campaign, what am I actually doing on the project? I am working with Nicolas Bellouin and Ellie Highwood to investigate what happens to the cloud radiative properties (i.e., how much energy from the sun and the earth’s surface passes through and out of the cloud) when the layer of smoke is elevated above the cloud; this is a scenario that we encountered numerous times during the campaign and that we observe in satellite observations. Previous studies suggest that the smoke may act to thicken the clouds, which makes them brighter and more reflective to incoming solar radiation (sunlight), thus acting to cool the climate. The stratocumulus clouds and smoke layers cover vast areas of the ocean, therefore small changes to the radiative properties of the cloud may have important impacts. My work will help us understand this effect in more detail and understand how sensitive it is to the properties of the smoke layer and cloud.

Smoke above cloud + Sunlight = Heat = Cloud response = Semi-direct radiative effect = Cooling? or Warming?

In my work I am using the MetOffice Large Eddy Model (the LEM) to simulate the evolution of a stratocumulus cloud deck with an elevated layer of absorbing smoke. The LEM has very high spatial resolution (imagine dividing the atmosphere into boxes – high resolution means lots and lots of little boxes) that allows us to simulate the main turbulent motions that drive the movement of energy, winds, and moisture in the atmosphere (typical climate-scale models have spatial resolutions ~ 200 times greater). It is also coupled to a radiation scheme that allows us to represent upward and downward fluxes of radiation through the atmosphere, smoke layer, and cloud. From this we can represent the additional heating that is caused by the smoke and any subsequent changes to the cloud field and radiative properties; we can then finally determine whether the smoke is acting to cool or warm the climate. This specific radiative effect is commonly referred to as the semi-direct effect and in the context of the IPCC AR5 report can be seen as a rapid adjustment to the instantaneous radiative effect of aerosols.

These simulations are giving us very interesting results that are currently being written up for publication (sneak preview: it’s not as simple as we previously thought). Watch this space!

A longer version of this blog can be found at

Posted in Africa, Atlantic, Atmospheric chemistry, Atmospheric circulation, Atmospheric optics, Climate, Climate change, Climate modelling, Clouds, Data collection, earth observation, Energy budget, Environmental hazards, Greenhouse gases, Measurements and instrumentation, Microphysics, Numerical modelling, Solar radiation, Weather forecasting, Wind | Leave a comment

Determining the Earth’s energy and water cycles

By Christopher Thomas

The Earth’s energy and water cycles govern the distribution and movement of energy and water in the atmosphere, oceans and land. Both energy and water are constantly being transported between different regions of the globe, and the two cycles are closely linked (‘coupled’) together. It is very important to determine the size and variability of these transports as well as the long-term average distributions of each quantity. This will enable us to detect trends due to climate change, as well as to provide insight into short term climatic variability such as the El Niño Southern Oscillation.

The Sun provides energy to the Earth via solar radiation; simultaneously, some energy is emitted back into space. It is therefore quite common to consider the energy imbalance at the top-of-atmosphere (TOA). There is currently a small downward imbalance which means the Earth is gaining heat. The incoming radiation is distributed between the atmosphere, land and oceans both vertically and horizontally. Water is transported around the globe by (e.g.) ocean currents, rivers, evaporation, and precipitation. The last two processes provide the crucial link between the energy and water cycles. When water changes state from vapour to liquid or vice versa it either releases or absorbs energy in quantities which are significant enough that they must be included in the energy cycle.

Our aim is to combine a variety of Earth observation (EO) data sets in order to determine the energy and water cycles with as much precision as possible. This work was pioneered by the NASA NEWS team (L’Ecuyer et al. 2015, Rodell et al. 2015) and as a first step we have reproduced their work. The Earth is divided into 16 regions (seven land and nine ocean) as shown in Figure 1. In each region we consider the energy and water content both in the atmosphere and at the surface, as well as the TOA radiation imbalance (no water is lost to space).

Figure 1: The regions used in this study.

Several complementary EO measurements of energy and water transport are combined in each of these regions. These measurements are expressed as fluxes (the rate of flow of energy through an area). The larger the flux in a particular region, the more energy is being added to (or removed from) that region. Both vertical and horizontal fluxes are exploited in this method. Figure 2 shows the initial net vertical energy flux at the Earth’s surface in each of the 16 regions. The initial net flux is obtained by simply adding the downward fluxes together and subtracting the upward fluxes from the total. The values shown in the figure are the average of about a decade’s worth of observations.

Figure 2: Initial values of net surface flux in each region. Positive values indicate heat gain from the atmosphere and negative values indicate heat loss to the atmosphere.

Now, we know that there are various physical balances that should be respected (such as conservation of energy). The measured fluxes don’t satisfy these balances so we have to modify each flux by adding or subtracting a particular amount to it. When doing this it is particularly important to consider the uncertainty on each flux, which quantifies how well we (think we) have measured it. Observations that are poorly measured have large uncertainties and will be allowed to move more than those that are well known. For example, the TOA radiation has been extremely precisely determined, so has a small uncertainty, but the horizontal movement of water in the atmosphere is poorly understood and has a large uncertainty. The latter can therefore be moved relatively further from its initial value when trying to ensure all the physical balances are respected.

The net surface fluxes obtained in this procedure are reproduced in Figure 3. Some features of the solution indicate that it may be possible to make improvements. For example, the North Atlantic (region 13 in Figure 1) is losing less heat than may be expected. A large amount of heat is transported northwards from the South Atlantic by a process known as the Atlantic Meridional Overturning Circulation (see e.g. Buckley and Marshall 2016). This sort of discrepancy has motivated us to look into ways to incorporate extra information into the solution and hopefully mitigate the problems. One method which seems promising is to account for large-scale correlations in satellite measurements of heat loss from the ocean to the atmosphere (technically known as latent and sensible heat fluxes). If the flux measurements over the oceans are positively correlated then they are more likely to all be measured too high or low at once. This could arise due to the data processing methods used to determine the fluxes from satellite data, for example.

Figure 3: Output values of net surface flux in each region. Positive values indicate heat gain from the atmosphere and negative values indicate heat loss to the atmosphere.

We have estimated the heat flux correlations over the oceans and if we include them in the fit we find that more heat is transported from the South to the North Atlantic, which consequently loses more energy to the atmosphere. This is a promising result and we are currently preparing a paper to describe it more fully. We also intend to explore additional possibilities such as using additional data sets, including reanalysis constraints, and enhancing the spatial resolution of the solution by dividing the globe into more regions. In the long term we have ambitions to include the carbon cycle alongside the energy and water cycles.


L’Ecuyer et al., 2015: The Observed State of the Energy Budget in the Early Twenty-First Century. J. Clim. 28(21), 8319–8346, 10.1175/JCLI-D-14-00556.1

Rodell et al., 2015:The Observed State of the Water Cycle in the Early Twenty-First Century. J. Clim. 28(21), 8289–8318,  10.1175/JCLI-D-14-00555.1

Buckley and Marshall, 2016: Observations, inferences, and mechanisms of the Atlantic Meridional Overturning Circulation: A review. Rev. Geophys. 54(1), 5–63, 10.1002/2015RG000493

Posted in Climate, Climate change, Data processing, earth observation, Energy budget, ENSO, Hydrology, Solar radiation, Water cycle | Leave a comment

Storylines of regional climate change

By Giuseppe Zappa 

An outstanding question for climate science is quantifying how global warming will regionally affect the aspects of climate that are most directly relevant to society, such as precipitation, windiness and extremes. But achieving this task is proving not to be simple. The main available tool consists in computer simulations performed using ensembles of climate models. These models are used to run scenarios in which greenhouse gas concentrations increase with time, so that the climate response to warming can be evaluated. But, for the aspects of climate that are controlled by the atmospheric circulation, there still remains substantial spread across the model projections thus leading to uncertainties in how regional climate will respond to global warming.

Figure 1:Decadal evolution of cold season (November to April) Mediterranean precipitation as a function of global warming in the simulations from 36 climate models from the 5th phase of the Coupled Model Inter-comparison Project (CMIP5) for the RCP8.5 emissions scenario. Precipitation and temperature changes are evaluated relative to the 1960 to 1990 mean. The thick line shows the multi-model. For presentation purposes, the horizontal axis starts at 0.2 K. 

Let’s consider Mediterranean precipitation as an example. Figure 1 shows that while the average of 36 climate model projections, as well as most of the individual models, indicate a future decline in winter Mediterranean precipitation, the magnitude of the precipitation reduction, even for a given warming of the planet, remains highly uncertain. Notably, the projected drying at 2 degrees warming in some models can be larger than the drying at 4 degrees warming in other models. Taking the multi-model mean provides a simple, and often adopted, approach to summarise the ensemble and communicate the regional projections to stakeholders and decision makers. But substantial information on the uncertainty is lost by simply averaging the model responses. So is this fully justified?

In a recent paper, Zappa and Shepherd propose to use an alternative storyline approach to characterise the uncertainty in regional climate projections from ensembles of climate models. To think in terms of storylines, it is necessary to realise that regional atmospheric circulation can be driven by remote aspects of climate. This is true on seasonal timescales, for example in response to the development of El Nino or La Nina events in the tropical Pacific, but it is equally true for the long timescales associated with climate change. In particular, Zappa and Shepherd identify two remote drivers of atmospheric circulation whose response to climate change is both uncertain and capable of influencing the European and Mediterranean climate: the magnitude of the upper tropospheric warming in the tropics and the strength of the Northern Hemisphere stratospheric vortex.

Figure 2:Four different plausible storylines of cold season Mediterranean precipitation change per degree of global warming. The different  storylines depend on the magnitude of the tropical amplification of global warming and on the strength of the stratospheric vortex response as indicated above the figures. See Zappa and Shepherd 2017 for more details. 

By applying a statistical framework to the climate models output, four different plausible storylines of Mediterranean precipitation change are identified for different combinations in the two remote drivers responses (see Figure 2). The patterns of regional precipitation change per degree of global warming within each storyline are found to be rather diverse: depending on the storyline, the Mediterranean precipitation response can be larger (Figure 2b) or weaker (Figure 2c), or it can be more focused on the eastern (Figure 2d) or on the western (Figure 2a) Mediterranean. A worst case storyline of Mediterranean climate change is identified for a large tropical amplification of global warming and a strengthening of the stratospheric vortex (Figure 2b), in which case the Mediterranean drying per degree of global warming is expected to be particularly enhanced. 

Until there is sufficient physical understanding or observational evidence to discard one of the above combination of driver responses, these four storylines should be considered equally plausible future realisations of Mediterranean regional climate change. It is worth to highlight that, until discarded, the worst case storyline could still be realised. This should be kept in mind when evaluating the risks of climate change and developing local adaptation plans. 


Zappa, G. and T.G. Shepherd, 2017: Storylines of atmospheric circulation change for European regional climate impact assessment. Journal of Climate,30, 6561-6577,

Posted in Atmospheric circulation, Climate, Climate change, Climate modelling, Greenhouse gases, Numerical modelling, Stratosphere, Troposphere | Leave a comment

How the Hadley Cells work

By Gui-Ying Yang

The Hadley Cell, named after British meteorologist George Hadley who discovered this tropical atmospheric overturning circulation, is one of the basic concepts in weather and climate. Figure 1 shows the zonal mean overturning circulation in a latitude height plane for Boreal summer June-July-August (JJA), based on 30 years (1981-2010) of ECMWF data. It is seen that the JJA Hadley Cell is dominated by its ascent near 10°N and descent near 20°S, with motion towards the summer hemisphere near the surface and a return flow towards the winter hemisphere in the upper troposphere. This classic picture of the zonally averaged Hadley Cell gives a smooth impression of the cross-equatorial flow moving from one hemisphere to the other. The basic theories of the Hadley Cell are based on angular momentum conservation with the additional consideration of some mixing and friction near the surface. However, angular momentum conservation from zero velocity at the equator moving to another latitude, φ, gives a zonal wind u=aΩsin2 φ/cos φ (134 m s-1 at 30o latitude and 700 m s-1 at 60o latitude) that is many times larger than that observed as seen in Figure 1. It is clear from Figure 1 that the angular momentum is far from uniform and the motion crosses angular momentum contours. Consistent with this, the actual subtropical jet maximum of about 40 m s-1 at 30oS is very much smaller than the value suggested by the theory based on angular momentum conservation in the upper branch of the Hadley Cell. This implies that eddy angular momentum mixing processes are actually of order one importance. In this study, we will reveal some interesting features associated with the Hadley Cell.

Figure 1:  The JJA zonal mean overturning circulation in a latitude-height cross section, based on 30 years of ERA-Interim data.  Colours indicate absolute angular momentum and blue contours indicate zonal winds.

Firstly, to investigate the nature of the JJA momentum flux, Figure 2 shows the steady and transient momentum flux in JJA 2009. It is seen that in addition to the expected strong transient momentum flux in the midlatitude, whose importance for the Hadley Cell has often been stressed, both steady and transient fluxes show a maximum in the tropical upper troposphere, extending from the region of tropical convection in the summer hemisphere into the sub-tropics of the winter hemisphere. The dividing line between tropical positive values and sub-tropical negative values looks almost identical in the steady and the transient, suggestive that there is a motion with SW-NE tilts north of about 15oS and NW-SE south of that with an amplitude that fluctuates in time. The latitude-longitude pictures at 150 hPa show that this signature comes predominantly from the Indian Ocean (not shown). This can be seen in the case studies below (Figure 5).

The Boreal winter picture indicates a similar tropical upper-tropospheric flux maximum but with the sign reversed as expected (not shown).

This indicates that angular momentum flux in the tropical upper troposphere, which has been neglected, is very important for the existence of the Hadley Cell.

Figure 2: Northward momentum flux of westerly wind in JJA 2009.

Then to examine the zonal and temporal variation of winds and convection in the Hadley Cell region, Figure 3 shows the JJA mean motion at (a) 200 hPa and (b) 950 hPa and Outgoing Longwave Radiation (OLR).  It is clear that the zonal average motion described by the Hadley Cell occurs in longitudinally confined regions that can be associated with the tropical convective heating regions (Low OLR). The lower tropospheric meridional motion contains the flow from the S Indian Ocean into the S Asian and the W Pacific regions of convection. It also shows the flow from the S Hemisphere into the E Pacific and Atlantic Inter Tropical Convergence Zone (ITCZ) heating regions. In the upper troposphere there is a return flow in each of these regions.  

Figure 3: Climatology JJA motion.

To illustrate the transient behaviour, Hovmöller of 5°N-10°S v at 200hPa (used to show the upper-tropospheric cross-equatorial motion) and  0-20°N OLR for 2009 JJA season are presented in Figure 4. The motion is seen to be localised in longitude and time.  The cross equatorial motion in the upper troposphere is strongest in transient waves associated with convective events over the Indian and W Pacific region.

 Figure 4: Transient motion in JJA 2009.

Finally, individual synoptic events in different longitudinal sectors are analysed. As a case study, Figure 5 (a),( b) show the upper and lower tropospheric horizontal winds for early July 2009 with contours of Potential Vorticity (PV) and OLR, respectively. On 6 July when convection is predominately in Indian sector, the lower tropospheric inflow is seen to have its origin from 30-40oS near 60oE. Two days later (8 July), in the upper troposphere, return flow reaches 30oS where it interacts with the S Hemisphere winter subtropical jet and the eastward moving synoptic waves on it, with a horizontal tilt consistent with that suggested by the momentum flux shown in Figure 2.  A filament of N Hemisphere positive PV moves towards the anti-cyclonic side of the S Hemisphere jet.

On 11 July, when convection is predominately in the Philippine sector, similar features are seen in a region shifted to the east.

Figure 5: (a) 370K PV and 150-hPa winds and (b) OLR and 950-hPa winds in early July 2009, 5 days apart with OLR and lower level winds leading the upper level features by two days.

In summary, this study gives evidence that:

(1) The existence of the Hadley cell involves not only the expected strong transient momentum flux in the midlatitude, but also the strong momentum flux in the tropical upper troposphere.

(2) The upper branch of the Hadley Cell is concentrated in certain longitudinal sectors and intensified cross-equatorial flow is associated with flaring in organised convection in those regions.  The tropical upper tropospheric motions associated with convection are crucial to the existence of the Hadley Cell.

(3) Filaments of the upper tropospheric air move from the summer to the winter hemispheres or are mixed in; they can reach the anti-cyclonic side of the winter subtropical jet and interact with the weather systems on it.

These observed features/processes have important implications on weather, climate and climate change, therefore it is important to know how well they are represented in weather and climate models. Also knowing that the cross-equatorial transfer of trace chemicals in the atmosphere occurs in filaments may have significant implications for atmospheric chemistry models, with almost undiluted summer hemisphere air moving deep into the winter hemisphere.

Posted in Climate, Climate change, Climate modelling, earth observation, Equatorial waves, extratropical cyclones, Tropical convection, Troposphere, Waves, Weather, Wind | Leave a comment

Mechanisms of Climate Change in the Indian Summer Monsoon

By Jon Shonk

Over one billion people are reliant on the rainfall of the Indian Summer Monsoon. During the wet season, which usually spans June to September, some parts of India receive over 90% of their total annual rainfall. Deficits or excesses of rainfall can have devastating effects, such as drought, inundation, crop failure and health issues. Bouts of extreme weather, such as short periods of very intense rain, can also have detrimental effects via flash flooding and the triggering of landslides.

It is therefore important to get an idea of how monsoon rainfall might change in a warmer future climate. Climate prediction uses numerical models to advance an initial global “snapshot” of the atmosphere and ocean forward in time using a supercomputer, and then examines the statistics of the weather over some period in the future. Nowadays, many institutions around the world run their own climate models. While these are all based around the same physical principles, the formulation and structure of the models can be very different. This means that the behaviour of models, even if initialised from the same global snapshot, can be quite different after 100 years of simulation.

The five maps in Figure 1 show the projected change in rainfall (averaged from June to August) over India in a future world that is 1.5 °C warmer than pre-industrial conditions (about 0.8 °C warmer than today), for five different climate models. There is a clear disagreement in the pattern of change, with no obvious consensus on which parts of India are likely to become wetter or drier.

Figure 1. Projected changes in rainfall as a result of a 1.5 °C warming, according to five climate models. Rainfall is averaged over June, July and August. Data from the HAPPI project (see Mitchell et al, 2017 for details).

So can we infer anything about the future Indian Summer Monsoon from these models? An advantage of using multiple models is that we can build an “ensemble” of predictions – that is, a number of plausible future climate projections. But the challenge is then how to statistically combine the projected changes to produce a single, clear, robust message.

The simplest option is to take an average across the projections from the five models (Figure 2a). The result is a weak pattern of slightly wetter conditions over eastern India and Bangladesh. However, the averaging process leads to areas where an increase of rainfall in one model cancels out a decrease in another, and understanding the reasons why models project such differences could provide extra clues as to how the monsoon might change.

Figure 2. Projected changes in rainfall, shown as the average across the same five models used in Figure 1. The changes for a warming of (a) 1.5 °C and (b) 2.0 °C are shown. Data from the HAPPI project.

By examining the behaviour of the models individually we can build an idea of the mechanisms by which the rainfall distribution changes in a warmer climate. This has been the focus of my recent work. I have also been looking at the differences in rainfall change between a world that is 1.5 °C warmer and one that is 2 °C warmer (Figure 2b). A paper on this should be ready soon…


Posted in Climate, Climate change, Climate modelling, drought, Environmental hazards, Flooding, Monsoons, Numerical modelling, Oceans | Leave a comment