Impacts of climate variability and change on the energy sector: A case study for winter 2009/10

By Emma Suckling

Secure and reliable energy supplies are an essential part of modern economic life. But the national and global infrastructures that deliver energy are changing rapidly in the face of new and unprecedented challenges, including the need to meet ever-increasing global demand for energy services, whilst reducing CO2 emissions caused by burning fossil fuels. Responding to these challenges will likely involve the development of new technologies, as well as increased deployment of weather-dependent renewables, such as wind and solar power, in the energy mix. This new energy landscape exposes stakeholders in the energy sector to a greater risk from weather and climate than ever before. Better understanding the impacts of climate variability on energy supply and demand therefore has the potential to aid policy and decision makers in evaluating risks.

The European Climatic Energy Mixes (ECEM) project is a Copernicus Climate Change Service (C3S), whose aim is to enable the energy industry and policy makers to assess the impact of climate variability and change on energy supply and demand over Europe. A proof-of-concept service – or Demonstrator is being developed, including datasets that bring together climate and energy data, produced in a consistent way covering a range of time scales and countries in Europe. The ability of the tool to provide insight into events, anticipate future risks and ask ‘what if’ questions is illustrated in the context of the unusually cold winter of 2009/10.

Record power demand in winter 2009/10
Many countries across Europe experienced unusually high levels of gas and electricity demand due to cold weather conditions during winter 2009/10 (December 2009 to February 2010). In the UK and France high levels of day-to-day weather-sensitive electricity demand is seen in the ECEM dataset (Figure 1), with demand exceeding 10-20% above normal levels for several days over the winter.

Figure 1: Daily fluctuations in modelled weather-sensitive electricity demand. Normalised anomalies expressed as difference in percentage from the long-term average (1979-2016).

A cold winter
Winter 2009/10 made headlines for being unusually cold across much of northern Europe and saw some of the lowest temperatures in the last 40 years in the UK. There was a strong contrast between conditions in northern and southern Europe (as seen in Figure 2 for winter mean temperature differences in 2009/10), consistent with a southward-displaced jet stream and a prolonged negative phase of the North Atlantic Oscillation [Reference 1].

Figure 2: Temperature anomaly (degC – differences from 1981-2010 mean) across Europe for winter 2009/10.

What if winter 2010 happened today?
The ECEM historical dataset also provides estimates of renewables energy supplies based on today’s energy mix and the historical climate drivers. This allows us to investigate the potential impacts of past climatic events if they happened today. A winter like 2009/10, which saw persistent cold and still conditions in the UK, would have a larger impact on the energy sector today due to the increase of renewables into the energy mix. For example, in 2010 renewables consumption was around 3% in the UK, rising to around 8% in 2015 (with wind power accounting for 4%) [Ref 2]. The low wind conditions in a repeat of winter 2009/10 would lead to a substantial reduction in wind power production over the season (Figure 3), which could lead to increased risks to electricity supply availability when combined with an increased demand due to low temperatures.

Figure 3: Estimated winter mean wind power production based on the historical wind speeds and today’s wind power generation capacity for the UK.

Anticipating cold, still winters and their impacts in future
Whilst winter 2009/10 was unusually cold compared to recent winters (i.e. the last 40 years), it was warmer than winter 1962/63, despite exhibiting very similar atmospheric conditions (a prolonged negative phase of the North Atlantic Oscillation, NAO-). It has been suggested that winter 2009/10 might have been even colder if the overall global warming trend observed in the 20th Century had not occurred [Ref 3]. Climate projections of winter temperatures over Europe generally show a warming trend out to the end of the century, suggesting that cold winters, such as 2009/10, may become less likely in future (Figure 4). This has implications for winter demand, which has a negative relationship with temperature over most of northern Europe. Projections of wind speed typically show no clear indications of any trend, however, with the increase of installed wind power generation over Europe it is likely that any future power system will be more sensitive to weather-dependent renewables generation than to temperature-driven demand. Gaining a better understanding of the impacts of climate variability and change on the energy sector is therefore an essential area of research [Refs 4-7].

Figure 4: Projections of winter mean temperature from the RCP4.5 climate scenario over the UK. The shaded region shows the smoothed upper and lower bounds from an ensemble of models, the red lines indicate the 1981-2010 winter mean temperature (top) and the 2009/10 winter temperature (bottom). The green line illustrates the variability from one model run from the full ensemble.

[1] G. Ouzeau, et al., 2011. European cold winter 2009-2010: How unusual in the instrumental record and how reproducible in the ARPEGE-Climate model? Geophysical Research Letters, 38, 11.

[2] The UK’s Energy Supply: security or independence? 26 May 2011

[3] J. Cattiaux, et al., 2010. Winter 2010 in Europe: A cold extreme in a warming climate.  Geophysical Research Letters, 37, L20704

[4] D. Brayshaw, et al., 2012. Wind generation’s contribution to supporting peak electricity demand: meteorological insights. Journal of Risk and Reliability, 266, 44-50

[5] D. Cannon, et al., 2015. Using reanalysis data to quantify extreme wind power generation statistics: A 33 year case study in Great Britain. Renewable Energy, 75, 767-778

[6] D. Drew, et al., 2015. The impact of future offshore wind farms on wind power generation in Great Britain. Resources, 4, 1, 155-171

[7] H. Bloomfield, et al., 2016. Quantifying the increasing sensitivity of power systems to climate variability. Environmental Research Letters, 11, 12

Posted in Climate, Climate change, Climate modelling, Renewable energy, Seasonal forecasting | Leave a comment

Exploring the impact of the Atlantic Multidecadal Variability (AMV)

By Dan Hodson

After 140 years of observations, we now know that the temperature of the surface of the Atlantic ocean slowly varied over time, cooling and warming over periods of decades (Figure 1). These slow variations in temperature sit atop the background global warming trend (A), the contrast with other regions of the globe can clearly be seen in spatial maps of SST difference (B). The term Atlantic Multidecadal Oscillation (AMO) was initially coined to describe these variations around the global mean trend, but recently the more general term Atlantic Multidecadal Variability (AMV) has been adopted by the community.

Figure 1. A) Black: Atlantic multidecadal Variability (AMV) (mean over black box). Green: mean over region outside black box
B) annual mean Sea Surface Temperatures: (1965–75) minus (1951–61).

The origin and mechanisms by which the AMV arises are still a matter of debate. It is ultimately impossible to deduce the origins from using observations alone (although we can hazard some educated guesses), so we have to turn to model studies. Some argue that the AMV arises due to internal ocean variability – involving variations in the heat transported by ocean dynamical processes, such as the Atlantic Meridional Overturning Circulation (perhaps responding to stochastic forcing from the atmosphere). Many coupled climate models do display AMV that arises due to this. 1 2. Others argue that models show that the historical AMV arose due to changes in external forcings, or question the role of ocean dynamics altogether.

Jon Robson has recently written about ongoing efforts to predict the evolution of the AMV by using ocean observations to initialize ocean models. These studies suggest an ocean-origin for the AMV is more likely. Whatever the origin of the AMV, and independent of our ability to predict it, we can still ask – what are the climatic impacts of the AMV? Again, we have to turn to models to start to answer this question. Multiple attempts have been made over the the past two decades to examine the possible impacts of the AMV on climate. Ten years ago we examined the idealized impact of a fix AMV pattern on climate in an Atmosphere-only model 1 2. We discovered significant, and potentially important, impacts on surface temperatures, rainfall and atmospheric circulation (Figure 2) – notably, these were consistent with the observational record in a number of regions.

Figure 2. (A to C) Observed JJA (Warm-Cold AMV periods). (A) Sea-level pressure.(B) Land precipitation (mm/day). (C) Land surface air temperature (°C). (F to H) As in (A) and (B) but Model response to AMV (warm – cold). D and E are AMV Warm – Cold composites from a model run with historical SSTs.

Motivated by this, during the DYNAMITE project, we repeated these experiments in a range of other atmosphere-only models, we discovered a range of similar responses, but a number of key uncertainties – e.g. the magnitude of the impact on rainfall.

Experiments such as these are the first step in elucidating the climatic impact of the AMV. However, since these experiments used atmosphere-only models with fixed sea surface temperature (SST), it wasn’t possible to investigate dynamical feedbacks – for example, how the atmospheric response to the AMV in turn affects the ocean, such feedbacks may ultimately modify the final atmospheric response. Modelling studies to date suggest that such feedbacks could be significant.

In order to address this, a new international multi-model experiment is underway to resolve these questions. It will run as part of CMIP6:DCPP – the Decadal Climate Prediction Project component of the fifth Coupled Model Intercomparision Project (there was no CMIP4) – the Decadal Climate Prediction Project . The experiments within DCPP will examine the impact of the AMV in coupled climate models. Each of the models in the experiment ensemble will allow SSTs in the models to evolve with the underlying ocean model, but will periodically nudge those in the Atlantic towards a warm AMV pattern. The idea behind this is to drive the models with a warm AMV, but without restricting the ocean coupling or responses. Reading are talking part in this international effort by using the MetUM-GOML2 coupled mixed layer ocean model developed by Nick Klingaman and Linda Hirons here in Reading. First results are just beginning to arrive, and it looks like we may have some interesting differences from the old AGCM results – most notably, the AMV appears to have a significant impact on the Pacific ocean across the globe. If these results are born in other models, it may point to a greater role of the Atlantic in modulating global climate than has hitherto been expected. Watch this space!


An anatomy of the cooling of the North Atlantic Ocean in the 1960s and 1970s Daniel L. R. Hodson, Jon I. Robson, Rowan T. Sutton, 2014: Journal of Climate, 27 (21), 8229-8243

Atlantic Ocean forcing of North American and European summer climate R. T. Sutton, D. L. R. Hodson, 2005: Science, 309 (5731), 115-118

Climate response to basin-scale warming and cooling of the North Atlantic Ocean R. Sutton, D. Hodson, 2007: Journal of Climate, 20 (5), 891-907 e-print

Climate impacts of recent multidecadal changes in Atlantic Ocean Sea Surface Temperature: A multimodel comparison Daniel Louis Richard Hodson, Rowan Timothy Sutton, C. Cassou, N. Keenlyside, Y. Okumura, T. Zhou, 2010: Climate Dynamics, 34 (7-8), 1041-1058

Aerosols implicated as a prime driver of twentieth-century North Atlantic climate variability Ben B. B. Booth, Nick J. Dunstone, Paul R. Halloran, Timothy Andrews & Nicolas Bellouin. Nature 484, 228-232 (12 April 2012)

The Atlantic Multidecadal Oscillation without a role for ocean circulation Amy Clement, Katinka Bellomo, Lisa N. Murphy, Mark A. Cane, Thorsten Mauritsen, Gaby Rodel, Bjorn Stevens Science 16 Oct 2015: Vol. 350, Issue 6258, pp. 320-324 DOI: 10.1126/science.aab3980

Decadal prediction of the North Atlantic subpolar gyre in the HiGEM high-resolution climate model Robson, J., Polo, I., Hodson, D.L.R. et al., 2017: Climate Dynamics

A Mechanism of Internal Decadal Atlantic Ocean Variability in a High-Resolution Coupled Climate Model Matthew B. Menary, Daniel L. R. Hodson, Jon I. Robson, and Rowan T. Sutton, Richard A. Wood, 2015: Journal of Climate 28:19, 7764-7785 

Posted in Climate, Climate change, Climate modelling, Numerical modelling, Oceans | Tagged | Leave a comment

Domestic implications of climate science

By Jonathan Gregory

I’m a climate scientist. I’ve been working in climate change research since 1990. During those years scientific information has become ever more detailed and convincing regarding the magnitude of climate change in both the past and the future due to human activities, principally the emission of carbon dioxide from fossil fuel combustion. I was an author of the most recent assessment (published in 2013) of the Intergovernmental Panel on Climate Change, which concluded, “Continued emissions of greenhouse gases will cause further warming and changes in all components of the climate system. Limiting climate change will require substantial and sustained reductions of greenhouse gas emissions.”

Although news reports about the IPCC may contain remarks such as, “In their latest international report, climate scientists warn that the world must cut greenhouse gas emissions to avoid disaster,” actually the role of the IPCC is only to provide policy-relevant information based on its assessment of published scientific work; it does not propose or advocate any policy. That is the job of policy-makers. One of the responses of the UK government is the Climate Change Act, which established a target for the UK to reduce its emissions by at least 80% from 1990 levels by 2050. This target is an appropriate UK contribution to global emission reductions consistent with limiting global temperature rise to as little as possible above 2°C.

UK carbon dioxide emissions come from many activities which use energy. Houses consume about 30% of the total. This arises mainly from burning gas in our boilers, for central heating and hot water, and partly from electricity use. The UK housing stock is old relative to most European countries, with many houses dating from Victorian times.

I’m a home-owner as well as a climate scientist, and this conjunction leads me to the conclusion that I ought to reduce my domestic carbon dioxide emissions. My house was built in 1873. It’s semi-detached, and has solid brick walls with no cavity. Over the last several years, I have been improving its energy efficiency. The energy “import” (from the gas and electricity mains) has fallen from about 30,000 kWh per year when I first moved in, to about 11,000 kWh per year recently (Figure 1).

Figure 1. Energy import per annum

This has been achieved through a variety of alterations, including insulation of roofs and walls, double and triple glazing, a condensing boiler, a wood-burning stove, and more energy-efficient electrical appliances. I am impressed by the performance of my new A+++ fridge/freezer, which has a greater capacity than the two old ones I used to run put together, but uses only 13% of the energy. That’s significant, because the fridges were the largest consumer of electricity in the house. I also have solar panels installed on the roof (Figure 2). These are of two kinds. The photovoltaic (PV) panels take up more space, and generate more than a half of our annual electricity consumption. The solar thermal panel heats all the hot water we need for showers and washing-up during the summer months. The solar thermal panel is better at collecting usable energy, because the efficiency of conversion of sunlight into electricity by PV panels is quite low.

Figure 2. Roof-mounted solar panels

The most dramatic and unusual undertaking to date has been to build a new wall on the outside of the gable-end wall of the house (Figure 3). The purpose of this was to create an insulated cavity. I decided to have it magnificently well-insulated, since the insulation itself is cheap compared with the cost of the work. I’m very pleased with it. Most people don’t notice it in daylight, because the bricks are a good match in colour, but you can see its effect in a thermal image (Figure 4) comparing my house with my neighbours’ on a cold day last winter. The side-wall of my house (on the left) is much colder than theirs, because less heat is leaking through it. The new wall has reduced the gas consumption by more than a third (it’s the downward step after 2010 in the graph).

Figure 3. Adding a new cavity wall to the existing house

Figure 4. Infrared image of Jonathan’s house side wall and his neighbour’s house side wall, showing the much lower heat loss (lower surface temperature) from the insulated wall (left)

The thermal conductivity of the insulating material (solid foam polyisocyanurate, or PIR) is 0.023 W m-1 °C-1. Hence a layer of thickness 190 mm has a u-value of 0.023/0.19=0.12 W m-2 °C-1, so for instance if the difference between the temperature inside and outside the house is 10 degC, the heat flux through the insulator is 1.2 W m-2. I don’t exactly know the thermal conductivity for the old bricks, but it’s probably over ten times greater than for PIR. It’s generally assumed that a traditional solid brick wall, two bricks or nine inches thick, has u of about 2 W m-2 °C-1. Thus, with the addition of the insulator, the side-wall of the house conducts between ten and twenty times less heat than before.

SuperHomes are old houses which have been refurbished by their present owners to reduce their carbon dioxide emissions by at least 60%. The SuperHomes scheme has a register of over 200 such properties. The aim of the scheme is to provide information about refurbishment for energy efficiency, through holding open days in SuperHomes each September. In August 2012, my house qualified as a SuperHome, but there’s still plenty more to be done!

Posted in Climate, Greenhouse gases | Tagged | Leave a comment

Heat waves of the past decade in Chinese mega-cities: a quick review

By Ting Sun

Although cities are often already warmer than their rural surroundings (the well-known “urban heat island” effect), heat waves (HWs), excessively hot periods, will not only enhance the urban and rural temperatures but also exacerbate the contrast between them (Li et al 2015), leading to aggravated thermal stress on urban dwellers (Sun et al 2016). I spent this summer in Beijing and experienced several sultry days in July, which, again, provided me with more sensible HWs than those HW papers I’ve been working on. As September came, Beijing began to cool down; this reminds me of the very cool summer in Reading last year (at least to me). So it seems a good time for a quick review of HWs in Chinese mega-cities of the past decade (2007–2017).

The megacities examined here are those with population larger than 5 million (MoHURD of China, 2016); there are 20 of these, with Shanghai, Beijing and Chongqing the top three populous mega-cities. Temperature observations from the Integrated Surface Database (Smith et al 2011) are employed in this work; however, due to the insufficient data continuity (i.e., consistent records shorter than 30 yrs), two cities (Hong Kong and Shenzhen) are excluded. As such, the following review is only for 18 Chinese mega-cities.  

First, let’s look back at the highest temperatures experienced by these cities during the past summer (Figure 1). Unsurprisingly, records over 35 °C (high temperature according to the Chinese Meteorological Administration, CMA) were recorded at all sites (except for Nanjing, very suspicious records here!) even with highs over 40 °C for several of them.


Figure 1. Daily maximum temperature of 18 Chinese mega-cities recorded during the 2017 summer months (JJA).

Then we move on to the HWs. Although CMA define a period with three or more consecutive days with daily maximum temperature (Tmax) over 35 °C as a HW event, given China is vast country with diverse climates, a location-specific approach (Meehl and Tebaldi 2004) is adopted here for HW identification. With T1 the 97.5th percentile of the observed  series and T2 the 81st percentile, a HW is defined as the longest period that satisfies the following conditions: (1) Tmax > T1 for at least 3 days; (2)   for the entire period; and (3) Tmax > T2 for the entire period, where Tmaxbar denotes the average of Tmax over the HW period.

A total of 973 HWs occurred during the past decade in the 18 mega-cities (Figure 2); in  2009, 2010 and 2017 all the 18 cities experienced HWs. Also, Hefei, Kunming, Changchun, Harbin, Hangzhou and Wuhan experienced HWs all the way through the past ten years, followed by Xi’an, Chengdu, Nanjing, Beijing and Shanghai that recorded HWs in nine of the ten years.

Figure 2. Occurrence of HWs (denoted by empty dots) of 18 Chinese mega-cities between 2007 and 2017.

By looking into the annual HW characteristics, we find more interesting facts. Regarding the annual frequency (i.e. number of HWs per year, Figure 3), 2009, 2010 and the very recent 2017 generally observed more HW events compared with other years. And, solely based on the trend, 2017, similar as 2009, looks to be on an upslope to more HWs.

Figure 3. Annual HW frequency of 18 Chinese mega-cities between 2007 and 2017.

The facts revealed by annual HW durations (i.e., number of HW days per year, Figure 4) are more striking. In addition to 2010 and 2017, 2013 emerges as another “significant” year; this is particularly true to me: I was in Zhejiang that year for the whole summer and underwent highs of 40 °C almost every day! By comparison with the HW frequency, it is clearly shown that HWs of 2013 were even stronger than those of 2010: though with fewer events, HWs persisted longer in 2013. Furthermore, 2017 outperformed other years with the most HW days.

Figure 4. Annual HW duration of 18 Chinese mega-cities between 2007 and 2017.

Following the successive warmest 2015 and 2016 since modern record keeping began in 1880, will 2017 hit a new record? Although an answer to it is not clear yet, this quick review (a far from a thorough investigation) highlights 2017 for the Chinese mega-cities as a remarkable year with the most annual HW days in the past decade. And, if such trend continues, it looks we will “welcome” more HW days in the coming years.


MoHURD of China, 2016 Facts of urban development in China: accessed 8 September 2017

Li D, Sun T, Liu M, Yang L, Wang L and Gao Z, 2015. Contrasting responses of urban and rural surface energy budgets to heat waves explain synergies between urban heat islands and heat waves. Environ Res Lett., 10, 054009

Meehl G A and Tebaldi C, 2004. More Intense, More Frequent, and Longer Lasting Heat Waves in the 21st Century. Science, 305, 994–7

Smith A, Lott N and Vose R, 2011. The integrated surface database: Recent developments and partnerships. Bulletin of the American Meteorological Society, 92, 704–8

Sun T, Grimmond C S B and Ni G-H, 2016. How do green roofs mitigate urban thermal stress under heat waves? J Geophys Res-Atmos, 121, 5320–35



Posted in China, Climate, Environmental hazards, Weather | Leave a comment

Abrupt climate change at the Royal Meteorological Society

By Joy Singarayer

Recently I was excited to be invited to attend and talk at a Royal Meteorological Society meeting in London on abrupt climate change since the last ice age. The scope of the meeting was to highlight mechanisms driving rapid environmental change during the last 21,000 years, the techniques used to quantify these (there are no direct observations from thousands of years ago, of course), and to discuss the implications for how the climate may evolve in the future. With the palaeo focus being perhaps a bit left field for the Society, we were pleased to find that the meeting was fully booked well in advance, and that the audience, like the speakers, had a good balance of both genders and early-career/established scientists.

At the last glacial maximum, around 21,000 years ago, global average temperature was lower by around 5 degrees Celsius, and sea levels lower by some 120 m, due to the expansion of continental ice-sheets. The process of deglaciation to reach the warm conditions similar to today took roughly ten thousand years, which seems orders of magnitude slower than future anthropogenic climate change projections. However, this process did not occur steadily and monotonically, but was punctuated by several rapid climate change episodes (see Figure 1), just as the last glacial period was.

Figure 1. Screenshot from Liz Thomas’ talk. NGRIP ice core oxygen isotope record for the last 40,000 years. Rapid warming of the Bolling-Alleröd is highlighted and the slower Younger Dryas cooling. MWP-1A and MWP-1B are ice-sheet Melt Water Pulse inputs into the global ocean.

How rapid were some of those changes? Liz Thomas from the British Antarctic Survey reviewed records obtained from ice cores. In some cores (e.g. Greenland North GRIP core) it has been possible to count annual layers within the ice, providing extremely high resolution and accurately dated information. The ratio of different oxygen isotopes (δ18O) in the ice water tells us about the temperature over the ice core region and shows that during the Bolling-Allerod warm transition around 14.7 kyr ago temperatures over Greenland increased by around 10 degrees Celsius in only 1-3 years. Dust contained in the layers originates primarily from low latitude deserts and suggests strongly that changes to low latitude atmospheric circulation and the hydrological cycle preceded the high latitude temperature rise. 

These rapid warmings were likely triggered by strengthening of the Atlantic Ocean overturning circulation. Further talks at the meeting examined these mechanisms through modelling (Lauran Gregoire) and ocean palaeoarchives (Andrea Burke), as well as exploring the impacts on early human societies (William Davies), and then to considering changes during the Holocene interglacial and the impacts of early civilisations, making a bridge to discussion of future anthropogenic change. In the final discussion (Figure 2), led by Paul Valdes, it was noted that while there is no true past analogue to future projected changes, and the mechanisms are different, some of the palaeoclimate changes were of similar magnitudes and speeds to future projections. In this sense they provide useful case studies for model evaluation and for understanding the potential responses of ecosystems to rapid change. They also suggest the importance of Earth System interactions (biosphere, cryosphere, atmospheric chemistry) in producing rapid changes – interactions that are now being more fully incorporated into those climate models being used to simulate future projections of climate change.

Figure 2. The panel discussion at the conclusion of the recent Royal Meteorological Society’s meeting on rapid climate change.

The meeting was organised by Ruza Ivanovic (University of Leeds) and chaired by Prof. Dame Jane Francis (British Antarctic Survey). The presentations (including recordings) and further details about the meeting may be found at the Royal Meteorological Society’s events page.



Posted in Climate, Climate change, Climate modelling, Conferences, Royal Meteorological Society | Leave a comment

The Undergraduate Research Opportunities Programme (UROP)

By Charlie Williams

The Undergraduate Research Opportunities Programme (UROP) is a scheme run by Careers at the University of Reading, enabling undergraduate students in the middle of their degree to work alongside an academic and gain hands-on research experience. They typically work for 6 weeks during the summer, although part-time options are available, and are paid a reasonable salary financed by the University. They are, essentially, a paid intern, working alongside a supervisor and assisting them on a given research project; the supervisor is responsible for writing the proposal, which is then assessed by Careers and funded if deemed appropriate.

Despite the scheme running since 2006, this was the first year I applied to UROP. My proposal was accepted, intended to look at the West African Monsoon (WAM) and how its behaviour has changed in the geological past, present and future. The primary research question was, in its simplest form: Can we use the past to shed any light on the future? This was designed to complement existing research, as part of the much larger PACMEDY project involving several members of the Department and many other institutions.

Research into WAM variability is of great importance, as West African societies are heavily reliant on rain-fed agriculture. Failure, or even just a weakening, of the monsoon can have devastating impacts such as drought, crop failure, famine and other resulting socio-economic issues. Under most scenarios of future climate change, it is believed that the WAM may provide more extreme heavy rainfall events, whilst at the same time becoming more erratic (Biasutti et al. 2008; Diallo et al. 2016). However, there is high uncertainty with many studies disagreeing as to even the sign of change of the WAM, let alone its magnitude or intensity, so one way of assessing our climate models’ projections is to look into the geological past, when analogous climate conditions existed. One example of this is the Mid-Holocene (MH), roughly 6000 years ago, when the WAM was significantly more intense and spatially more extensive, due to a different orbital configuration (Gaetani et al. 2017). This can be seen in Figure 1, where a stronger monsoon extending across West Africa is shown , with a reduction in mean rainfall along the Guinean coast.

Figure 1: Annual mean JJA rainfall differences between the MH and the pre-industrial era, as simulated by the UK Met Office Hadley Centre’s climate model, HadGEM2-ES

The project, therefore, was to compare the behaviour of the WAM over three separate climate states, something currently lacking within the scientific community. Over the summer, my intern and I worked together, focusing in particular on a reasonably well understood physical process within the WAM, namely the West African dipole; this is one of the main patterns of rainfall variability across West Africa, and is characterised by positive (negative) rainfall anomalies over the Sahel and negative (positive) rainfall anomalies along the Guinean coast, associated with negative (positive) SST anomalies in the Gulf of Guinea and equatorial eastern Atlantic (Lough 1986; Janowiak 1988; Cook & Vizy 2006). Several interesting results were found, and although the questions were not entirely resolved within the duration of the UROP placement, sufficient progress was made such that the findings are now being written up for publication.

There are numerous benefits to UROP, both for the student and for the supervisor. For the student, they will hopefully gain many transferable skills, interact with the research community and learn what it is like to be an academic. If the student is interested in a career in research, or is just contemplating postgraduate studies, it provides them with a taster of what full-time research can be like. For the supervisor, it allows an extra ‘pair of hands’ to join your team, for example by conducting research to aid and supplement your existing research or by allowing a pilot project to run. Personally, I found it a very rewarding experience, and very much hope to get a decent paper out of the project – with my intern as co-author, of course!


The primary acknowledgement, naturally, goes to my UROP student, Hana E. Beckwith. Secondly, of course, very many thanks to UROP and Careers for funding this project – details of the scheme at


Biasutti, M., Held, I. M., Sobel, A. H. & Gianni, A., 2008. SST Forcings and Sahel Rainfall Variability in Simulations of the Twentieth and Twenty-First Centuries. J Clim. 21 (14): 3471-3486

Cook, K. H. & Vizy, E. K., 2006. Coupled Model Simulations of the West African Monsoon System: Twentieth- and Twenty-First-Century Simulations. J Clim. 19: 3681-3703

Diallo, I. et al., 2016. Projected changes of summer monsoon extremes and hydroclimatic regimes over West Africa for the twenty‐first century. Clim Dyn. 47: 3931-3954

Gaetani, M. et al., 2017. West African monsoon dynamics and precipitation: the competition between global SST warming and CO2 increase in CMIP5 idealized simulations. Clim Dyn. 48: 1353-1373

Janowiak, J. E., 1988. An investigation of interannual rainfall variability in Africa. J Clim. 1: 240-255

Lough, M. J., 1986. Tropical Atlantic sea surface temperatures and rainfall variations in sub-Saharan Africa. Mon Wea Rev. 114: 561-570


Posted in Africa, Climate, Climate change, Climate modelling, Monsoons, University of Reading | Leave a comment

Simulating the effect of electrical charge on cloud drops using Direct Numerical Simulation

By Torsten Auerswald

In the atmosphere, clouds develop when water vapour condenses leading to the formation of cloud drops. This process is usually supported by the presence of condensation nuclei which allow drop formation at low supersaturations. Aerosol particles in the air can act as such condensation nuclei. Once the formation of cloud drops is initiated, several physical processes lead to a growth of the drops. Small drops mainly grow by condensation of water vapour from the surrounding atmosphere onto the drop, while for larger drops collision and subsequent coalescence of drops are the main mechanism for growth.

The collision process of these larger drops is influenced by the size of the drops, as larger drops fall faster than smaller ones. Therefore, larger drops can collect smaller drops due to different terminal velocities. Another important driver for drop collision is turbulence which can lead to an increase in collisions due to the inertia of the drops.

However, not every drop collision results in coalescence. Two colliding drops can bounce off from each other or break up into smaller drops. It can be shown that successful coalescence is more likely when the colliding drops are electrically charged. Charging of drops can occur naturally due to background radioactivity and cosmic rays, which form ions in the atmosphere and cause aerosol particles to carry a slight charge. Thus, by charged aerosol particles acting as condensation nuclei, charged cloud drops are formed. Because of the polarisation of one drop from the charge carried by another drop, even drops with like charges will experience attractive forces, given that their separation distance is sufficiently small. This effect increases the efficiency of the collision and coalescence processes in the cloud and therefore can act to accelerate the drop growth and ultimately the production of raindrops.

In our project we are developing a model to simulate the behaviour of cloud drops in a warm cloud. For a small volume of such a cloud the turbulent flow is simulated using Direct Numerical Simulation (DNS) and the collision of drops in the turbulent flow is studied. By introducing charged drops into the simulation and considering the electrical forces resulting from the charge, we will be able to investigate the influence of electrical charge on the collision rate and size distribution of cloud drops and the production of rain drops.

To illustrate this type of simulation, the figure below shows a snapshot from a simulation where cloud drops are moving in a flow with periodic boundary conditions and growing by collision.To see the animation click on this link.

Posted in Aerosols, Climate | Tagged | Leave a comment

Forecasting the Indian monsoon

By Arathy Menon

The South Asian monsoon, which brings rainfall to India and the neighbouring countries during the boreal summer season, is a major atmospheric circulation system. India receives more than 80% of its annual rainfall during the monsoon season, generally occurring from June to September. However, the monsoon onset can also occur as early as May and the withdrawal can occur as late as early October, while the season itself is comprised of several wet and dry periods which occur as a part of the intraseasonal variability. Any variability in the timing, duration and intensity of the monsoon can have a significant impact on rain-fed agriculture that contributes a major portion of India’s GDP.  Irrespective of recent growth in service and industrial sectors, agriculture is still the predominant occupation in many rural regions of India. The Indian monsoon also has significant impact on the coal and steel industries, thereby affecting the world economy. Hence understanding and predicting the monsoon is vital.

The forecasting centres around the world try to predict the monsoon onset and amount at least a season in advance. However, it is notoriously difficult to predict the timing of the onset as well as the overall seasonal rainfall. The models participating in the latest version of the Coupled Model Intercomparison Project showed considerable improvement in simulating the mean rainfall and variability of the South Asian monsoon compared to earlier models; however, a dry bias still exists over India in most of these models (Sperber et al., 2013). The ability to predict the monsoon at least a season in advance is limited as it is difficult to represent tropical convection (the processes leading to monsoon rainfall) in our forecast models, and we lack a proper understanding of the way in which land and ocean surfaces alter the atmosphere on small scales to initiate monsoon storms. Hence detailed observations of the lower layers of the atmosphere and surface are needed to understand such processes.

In order to achieve this, an intensive field campaign was conducted in India during summer 2016 as part of the INCOMPASS project (refer to Dr Andrew Turner’s blog for more detail). During summer 2016, we took the UK’s Atmospheric Research Aircraft (operated by the Facility for Airborne Atmospheric Measurement), a modified BAe 146, on its first-ever mission to India to gather new monsoon observations (Figure 1). The flight observations were accompanied by ground-based observations from towers that measure the surface temperature, soil conditions and fluxes of temperature and moisture into the atmosphere as well as weather balloons which were launched from a network of weather stations spread over India. We are now using the data collected to challenge and improve our forecast models, at the UK Met Office and India’s National Centre for Medium Range Weather Forecasting.

Figure 1. The FAAM atmospheric research aircraft, a modified BAe 146

Climate models generally have a coarse spatial resolution, which means that atmospheric and oceanic processes operating at a smaller scale (for example convective precipitation) cannot be predicted by physical equations but instead we have to resort to a process called parametrization. This will help us to represent the effects of finer scale processes such as clouds on the coarser scale meteorology. Due to this inadequacy, climate models when used to simulate the monsoon cannot capture many important features of the monsoon. The most effective way of solving this problem is to use a model with finer resolution (grid spacing). The Met Office Unified Model (MetUM – Brown et al., 2012) uses the same dynamics and physics both at numerical weather prediction and climate projection scales. Hence it is a perfect tool to better understand the physical mechanisms involved in monsoon rainfall and to determine the role of model resolution in improving monsoon forecasts.  

For our use, we have customized the MetUM with a ‘nested suite’, which is a fine-scale regional model placed inside a coarser resolution global model. Using this method, we can operate a very fine resolution (4 km grid spacing) model for South Asia and outside this ‘nest’ the model will have coarser grid resolution (~17 km over the rest of the globe). The 4 km grid-spacing is common for forecasts in the UK but is more computationally challenging for a country the size of India. Figure 2 shows the impact of the grid spacing on forecasts of monsoon rainfall during 2016.  It features a monsoon depression, a low-pressure system, that formed in the Bay of Bengal during early July 2016 and shows that it is captured well by the models. However, the details of the clouds and rainfall are much finer. The individual convective cells are apparent in the finer resolution version of the model. This type of monsoon storm is often implicated in heavy monsoon rains and flooding, such as the recent events of August 2017 in Mumbai.


Figure 2. Rainfall rate (shaded), snowfall rate (shaded), total cloud fraction (shaded) and mean sea level pressure (contours) from 4 km resolution model (left) and global model (right) on 7 July 2016. An animation for the period 1 to 7 July 2016 is available via this link: In the animation, spiral bands of rainfall associated with the monsoon depression are seen over eastern parts of India on 6th and 7th July.

Currently, we are running two experiments for the 2016 monsoon with this setup contrasting them by using different land surface conditions such as soil moisture and vegetation types, which will eventually improve our understanding of the impact of changing land-surface conditions on monsoon rains. As the project continues we will pioneer development of even-finer resolution models, down to 100 m scales, which will allow us to examine storm development and frontal weather system structures with high fidelity.

By improved theoretical understanding of the physical processes of the monsoon and an improvement in rainfall prediction, the work will allow improved agricultural planning and security of the food supply, benefiting the Indian economy.


Brown, A., Milton, S., Cullen, M., Golding, B., Mitchell, J., & Shelly, A., 2012. Unified modeling and prediction of weather and climate: A 25-year journey. Bulletin of the American Meteorological Society,  93(12): 1865-1877.

Sperber, K. R., Annamalai, H., Kang, I. S., Kitoh, A., Moise, A., Turner, A., Wang, B., & Zhou, T., 2013. The Asian summer monsoon: an intercomparison of CMIP5 vs. CMIP3 simulations of the late 20th century. Climate Dynamics, 41(9-10), 2711-2744.

Posted in Climate, Climate modelling, Monsoons, Numerical modelling, Seasonal forecasting | Leave a comment

Recent progress on decadal prediction in the North Atlantic

by Jon Robson

The North Atlantic is a region of the Earth that is characterised by pronounced multi-decadal variability in surface temperatures – a phenomenon that has become known as Atlantic Multi-decadal Variability (AMV, see Sutton et al for a short review). The North Atlantic also appears to be one of the most predictable areas on Earth, with surface temperature and ocean heat content in the subpolar North Atlantic (~50-70°N) apparently predictable up to a decade in advance. This is encouraging given that AMV has also been associated with considerable societal impacts, including modulating the number of Atlantic hurricanes and rainfall over the Sahel (see Figure 1f and 1g).

Figure 1. An overview of Atlantic Multi-decadal Variability (AMV). (a) shows sea surface temperature anomalies for global (blue) and Atlantic (red) and the resulting AMV index (black). (b) and (c) shows proxies of ocean strength in the North Atlantic and sulphate aerosol precursor emissions. (d) shows the spatial extent of the AMV pattern. (e), (f) and (g) shows variability in winter time North Atlantic Oscillation, Hurricane accumulated energy and Sahel Rainfall. Thick lines show 10 year running means.

We recently argued that the high level of predictability of subpolar North Atlantic temperature is consistent with the initialisation of – but not necessarily the prediction of – the thermohaline component of the ocean circulation. Put another way, hindcasts appear skillfully to predict large changes in temperature only when they are started from an anomalous ocean circulation, rather than being able to predict the onset of an anomalous ocean circulation itself. Nevertheless, it is the slow evolution of the anomalous ocean circulation which allows us to predict North Atlantic upper-ocean temperature in advance. The prediction of upper-ocean temperatures can also lead to skillful predictions of other variables; for example, recent advances in 2-5 year predictions of Sahel rainfall and summer surface temperature over China are both related to improved predictions of the North Atlantic.

The particular view of the slow changes in the North Atlantic being governed by slow changes in the thermohaline circulation hasn’t really moved on from the main paradigms of the early 2000s. However, there have been a number of challenges to this view. In particular, multi-decadal changes in regional forcings (particularly Anthropogenic Sulphate Aerosols), or local thermodynamic coupling of the atmospheric variability, have both been proposed to be the main controlling factor of AMV. Long-story-short, we know that many models’ simulation of the Atlantic and AMV is deficient when compared to observations, and the role of the external forcing, in particular, is a major uncertainty. For decadal predictions, the forcings certainly provide skill, particularly in the tropical Atlantic. However, apart from the long-term warming trend, little is known about the role of different forcing factors (e.g. volcanic or anthropogenic aerosols, or solar), nor do we understand how the forcings are leading to skill, or even if the forced responses are realistic.

So where will progress come from over the next few years? Well, interesting changes abound in the North Atlantic at the moment. The subpolar North Atlantic may be transitioning into a cold state similar to that last observed in the 1970s-1980s (see Figure 2), and many people (including yours truly) have published tentative predictions of a further cooling of the North Atlantic. The Atlantic is also now observed at unprecedented levels of detail (for example the RAPID and OSNAP programs), so these changes will be watched closely.

Figure 2. 0-700 m heat content anomalies in the North Atlantic subpolar gyre region computed from the NODC data set. Figure created using KNMI climate explorer.

On the modelling side, recent advances in predicting the North Atlantic Oscillation on multi-year timescales could open the doors to further Atlantic-wide improvements. Near-term Climate Prediction is also now a World Climate Research Program “Grand Challenge”. Finally, international modelling experiments like the Decadal Climate Prediction Project (DCPP, a CMIP6 endorsed MIP) will continue the exploration of “near-term” climate prediction. DCPP will also further co-ordinate the real-time decadal prediction efforts of the community, as well as more process focused sensitivity studies. More generally, the wider CMIP6 activities (from highResMIP to VolMIP) also offer new opportunities.

Finally, the UK ACSIS project  is beginning better to co-ordinate and integrate the UK’s scientific expertise in observations and modelling across atmosphere (including composition), ocean and cryosphere, in order to tackle the multi-faceted, multi-disciplinary problem of understanding multi-decadal timescale variability in the North Atlantic.

So, taken altogether, there is a lot of North Atlantic Science to look forward to; I just wish I could find the time to look at all the things I want to!


Sutton, R. T., McCarthy, G. D., Robson, J., Sinha, B., Archibald, A. and Gray, L. J., 2017. Atlantic Multi-decadal Variability and the UK ACSIS Programme. Bulletin of the American Meteorological Society. ISSN 1520-0477 doi: 10.1175/BAMS-D-16-0266.1

Yeager, S. G. and Robson, J. I., 2017. Recent progress in understanding and predicting Atlantic decadal climate variability. Current Climate Change Reports, 3 (2). pp. 112-127. ISSN 2198-6061 doi: 10.1007/s40641-017-0064-z

Monerie, PA., Robson, J., Dong, B. et al., 2017. A role of the Atlantic Ocean in predicting summer surface air temperature over North East Asia? Climate Dynamics. doi: 10.1007/s00382-017-3935-z

Robson, J., Polo, I., Hodson, D. L. R., Stevens, D. P. and Shaffrey, L. C., 2017. Decadal prediction of the North Atlantic subpolar gyre in the HiGEM high-resolution climate model. Climate Dynamics. ISSN 0930-7575 doi: 10.1007/s00382-017-3649-2

Sheen et al, 2017. Skilful prediction of Sahel summer rainfall on inter-annual and multi-year timescales, Nature Communications 8, doi: 10.1038/ncomms14966

Dunstone N, Smith D, Scaife A, Hermanson L, Eade R, Robinson N, Andrews M, Knight J., 2016. Skilful predictions of the winter North Atlantic oscillation one year ahead. Nat Geosci 9:809–814. doi:10.1038/ngeo2824

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

Without the Tibetan Plateau, what would happen to the Asian summer monsoons?

By Mike Wong

The Tibetan Plateau is the highest and most extensive plateau in the world, with an average elevation exceeding 4000 metres and stretching over 2.5 million square kilometres. While it is often called the ‘rooftop of the world’, it also serves as the ‘water tower of the world’. Many major rivers in Asia, including the Yangtze, Mekong and the Ganges, originate from the Tibetan Plateau and support the livelihoods of more than 40% of the world’s population living in rapidly developing economies such as China and India.

The Tibetan Plateau is also a crucial component of the Asian climate. During late spring to early summer, its vast and elevated surface heats up rapidly and acts as a highly effective heat source for the atmosphere above. The heating from the plateau’s surface is long believed to be vital in creating ascent and supporting the meridional overturning circulation of the Indian Summer Monsoon.

However, recent studies (e.g. Boos and Kuang 2010, 2013, Wu et al. 2012) questioned the impact of the Tibetan Plateau on the Indian summer monsoon. Using idealised orography in which the Tibetan Plateau is removed, keeping only the Himalayas, these studies demonstrated that the Indian summer monsoon can be maintained in global climate model simulations. Therefore, the current consensus in the literature is that the orographic sheltering provided by the Himalayas is equally important, or perhaps more important, than the elevated surface heating from the Tibetan Plateau in maintaining the Indian summer monsoon.

Funded by the Climate Science for Service Partnership China (CSSP-China), the MESETA project aims to investigate further the role of the Tibetan Plateau in maintaining the Asian summer monsoons. Following previous studies, the Met Office’s global climate model HadGEM3 is used to perform various simulation experiments with idealised orography. Three different modifications to the orography are used to demonstrate the impact of the Tibetan Plateau, Himalayas and the Iranian Plateau on the Asian monsoons (Figure 1a-d). The simulations are performed at N96 resolution (200 km grid spacing at the equator) covering the period between 1981 and 2001, providing 20 years’ worth of data to derive the climatology of the summer monsoons. First, a control experiment is performed using the default orography and Figure 2a shows the summer (June-August) average precipitation and 850 hPa wind. Results from each sensitivity experiment is then compared to the control.

Figure 1. Surface orography used in each experiment:  a) Control; b) No Tibetan Plateau; c) Himalayas and Iranian Plateau only and d) Himalayas only.

The first experiment, No Tibetan Plateau (Figure 2b-c), focuses on the impact of the Tibetan Plateau and Himalayas on the monsoons by removing both terrains from the model. Compared to the control experiment, the Indian summer monsoon clearly weakened as indicated by the easterly anomalies over the Arabian Sea. Although there is more rainfall over India compared to the control, it is mostly likely related to HadGEM3’s bias in simulating summer rainfall in the region. There is also a reduction in rainfall over most of China, suggesting a weakened East Asian summer monsoon. Therefore, this experiment shows that without the Tibetan Plateau and the Himalayas, both the Indian summer monsoon and the East Asian summer monsoon will be a lot weaker in intensity.

Figure 2. Summer precipitation and 850 hPa wind (left column) and difference relative to control : a) Control; b-c) No Tibetan Plateau; d-e) Himalayas and Iranian Plateau only; f-g) Himalayas only.

However, things improved greatly in the second experiment when the Himalayas were put back into the model (Himalayas and Iranian Plateau only, Figure 2 d-e). Low level circulation over the Arabian Sea and India is more consistent to the control as the easterly anomalies reduced, while summer rainfall over India is also more similar to the control. Therefore, it seems that the Indian summer monsoon can indeed be maintained even if the elevation of the plateau is drastically reduced as long as the Himalayas are intact. In contrast, the East Asian summer monsoon is more sensitive to the presence of the Plateau as most of China is still receiving less summer rainfall than the control.

While the first two experiments demonstrated the crucial role of the Himalayas, the third experiment, Himalayas only (Figure 2 f-g), focuses on the importance of the Iranian Plateau. In this experiment, the Iranian Plateau is removed leaving only the Himalayan ridges in the model such that their contribution can be isolated. Without the Iranian Plateau, the summer westerlies over the Arabian Sea are weakened and the region is affected by north-easterly wind anomalies, bringing dry continental airmass into the region. Although the weakening of the westerly monsoon is not as significant as in the No Tibetan Plateau experiment, the results here show that the Iranian Plateau also exerts considerable influence on the Indian summer monsoon.

To summarise, the idealised experiments here show that the Indian summer monsoon is not sensitive to the elevation of the Tibetan Plateau as long as the Himalayas and the Iranian Plateau are present. In contrast, the East Asian summer monsoon is more sensitive to the presence of the Tibetan Plateau. Also, it is necessary to consider the impact of model bias as the monsoons in the control experiment are not perfect reconstructions, especially in terms of rainfall. To further examine how much of the results are model dependent, some of the experiments will be repeated in other climate models through the forthcoming Global Monsoon Model Inter-comparison Project (GMMIP). So, if one day the Tibetan Plateau somehow disappeared, don’t worry, the Indian summer monsoon will probably be fine (too bad for East Asian summer monsoon though …).


Boos WR and Kuaang Z., 2010. Dominant control of the south Asian monsoon by orographic insulation versus plateau heating. Nature, 463 (7278): 218-222.

Boos WR and Kuang Z., 2013. Sensitivity of the south Asian monsoon to elevated and non-elevated heating. Scientific reports 3.

Wu G, Liu Y, He B, Bao Q, Duan A, Jin FF, 2012. Thermal controls on the Asian summer monsoon. Scientific Reports 2.

Posted in Climate, Climate modelling, Monsoons, Numerical modelling | Leave a comment