How can we contribute to extreme event attribution in the Arctic?

By: Daniela Flocco

News of broken temperature records, droughts and extreme climate events are nowadays constantly present in newspapers and on social media. The study of the connection between extreme and global climate changes has become subject of an area of research called ‘extreme event attribution’, defined as the science of detecting whether anthropogenic (human-made) global warming contributed to the occurrence of extreme events. Scientists at National Oceanic and Atmospheric Administration (NOAA) have produced yearly reports since 2011 with the scope of explaining the causes of the previous year’s extreme events (see for more details).

Extreme climatic events have become of wide interest for their economic consequences, especially when they concern urban areas of the globe (Frame et al. 2020). The anthropogenic impact on extreme events can be also observed in less populated regions such as the Poles, even though it is more difficult to estimate its “cost” in the short term. A recent study (Kirchmeier-Young, et al., 2017) looked at the recent extreme Arctic sea ice September minima (focusing on the record-minimum of 2012) and assessed the human impact on these events. They found that the occurrence of extreme sea ice extent minima is consistent with a scenario including anthropogenic influence and is extremely unlikely in a scenario excluding anthropogenic influence. They also state that the inclusion of anthropogenic forcing is a necessary but not a sufficient cause at present to explain the observed sea ice extent lows.

Attribution of extreme events is strongly based on statistical studies of model forecasts and therefore relies on high resolution, physically-sophisticated models. This is particularly challenging in a changing climate where the parameterization of physical processes need to be able to capture the behaviour of unprecedented scenarios and produce representative results. In fact, the statistical analysis needed for event attribution requires climate models that have skills in forecasting the expected behaviour with respect to less predictable, rare events (Nature News, 2012). 

Researchers are engaged in a common effort to improve models performance and assess it. An example of how the improvement of the physics in a sea ice model can lead to improvement in prediction skills is work that our group (Centre for Polar Observations and Modelling), has carried out during the past few years: the implementation of a melt pond parameterization in the sea ice component of a global climate model and the analysis of the consequent improvements (Flocco et al., 2012, Schröder et al., 2014).

Figure 1: Melt ponds on Arctic sea ice (©NASA/Kate Ramsayer).

Changes in the Arctic and the Antarctic are faster and amplified with respect to the lower latitudes. A contributor of the ‘polar amplification’ is the so called the ice-albedo feedback: sea ice reflects almost entirely the solar radiation because of its high reflectance (albedo). When sea ice melts, larger areas of the ocean become exposed to sunlight; these absorb large part of the solar radiation inducing further melt. This is true also on the sea ice itself where melt ponds, puddles of water forming in spring in topographic lows from sea ice and snow melt, cause a strong increase in sea ice melt forming more melt ponds (Fig. 1). This process links the presence of ponds, in particular in the early melt season, to the amount of summer ice melt and consequently the amplitude of the minimum ice extent in September.

Figure 2: Annual cycle of Arctic mean fraction of sea-ice area covered by exposed melt-ponds in our CICE simulation. (Schroeder et al., 2014).

Figure 3: Predicted ice extent verified by use of SSM/I data for the period 1979–2013 (Schroeder et al., 2014).

The presence of melt pond on sea ice has increased over the past decades (Fig. 2), making it crucial to develop a parameterization suitable for a climate model that would be able to deal with a changing sea ice state. In fact, the increase in melt pond presence could be thought of as a proxy for air temperature rise. The inclusion of the new melt pond physical description allows skilful predictions of the sea ice extent minima in September depending on the presence of melt ponds in May (Fig. 3) and in particular, the improved model was able to predict with high confidence the sea ice extent minimum of 2012.


 Flocco, D., D. Schröder, D. L. Feltham, and E. C. Hunke, 2012: Impact of melt ponds on Arctic sea ice simulations from 1990 to 2007, J. Geophys. Res. 117, C9,

Frame, D.J., M. F. Wehner, I. Noy, and S. M. Rosier, 2020:  The economic costs of Hurricane Harvey attributable to climate change. Climatic Change 160, 271–281,

Kirchmeier-Young, M. C., F. W. Zwiers, and N. P. Gillett, 2017: Attribution of Extreme Events in Arctic Sea Ice Extent. J. Climate, 30, 553–571,

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

Nature news: Nature 489, 335–336 (20 September 2012) doi:10.1038/489335b





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

Towards a marginal Arctic sea ice cover

By: Danny Feltham

As the winter night descends on the polar oceans, the surface mixed layer cools and begins to freeze, forming a floating layer of sea ice. Sea ice is a complex and dynamic component of the climate system; it is strongly influenced by, and in turn influences, air and ocean temperatures, winds and ocean currents, and undergoes large seasonal changes, growing in extent and thickness in winter, and receding to a minimum in late summer.

The planet is warming at ~1oC per century, and amplification processes have roughly doubled the Arctic regional warming rate in recent decades. The strong decline of Arctic sea ice is a striking indicator of climate change, with the last 15 years (2005—2019) seeing the 15 lowest September Arctic ice extents in the satellite record. This decline has been a wake-up call to scientists, policy-makers, and the general public. Studies show that the loss of sea ice has already contributed to Arctic amplification of global warming, has influenced biological productivity, species interactions and disease transmission, and is impacting indigenous peoples, trade, and oil exploration, including the promotion of a growing polar ecotourism industry.

Figure 1: Schematic cartoon of the Arctic sea ice food web.  Credit to Hugo Ahlenius.

The sea ice cover, of either pole, features a dense inner pack ice zone surrounded by a marginal ice zone (MIZ) in which the sea ice properties are modified by interaction with the ice-free open ocean, particularly ocean wave-ice interaction that can break up the ice cover. (See Figure 1.) The MIZ is some 100 km or so wide and is a region of low ice area concentration consisting of a disperse collection of small sea ice floes: the reduced sea ice cover exposes greater areas of the ocean to the atmosphere, and intensifies and prolongs air-ocean exchanges of heat, moisture, and momentum, altering the circulation and properties of air, ocean, and ice, air-sea gas exchange, and carbon exchange across the air-sea interface.

The conspicuous reduction of Arctic sea ice extent, combined with the observations that the MIZ is getting wider over the last decade, has often led to the impression that the MIZ is getting larger and, in the science journalism literature (and quite a few scientific papers also), one often comes across the assumption (assertion) that the “MIZ is increasing”, often in conjunction with comments on what this will mean for the future. (I will not mention names here to spare some blushes, except to note that I have sometimes found myself guilty of such woolly thinking and found myself in good company.)

Figure 2: Arctic sea ice extent (solid line) and MIZ extent (dashed line) from model and four remote sensing products (see legend). MIZ extent is defined as the area of ocean with sea ice area fraction of between 15 and 80%. Sea ice extent is the area of ocean with ice area fraction above 80%. An error bar of 10% has been applied to all observational products.

A recent study by Rolph et al [2020] has analysed sea ice concentration data from a range of sources and, using a commonly-used definition of the MIZ extent as the area of that region of the ocean with ice area fraction between 15 and 80%, analysed changes in the Arctic MIZ extent for the first time. While there are some significant caveats concerning the accuracy of ice concentration data, particularly in the summer, a conclusion from this study is that there is little evidence that the MIZ extent is increasing or decreasing and, in fact, appears to be fairly constant over the last three to four decades. You can see this from the top panel of Figure 2 which shows Arctic sea ice extent as estimated from various means (solid lines) and the MIZ extent (dashed lines).

What appears to have been happening is that, on a monthly average basis (to average over wind-induced fluctuations of the ice cover) there has been a decadal trend for the central pack ice of the Arctic Ocean to recede and move north and the MIZ has moved north with the pack. While the MIZ region has widened at a rate of ~1.5 km/year, its extent has remaining roughly constant. This is possible because the perimeter of the MIZ has, on average, decreased in proportion to the increase in width. (This is further evidence, should one need it, that the Earth really is round!)

So, the MIZ has been migrating north but not changing in area. Does this matter? As indicated above, the MIZ is a region of enhanced air-ocean heat, moisture and momentum exchanges and the location of these exchanges is relevant to local weather and oceanography. But, perhaps more dramatically, the MIZ is also a region of marine primary production, delivery of nutrients to the euphotic zone, and a hunting platform for polar bears and indigenous communities (Figure 1). Movement of the ice edge northwards is transforming the lives of local peoples and wildlife.

While the MIZ extent may not have been changing in recent times, the fraction of the Arctic ice cover that is the MIZ has been increasing, see the bottom panel of Figure 2.  So, among other things, it seems the processes that dominate in the MIZ, such as wave-ice interaction, are becoming increasing important for the remaining Arctic ice cover whereas they have in the past been of only marginal significance (pun intended).

Figure 3: Left: current and projected changes in the MIZ [Strong and Rigor, 2013; Aksenov et al, 2017]. Right: projected MIZ in the 2030s in summer (June-August) [Aksenov et al, 2017].

If the increasing trend of MIZ fraction were to continue, one may expect the entire Arctic Ocean to eventually become marginal (before being eliminated entirely). Figure 3 shows a projection of Arctic MIZ area fraction and a snap-shot map in 2030. The implications of loss of Arctic sea ice cover are still being worked out in climate modelling and field studies (notably the recent US Office of Naval Research MIZ field program) but, likely as not, there will be as many unknown unknowns as known unknowns. One thing, however, seems fairly clear: the nature of air-sea ice-ocean exchanges and feedbacks will alter in the coming decades and these interactions will depend on physical representations of MIZ sea ice processes that have never needed to be included in models before.


Aksenov, Y., E. E. Popova, A. Yool, A. J. G. Nurser, T. D. Williams, L. Bertino, and J. Bergh, (2017) On the future navigability of Arctic sea routes: High-resolution projections of the Arctic Ocean and sea ice, Mar. Policy, 75, 300-317,

Strong, C. and I. G. Rigor, (2013) Arctic marginal ice zone trending wider in summer and narrower in winter, Geophys. Res. Lett., 40(18), 4864–4868,

Rolph, R. J., D. L. Feltham, and D. Schroeder, (2020) Changes of the Arctic marginal ice zone during the satellite era. The Cryosphere. ISSN 1994-0424 doi:  (In Press)



Posted in Climate | Leave a comment

Covid-19: Using tools from geophysics to assess, monitor and predict a pandemic

By: Alison Fowler, Alberto Carrassi, Javier Amezcua

The emergence of a new coronavirus disease, known as Covid-19, that could be transmitted between people was identified in China in December 2019. By 3rd March 2020 it had spread to every continent except Antarctica, totalling 92,840 confirmed cases and 3,118 deaths.

As scientists worldwide scrambled to understand this new virus, a fundamental and immediate question was how many more people are likely to die and what impact can governmental interventions have?

To answer this question, we have two valuable resources available to us. The first are numerical models, which have identified the key equations that can be used to explain a pandemic. The second are observational data, which detail the number of deaths and hospitalisations due to Covid-19 that have occurred to date. Neither models nor observations are perfect but by combining (assimilating) them, we can utilise the best parts of both whilst minimising their flaws. A huge benefit of data assimilation is that it also provides a robust estimate of the uncertainties of the output, offering an understanding of the worst, best as well as the most likely situation.

Assimilating observational data with models is routinely performed in geophysics. In fact, data assimilation is fundamental to modern day weather forecasting. Evidence for this is provided by the step change in the accuracy of weather forecasts that has been possible with the increasing availability of information from satellites orbiting the earth.

 Can data assimilation tools that have been developed for the geosciences be applied to pandemic modelling?

A team of scientists from 8 different countries (Argentina, Canada, England, France, Netherlands, Norway, Brazil and the United States of America) diverted their attention from geophysics for a few months to examine this very question. Each employed a state-of-the-art data assimilation tool typically used for geophysical problems to explain and predict the course of the pandemic in their own host country. The evolution of the epidemic is seen to vary widely between these 8 countries. Factors affecting this include differences in location (e.g. hemisphere), population densities, social habits, health-care systems, and importantly the government interventions employed. 

It was found that by using data assimilation to derive key parameters of the pandemic we could fit a classic metapopulation model to explain the reported deaths and hospitalisations in each of the 8 countries. The model itself is a version of the Susceptible-Exposed-Infected-Recovered (SEIR) compartment model that has been adapted to Covid-19 by including age-stratification and additional compartments for quarantine and care-homes. This is analogous to compartmental models often used in geophysics such as those used for studying carbon dynamics. Using this approach, we were able to successfully represent the impact of the (very different) interventions taken in the 8 different countries; visualising the rapid drop off in person-person transmission on different dates of lockdown.

Given the success of data assimilation to explain the reported deaths, the next step is to provide predictions under different possible scenarios. For England we took three possible scenarios from the 1st June when lockdown began to be eased. These were defined in terms of, the now familiar variable, the R number, which quantifies the average number of people an infected person will pass the virus onto. The three values that were chosen were 0.5 (reduction in number of cases with time), 1 (steady number of cases with time) and 1.2 (increase in number of cases with time). 

As of 1st June, approximately 45,000 deaths were attributed to Covid-19 in England in all settings (source, ONS). Our projections under the three different scenarios predict that by the 1st September the total deaths will be 57,000±1,900 (R=0.5), 63,600±2,700 (R=1) and 76,400±4,900 (R=1.2).  Given how widespread Covid-19 already is in England, these results highlight the potential of measures, which reduce a large amount of person-person contact, to save tens of thousands of lives. The uncertainty in the numbers reflects the uncertainty in the simple model and the uncertainty in the reported values. The collection of data on deaths, hospitalisation and number of positive cases is marred by a myriad of political and social complications, problems we do not normally need to consider when measuring winds and rainfall.

Figure: Evolution of the Covid-19 epidemic in England. Top: Total deaths. Middle: Number in hospital.  Bottom: The estimated R value (average number of person-person transmission). The black dots show the reported values up to 5th June for deaths (source, ONS) and up to 12th June for number in hospital (source, daily Gov. Press Conference). Blue lines indicate the initial estimates and the red lines indicate the values after assimilation, with the bold line indicating the most likely value. After 1st June three predictions are made based on three different R values.


Geir Evensen, Javier Amezcua, Marc Bocquet, Alberto Carrassi, Alban Farchi, Alison Fowler, Peter Houtekamer, Christopher K. R. T. Jones, Rafael de Moraes, Manuel Pulido, Christian Sampson, Femke Vossepoel: An international assessment of the COVID-19 pandemic using ensemble data assimilation. Submitted to Foundations of Data Science. Preprint on medRxiv. doi:

Posted in Climate, data assimilation, Numerical modelling | Leave a comment

Finding the skill of forecasts of extreme precipitation in Southeast Asia

By: Samantha Ferrett

Forecasting weather in Southeast Asia

Southeast (SE) Asia is prone to high‐impact weather and is often subject to flooding and landslides as a result of heavy rainfall. Just last month Indonesia was hit by heavy rainfall that resulted in floods and landslides because of a rainy season that lasted longer than was initially forecast. Global computer models used for Numerical Weather Prediction (NWP) have been known to fail to accurately capture Maritime Continent rainfall, limiting predictions of high‐impact weather in the region. I am a Research Scientist on a Weather and Climate Science for Service Partnership (WCSSP) Southeast Asia project, “FORecasting for SouthEast Asia” (FORSEA), that aims to improve forecasts in SE Asia to reduce social and economic losses from high impact weather events. In this blog, I will provide an overview of some of my recent work that examines how well newly developed ensemble forecasts reproduce extreme precipitation in SE Asia.

What is an ensemble forecast?

A deterministic forecast is a single forecast from a computer model using one initial condition and producing one final estimate of future weather. The initial condition is an estimation of the observed weather at the start of the forecast. There are multiple reasons for an incorrect forecast. For example, one cause is that the model may not be able to fully replicate processes that drive weather in the real world. This is why forecasts are just an estimate of future weather. Unfortunately, there is also some uncertainty even to the observed weather that can result in large errors in the final forecast, even with a ‘perfect’ forecast model. An ensemble forecast consists of multiple forecasts from the same model, each with slightly different initial conditions, representing the uncertainty in observations. This results in an ensemble of estimates of future weather that can then be used to gain an understanding of the uncertainty of the forecast.

Are convection-permitting ensemble forecasts worth it?

Figure 1: Schematic of rainfall in a coarser resolution forecast model (left) and rainfall in a high-resolution convection permitting model (right). Darker blues indicate more rainfall.

A forecast model divides the region to be forecast into a grid. Convection-Permitting (CP) forecasts are those that use NWP models with such small grid sizes that they can better represent processes associated with rainfall. A schematic showing the difference between a coarser grid and a high-resolution grid used in CP models is shown in Fig. 1. The downside is that CP models, and ensembles, are more computationally expensive. Modellers face a difficult task in striking a balance between cost and benefit; this is where those of us who analyse such models hope to be useful! It’s important for the modelling community to know if the resources invested in these forecasts are worth it.

In my work I examine how “skilful” forecasts of extreme rainfall are for CP ensembles of forecasts in Malaysia, Indonesia and the Philippines. These are ensembles of 17 forecasts at a resolution of 4.5km (like the schematic in Fig. 1 shows) and were run by the Met Office between October 2018 to March 2019. SE Asia has a strong daily cycle of precipitation where precipitation is over land during the day and moves over ocean during the night. A question to answer is if these normal daily variations of rainfall remove the need for CP forecasts – is rainfall so dominated by the daily cycle that there is no need for these high resolution forecasts?

Figure 2: Fractions Skill Score (FSS) of 3 hourly accumulated precipitation at 8pm-11pm local time (Malaysia) exceeding 95th percentile aggregated over all forecasts in Oct 2018-Mar 2019 as function of spatial scale (x-axis) a) Malaysia, b) Indonesia and c) Philippines. The horizontal line shows the FSS=0.5 “skilful” threshold. Lines show results from the ensemble forecast for 1, 3 and 5 days into the forecast (black, mid grey and light grey solid lines) and results from a forecast based on observed weather from 1, 3 and 5 days before the day to be forecast (black, mid grey and light grey dashed lines).

I compare the skill (using a metric called the Fractions Skill Score) of the ensemble forecasts, shown by the solid lines in Fig. 2, to a “persistence” forecast, shown by the dashed lines in Fig. 2. The persistence forecast does not use a model but instead uses observed weather from the days prior to the day being forecast to estimate the weather. The forecast is considered skilful at the spatial scale shown on the x-axis if the metric exceeds the threshold shown by the horizontal black line. The ensemble forecast is much more skilful. The larger skill at lower spatial scales means that smaller scale features can be more accurately forecast by the ensemble. Even skill five days into the ensemble forecast (shown by light grey solid line) is higher than that of the first day of the forecast based on observations (black dashed line). This means there is value in using such a forecast in all three regions.

It’s not over…

This is promising news for the use of CP models in the tropics, but questions still remain to be addressed in FORSEA:

  • How do common large scale features known to modulate SE Asia rainfall, such as the Madden Julian Oscillation or equatorial waves, influence forecast skill?
  • Shall we go smaller? This suite of forecasts also includes sub-kilometre scale forecasts. Is there benefit to using these?


Clark, P., N. Roberts, H. Lean, S. P. Ballard and C. Charlton‐Perez, 2016: Convection‐permitting models: a step‐change in rainfall forecasting. Met. Apps, 23, 165-181.

Ferrett, S., G.‐Y. Yang, S. Woolnough, et al., 2020: Linking extreme precipitation in Southeast Asia to equatorial waves. Q J R Meteorol Soc., 146, 665– 684.

Love, B.S., A. J. Matthews, and G. M. S. Lister, 2011: The diurnal cycle of precipitation over the Maritime Continent in a high‐resolution atmospheric model. Quarterly Journal of the Royal Meteorological Society137, 934– 947,

Roberts, N.M. and H.W. Lean, 2008: Scale-Selective Verification of Rainfall Accumulations from High-Resolution Forecasts of Convective Events. Mon. Wea. Rev., 136, 78–97,

Posted in Numerical modelling, Rainfall, Weather forecasting | Leave a comment

Sunshine, cracks, fly-tipping, aphids: Phew, what a May!

By: Roger Brugge

Those spending more time in their gardens and in local parks and woods in and around Reading due to the current lockdown situation will have noticed how the ground surface had been resembling the look and feel of a badly-laid concrete patio – grey in colour, very hard, somewhat uneven and with some rather large cracks. It has very definitely dried out since I was sliding around in the mud in my local woods while walking my dogs this past winter.

Cracks in the ground in the Maidenhead Thicket woods.

Indeed, in my home location of Boyn Hill in west Maidenhead, May 2020 was the second driest May since before 1859.

Monthly rainfall totals during May for the Boyn Hill area of Maidenhead, 1859-2020.

Compared to an average May rainfall of 53 mm, just 2.7 mm fell this year on two days. In fact, the rainfall spells lasted barely 40 minutes in total. If we go back to 1990, we find our driest May when 2.1 mm fell – while in 1979 we had our wettest May with 126 mm of precipitation.

In Reading, May 2020 has been even drier, with just 1 mm of rainfall – the driest May in a record back to 1901, with other dry Mays having occurred in 1990 (3.1 mm), 1956 (5.1 mm) and 1919 (5.9 mm). However, go into Oxfordshire and the weather station at RAF Benson recorded just a ‘trace’ – i.e. no measurable rainfall – during May 2020. This lack of rain to wash away pests has led to an abundance of black fly on my broad beans and aphids in the garden!

Locals may remember how the River Pang (which flows through nearby Pangbourne) ran dry as the summer months of 1990 brought a summer heatwave. In Windsor, people were encouraged to share their bath water (after use) with trees in the Castle gardens. Hosepipe bans, tarmac turning to treacle and poisonous blue-green algae at Dinton Pastures Country Park were other features of the summer along with a maximum temperature of 35.5 °C on 3 August – still the fourth highest maximum temperature of all time in the Reading record. One wonders what lies ahead for this summer.

But, be warned. The summer of 1956, in contrast, was rather poor – in Reading only six days in July and one in September reached 25 °C. In 1919, 30 °C was reached on a total of three days at the University (but as late as August and on 11 September – one of the warmest September days in the Reading record). Looking back further, the very dry May of 1895 was followed by a wet second half to the year in Berkshire, with similar features to the rainfall distribution occurring in 1896 after another very dry May.

In Reading, one of the measurements made at 0900 UTC each day by our weather observers is that of the state of the ground, using the code shown in this table (assuming that there is no lying snow):Codes used to describe the state of ground (without lying snow) in the UK.

During late spring and summer in Reading the state of the ground at 0900 UTC is usually described as dry or moist; it might be wet following a heavy downpour within the previous 24 hours or be described as being ‘dry with cracks’ after a warm, sunny, dry spell lasting (say) 7-10 days.

May 2020 has seen our observers noting that the soil was moist (code 5/1, on the 1st due mainly to some rainfall at the end of April), dry (code 5/0, on the 2nd-4th) but extremely dry with cracks thereafter – so 27 days with cracked soil). A typical year has about 12 days when a code of 5/9 is recorded – and we had 25 such mornings in April 2020 also!

Following the wet spell from late September 2019 to early March 2020, such a ground state is quite surprising. But then spring 2020 has also been the sunniest on record in Reading (in a record back to 1956) and across the wider area of Central Southern and South-East England since before 1929:

Spring sunshine totals in Cent S and SE England (black, 1929-2019, data courtesy of the Met Office) and at the University of Reading (red, 1958-2020).

Monthly sunshine totals this spring (as measured by a Campbell-Stokes sunshine recorder) in Reading have been as follows:

  • March 2020 – 170.2 h (seventh sunniest March on record; the sunniest was 178.7 h in 2007)
  • April 2020 – 250.9 h (the sunniest April on record; previous record was 234.0 h in 1984)
  • May 2020 – 345.1 h (the sunniest month of any name on record; previous record 305.6 h in June 1975)

This makes a spring total of 766.5 hours – the sunniest spring on record at the University, while only three summers have been sunnier than spring 2020 in Reading.

The University’s Campbell-Stokes sunshine recorder. A glass sphere focuses the sun’s rays on to a graduated card and the length of the burn trace on the card corresponds to the duration of sunshine.

Daily mean sea level pressure at 0900 UTC during spring 2020 at the University of Reading.

These sunny conditions are not unrelated to the generally high pressure we have had for much of spring. At 0900 UTC the mean sea level pressure in Reading has often been in the range 1020-1040 hPa, except for two notable cyclonic spells in early March and at the end of April/beginning of May. Winds have also often been between the drier north-east and south-east, rather than the more usual south-westerly quadrant of the compass. High pressure, associated with generally descending air, often helps to reduce cloud cover in the spring/summer, leading to a reduction in precipitation and, hence, sunnier conditions.

Daily wind direction at 0900 UTC during spring 2020 at the University of Reading.

In May 2020, the average pressure at 0900 UTC was 1022.1 hPa – the third highest May value (after May 1991 and May 1944) in the Reading record since 1908.

Clear skies can sometimes lead to cool nights as well as warm days. In fact, during a cool northerly flow of air in mid-May the grass minimum temperature in Reading fell to -9.3 °C on the 15th, thereby doing severe damage to delicate crops such as potatoes. My potatoes suffered too – although meteorological foresight did allow me to carry out what looked like a bad case of fly-tipping on my allotments in a successful attempt to minimise the damage by covering most of the early crop:

Anti-frost measures on the author’s allotments in Maidenhead, mid-May 2020.

This spring in Reading the mean daily maximum temperature has been 2.5°C above average, while nights have been slightly cooler than normal overall, with the mean daily minimum temperature 0.2°C below average, with 50 ground frosts (compared to the 38 we would normally expect in spring). Overall, it was the warmest spring since 2017.

In summary, provided that we have been willing to water our gardens and allotments, the weather conditions have generally been quite conducive to enduring the prevailing lockdown conditions – indeed, my dogs have often picked up the whiff of evening barbeques!

Posted in Climate, Rainfall, Weather | Leave a comment

Predicting the impacts of tropical cyclones – when should we trust the forecasts, and when should we not?

By: Rebecca Emerton

Between September 2018 and May 2019, a record-breaking 15 tropical storms moved through the southern Indian Ocean. This marked the first season where two intense tropical cyclones, Idai and Kenneth*, made landfall in Mozambique, causing devastating flooding and affecting more than 3 million people, with over a thousand fatalities. As I write this, Cyclone Amphan, the strongest tropical cyclone on record in the Bay of Bengal, is heading towards the coast of India and Bangladesh, causing significant concerns due to extreme winds, rainfall and flooding from storm surge – all while also trying to prevent the spread of COVID-19. The anticipation and forecasting of natural hazards such as tropical cyclones and the flooding they cause, is crucial to preparing for their impacts. But it’s important to understand how well forecast models are able to predict these events, and the limitations of the forecasts – this information informs decision-making by national meteorological agencies, humanitarian organisations and the forecast-based financing community on how best to interpret forecasts of tropical cyclones. When should disaster management agencies trust a forecast for a landfalling tropical cyclone, and when should they not?

Satellite image of intense tropical cyclone Idai, 12th March 2019, overlaid with the ECMWF (European Weather Centre) forecasts of the tropical cyclone’s track towards central Mozambique.

I’m currently working on a research project called PICSEA** (Predicting the Impacts of Cyclones in South-East Africa), which aims to provide insight into how well different forecast models are able to predict tropical cyclones and their associated hazards in the south-west Indian Ocean. In this blog post, I’ll talk about our visits to the national meteorological services we’re working with in south-east Africa, and highlight some of our key results so far, for how well we’re able to predict tropical cyclones in the region.

Working with national meteorological services in south-east Africa

We’re working with the Mozambique, Madagascar and Seychelles national meteorological services and the Red Cross Climate Centre on research that we hope will be of use to forecasters faced with providing warnings for tropical cyclones, and humanitarian organisations tasked with taking early action ahead of and during these events. In 2019, we were able to visit all three meteorological services, to meet the forecasters and researchers, to learn about their methods and learn from their local knowledge and experience. We were also able to discuss our research plans to gain their perspective and ideas on what questions were most important to tackle, and how best to collaborate on answering those questions.

Left: visiting the Limpopo River in Mozambique, discussing flooding from cyclones and spotting wild crocodiles. Right: at a cyclone discussion meeting at the Technical University of Mozambique (UDM) in Maputo, organised by the Universities of Reading, Bristol and Oxford, in collaboration with the national and regional hydrological services, the national meteorological and disaster management agencies, the Red Cross (Cruz Vermelha Moçambique) and UDM.

I came away from these visits with a much better understanding of the procedures around forecasting tropical cyclones in each country and the challenges faced (ranging from which forecasts to trust, to communicating forecasts and warnings in areas where many different local languages are spoken), alongside some great ideas for our research – which models to focus on, which questions would be the most useful to answer (and which would be less useful!), and how best to communicate and visualise our results. A further trip to Mozambique also gave us the opportunity to meet with representatives from the national meteorological, hydrological and disaster management institutes, the Red Cross, and academics from the Technical University of Mozambique, to discuss experiences of forecasting and responding to Cyclones Idai and Kenneth, and furthering the collaborations between the various national and international organisations. I also took a little time off to explore a part of the world I’ve been wanting to visit for a very long time!

Taking some time off to go hiking in the Seychelles (top left) and explore Madagascar’s rainforests, mountains, deserts and beaches, spotting lemurs (top right & bottom left) and baobabs (bottom right).

What does the research show?

Our research has focussed on evaluating how well the UK Met Office and European Centre for Medium-Range Weather Forecasts (ECMWF) ensemble and high-resolution deterministic models are able to predict tropical cyclones in the south-west Indian Ocean. We’re looking into how closely the models are able to predict the path the tropical cyclones will take, the amount of rain they’ll produce, and whether the models correctly capture the strength of the tropical cyclones. In the next couple of paragraphs, I’ll briefly discuss some highlights from our results for the UK Met Office forecasts (while we continue to process data for ECMWF!).

We’ve been looking at forecasts of tropical cyclones over the past 10 years, and we’ve seen significant improvements in the accuracy of the forecasts over that time, for both the path/track of the storms, and also the intensity/strength. We know that forecast models in general can often struggle to predict the intensity of tropical cyclones, typically under-estimating their strength, but it’s encouraging to see significant improvements in the forecasts of pressure and wind speed as the models are upgraded.

Change in track error over the past 10 years, for an example predicted storm. If, 3 days in advance, you were to predict a tropical cyclone to be in the centre of the circle, then based on the typical errors across many forecasts, the actual position of the storm in 3 days’ time could end up being anywhere within the blue circles. The left map shows the UK Met Office deterministic forecast model from 10+ years ago (July 2006 – March 2010), the right map shows the current version of the model (running since July 2017). The dark shaded circle indicates the average track error, and the lighter circle indicates the error for the worst forecast of any storm using that model.

Errors in UK Met Office deterministic forecasts of tropical cyclone rainfall, in three different versions of the model from N320 (July 2006 – March 2010), N512 (March 2010 – July 2014) and N768 (July 2014 – July 2017). Significant over-estimations of the rainfall around Madagascar in the N320 version of the model are improved in later versions of the model, but results vary by region, with the forecasts predicting too much rainfall in some areas, too little in other areas.

We’ve also been looking into whether certain background tropical conditions impact the accuracy of tropical cyclone forecasts. We’ve found that there are certain phases of the Madden-Julian Oscillation (which is the eastward movement of a large region of enhanced and suppressed tropical convection, and the different phases describe the location) during which forecasts of tropical cyclone intensity are improved (phases 1 and 2  for forecasts up to 4 days ahead) compared to other phases. With further research, we hope to identify other such potential ‘windows of opportunity’ for more accurate forecasts, and also highlight times when forecasts may be less accurate than usual – information that is also key for forecasters and decision-makers.

What’s next?

While there’s always plenty more research to be done, we also plan to work with the Red Cross Climate Centre and our project partners in Mozambique, Madagascar and the Seychelles, to produce forecast guidance documents and infographics. These could be used to communicate the outcomes of the research alongside other key information, in a way that can be used by the forecasters who are monitoring, predicting and providing warnings for tropical cyclones in the region. As well as looking at forecast models, national meteorological services also receive regional forecasts and warning bulletins for tropical cyclones from the Regional Specialised Meteorological Centre, which for the south-west Indian Ocean is Météo-France in La Réunion.

While the current pandemic situation may have sadly changed our plans to host visiting scientists here at the University of Reading for two months to work together on this (we had the pleasure of working with Hezron Andang’o and Lelo Tayob from the Seychelles and Mozambique meteorological services before their time in Reading was cut short back in March), I look forward to continuing our work remotely with all of our partners for now, and hopefully to visiting each weather service again in the future!

* Read more about the forecasts and response to Cyclones Idai and Kenneth in this presentation and keep an eye out for our upcoming paper:

Emerton, R., Cloke, H., Ficchi, A., Hawker, L., de Wit, S., Speight, L., Prudhomme, C., Rundell, P., West, R., Neal, J., Cuna, J., Harrigan, S., Titley, H., Magnusson, L., Pappenberger, F., Klingaman, N. and Stephens, E., 2020: Emergency flood bulletins for Cyclones Idai and Kenneth: the use of global flood forecasts for international humanitarian preparedness and response, under review

**PICSEA is a SHEAR Catalyst project funded by NERC and DFID, led by Nick Klingaman with Kevin Hodges, Pier Luigi Vidale and Liz Stephens, and international project partners Mussa Mustafa (INAM, Mozambique), Zo Rakotamavo (Météo Madagascar), Vincent Amelie (SMA, Seychelles) and Erin Coughlan de Perez (Red Cross Climate Centre). Oh, and I do the data analysis!

Posted in Africa, Tropical cyclones, Weather forecasting | Leave a comment

Sea Surface Temperature Climate Data Record

By: Owen Embury

Oceans cover over 70% of the Earth’s surface and knowing its temperature is crucial for understanding both weather and climate. Historically, sea surface temperatures (SSTs) have been measured in situ – from ships and automated buoys – however, in recent decades we have also been able to use satellites to make observations of the Earth system. While satellites make indirect observations of the SST by measuring the thermal radiation emitted by the sea, they can provide significantly more coverage than in situ observations as a single satellite can see the whole Earth over a day or so. We now have a global, gap-free, 38-year time-series of SST generated as part of the European Space Agency (ESA) Climate Change Initiative (CCI) for SST and continued under the Copernicus Climate Change Service (C3S).

Our SST data are generated from infra-red sensors flown on-board polar orbiting satellites. These satellites circle the Earth in a low orbit (around 800 km) travelling in a north/south direction passing over each pole and taking about 100 minutes for each orbit. Each satellite will be able to see the whole of the Earth’s surface over a day or so as the Earth rotates. We use two main types of infra-red sensors to measure the SST. The first are the Advanced Very High Resolution Radiometers (AVHRRs) – a series of meteorological instruments which have been in use since the late 1970s. By modern standards, these instruments are no longer advanced or high resolution! The second set of instruments are the Along Track Scanning Radiometers (ATSRs) and Sea and Land Surface Temperature Radiometers (SLSTRs). The instruments were specifically designed to make high accuracy climate observations of the surface temperature and can be used to improve the accuracy of the AVHRR observations.

Figure 1: Global mean diurnal cycle of daily SST anomalies from drifting buoys for different wind speeds. The annual mean diurnal cycle is displayed in black, Northern Hemisphere winter (DJF) in blue, spring (MAM) in green, summer (JJA) in red and autumn (SON) in orange. The bars in the top left corner show the associated uncertainties. Reproduced from Morak-Bozzo et al. 2016.

When generating a climate data record, we are interested in how the SST changes over long time periods (years and decades); however, the SST changes throughout the day as the ocean warms during the day and then cools again overnight (the diurnal cycle). In order to separate these two effects, we provide the satellite-based temperature at a standardised time of day of 10:30 am or 10:30 pm local time (which is a good estimate of the daily average temperature). Figure 1 shows how the SST changes through the day at different wind speeds. The cycle is largest at low wind speeds where the surface warms rapidly from dawn to mid-afternoon. The cycle becomes smaller as the wind speed increases, because the wind causes the surface water to mix with lower water resulting in a more constant temperature.

Figure 2: – SST product levels. L2P data are the original satellite viewing geometry with gaps due to cloud cover and land. These are gridded to L3U, and then combined to produce daily L3C. Data from multiple sensors are combined and interpolated to produce a gap-free L4 SST.

SST data are provided on multiple “levels”, each adding a bit more processing to make the data more convenient for users. The lowest level data containing SST is known as L2P which contains the full resolution imagery as observed by the satellite. When displayed as an image, it forms a long rectangle along the satellite’s direction of travel. To make the data easier for users, we first grid it onto a global 0.05° latitude-longitude grid (L3U), where we can see the satellites orbit around the Earth. Next, all the data from one day is collated to produce L3C at which point we have near-global coverage apart from gaps due to clouds. Finally, we produce a gap-filled estimate of the daily-mean SST which combines data from multiple satellites and interpolates into the gaps to give a Level 4 analysis product. These SST product levels are shown in Figure 2.

Data and References:

Our SST data are available from the ESA Open Data Portal and Copernicus Climate Change Service

Merchant, C. J., and Coauthors, 2019: Satellite-based time-series of sea-surface temperature since 1981 for climate applications. Sci. Data, 6. 223. ISSN 2052-4463 doi:

Morak-Bozzo, S., C. J. Merchant, E. C. Kent, D. I. Berry, and G. Carella, 2016: Climatological diurnal variability in sea surface temperature characterized from drifting buoy data. Geosci. Data J., 3 (1). pp. 20-28. ISSN 2049-6060 doi:

Posted in Climate, Oceans, Remote sensing | Leave a comment

Measuring Lake Water Temperature From Space

by: Laura Carrea

‘Climate change’, and ‘global warming’: these have been two of the most referenced terms in the media in the past few years.  These words sometimes generate controversy and discussion not only on social media or between friends, but also among politicians. Those that do not support climate change measures often state that they “do not believe in it”.

In order to form an opinion on what to ‘believe’ it is crucial to refer to factual evidence from the natural world. Measurements are a clear and objective reflection of what happens in nature and, by definition, they should not create debates. You would be unlikely to hear discussions on whether the temperature this morning was 15°C if it was measured with a reliable thermometer.

My work is on measuring temperatures of inland water, which means lakes, reservoirs, lagoons and rivers, and I try my best to cover lakes all around the world in the most diverse environments. Currently, I am monitoring more than 1,000 lakes throughout the world.

How do I measure temperatures for such a wide range of lakes?

I mainly use satellites, but also the kindness of many people and institutions around the world who measure the temperature of the water directly on the lake.

Satellites are a “modern” way to record temperatures of the whole globe over a day or so.  However, they cannot ‘see’ through clouds, leaving some very cloudy areas unrecorded. The satellites I am using have been designed and launched by the European Space Agency (ESA) and by the National Oceanic and Atmospheric Administration (NOAA) with the purpose of monitoring the temperature of the sea and, in general, the surface of our planet.

Given the measurements, my question is: ‘Is the temperature of the water in lakes, reservoirs, lagoons and rivers increasing or decreasing?’

First, I calculated what is called a climatology – a 20-year average of the data (in this case, for 1996 to 2015) of the temperatures for each lake I have investigated. This average value constitutes what it is called ‘the reference’ or ‘baseline’.

I have then collated temperature measurements from 1995 until 2019 and calculated the difference between the temperature of the lake and the climatology during the hottest months. In this way, I have a precise idea whether for example in 2016 a particular lake was warmer or cooler than the 20-year average. This difference is called an anomaly.

I have studied more than 900 lakes across the globe, of which 127 were in Europe. I report here some of my findings for the year 2018 (Carrea L. et al., 2019), but I contribute every year to reports on the state of the climate.

I have looked at the 2018 anomalies for all the lakes I have studied and for the European lakes only. Figure 1 below shows coloured circles in locations that are approximately the position of the lakes I have investigated on the globe. For some areas, like North America and Tibet, the dots in the figure are aligned rather than overlapped. This is a way to display all the lakes in areas where the density of the lakes is very high.  A blue dot indicates a negative anomaly, which means that the lake was cooler than the 20-year average while a red dot indicates a positive anomaly which means that the lake was warmer.

This is what I have found:

  • In 2018, lakes in Europe, Tibet, New Zealand and mainly the east of the United States were warmer than the 20-year average while the rest of the lakes were cooler or had similar temperature to the 20-year average.
  • In 2018 in Europe, the vast majority of lakes were warmer than the 20-year average where the warmest lake was in Germany (lake Constance being 1.68°C warmer than average) and the least warm lake was in Iceland (0.66°C cooler than average).

Figure 1: 2018 lake average temperature anomalies in the warm season. Source:  Carrea et al. (2019)

To understand the thermal behaviour of the lakes through the years, I have plotted for each year the value of the anomaly for the European lakes only, and for all the lakes together. The plot can be seen in Figure 2 and from it we can deduce the following:

  • In 2018 the European lakes were overall 0.83°C above the 20-year average. This is the highest value since 1995 at least.
  • Putting together all the lakes I have investigated (European and non-European), overall in 2018 they were 0.17°C above the 20-year average.
  • Excluding the European lakes, I found that the rest (800 lakes) were only 0.06°C above the 20-year average, which is much less than the anomaly for European lakes.
  • Since 1995 the overall temperature of the lakes is increasing, and I have found that the European lakes are warming faster than all lakes globally.

Figure 2: Average lake surface water temperature anomalies per year for 923 lakes worldwide and or 127 European lakes. Source: European State of the Climate in 2018.

For Europe, 2018 was a very hot year, which was also confirmed by other sources of data (Toreti et al. 2019), but clearly, lakes are warming and some of them at a very worrying pace.

What is the problem with warming lakes? The temperature of the water has a strong impact on the lake ecosystem and can, among other things, lead to an increase in the development of toxic cyanobacteria blooms. The resulting poor water quality endangers the ecosystem and therefore the life that it supports, including human life.

Data and References:

 The data used here are from the GloboLakes (Carrea L. et al. (2019)) and Copernicus Climate Change Service projects. The findings reported here are part of the State of the Climate Report for the Bulletin of the Meteorological American Society (Carrea et al. 2019) and for the European State of the Climate of the Copernicus programme (

Carrea, L.; Merchant, C.J. (2019): GloboLakes: Lake Surface Water Temperature (LSWT) v4.0 (1995-2016). Centre for Environmental Data Analysis, 29 March 2019.

Carrea, L.; Embury, O.; Merchant, C.J. (2015): GloboLakes: high-resolution global limnology dataset v1. Centre for Environmental Data Analysis, 21 July 2015.

Carrea, L., Woolway, R. I., Merchant, C., Dokulil, M. T., de Eyto, E., DeGasperi, C. L., Korhonen, J., Marszelewski, W., May, L., Paterson, A. M., Rusak, J. A., Schladow, S. G., Schmid, M., Verburg, P., Watanabe, S. and Weyhenmeye, G. A. (2019) Lake surface temperature [in “State of the Climate in 2018”]. Bulletin of the American Meteorological Society, 100 (9). pp. 13-14. ISSN 1520-0477 doi:

Toreti A., Belward A., Perez-Dominguez I., Naumann G., Luterbacher J., Cronie O., Seguini L., Manfron G., Lopez-Lozano R., Baruth B. et al. 2019. The exceptional 2018 European water seesaw calls for action on adaptation. Earth’s Future. 7(6):652–663.

Posted in Climate, Climate change, earth observation, Remote sensing | Leave a comment

Do cuts in particle pollution accelerate climate change?

By: Richard Allan 

Glorious weather over much of the UK in April 2020 replaced the seemingly relentless winter rain, sodden ground and flooding, making the lockdown bearable for those lucky enough to be able to enjoy time outside. I was struck by the vivid, brilliance of the blue skies and the uncharacteristically pristine air (at least before I was engulfed by smoke wafting up from my neighbour’s barbeque).  Improvements in air quality globally (also seen locally in Reading) resulting from the temporary suppression of economic activity from COVID-19 lockdowns (Figure 1) offer a tantalising glimpse of a cleaner, healthier world possible in a low-carbon future. While human-caused CO2 emissions will temporarily be suppressed there will be no respite from warming of climate – and less particle pollution haze can in fact increase warming as more sunlight reaches the surface.

Figure 1: Reduced column amount of NO2 air pollution over Europe in March/April 2020 compared with 2019 is also a good indicator of changes in harmful particle pollution or aerosols which shade the surface from the sun (Image: © European Space Agency/Copernicus Sentinel-5P satellite data processed by KNMI/ESA)

The effect of longer-term changes in particulate matter (or aerosols) on the atmosphere can already be detected in observations over Europe and China (Figure 2). Measurements show decreasing sunlight reaching the surface (a “dimming”) up to the 1980s as increased particle pollution blocked out the sun. However, cuts in pollution, first in Europe and later in China, can explain a reversal of this trend (a “brightening” at the surface). Aerosols particles act as mirrors by reflecting sunlight back to space but this research also finds changes are linked with how much sunlight is captured (absorbed) by aerosol in the air aloft. These altered heating patterns through the atmosphere and from region to region can further influence rainfall patterns as well as surface temperature.

Figure 2: A schematic depicting measurements of sunlight reaching the surface and intercepted by the atmosphere (compared with 2000-2015 averages shown as horizontal lines) over Europe and China (based on Schwarz et al. 2020 Nature Geosci. Fig. 2).

The extent of regional changes in particle pollution are also evident from satellite imagery since 2010 showing declining pollution in parts of China yet increases across India (Figure 3). And continued rapid cuts to air pollution in Europe and Asia can potentially lead to a short-term acceleration in climate change. Computer simulations show a more rapid increase in European and Asian heatwaves by 2050 associated with sharp cuts in air pollution across parts of Asia. The worst-case-scenario indicates that the hottest day of the year may be up to 4°C hotter by 2050, compared to the present day, with 30-40% of this increase due to air pollution cuts. Research under review also predicts more rapid increases in rainfall during the tropical monsoon seasons in response to dramatic cuts in Asian air pollution. However, it confirms that acceleration of climate change is limited to the next few decades, after which the effect is swamped by the dominant response to greenhouse gas increases. Continuing to emit greenhouse gases into the atmosphere at the current rate will drive far larger and more sustained temperature rises that along with changes in the global water cycle will cause serious impacts for our societies and the ecosystems upon which we depend. This underlines the importance of rapidly reducing greenhouse gas emissions and expanding upon the Paris Agreement targets in the postponed Glasgow COP26 meeting.

Figure 3: Satellite data suggests that aerosol pollution was already declining in China but increasing in India (from Samset et al. 2019 Nature Geosci. Fig. 1)

The recent improvements in air quality and temporary slowing of CO2 emissions resulting from the COVID-19 lockdowns is no silver lining to the awful damage from the pandemic on families and societies. Yet there is an opportunity to re-evaluate what is important for life and society, changing how we work and live to follow a less dangerous and healthier lower carbon path. It may even offer a unique natural laboratory to better understand how pollutant particles affect weather systems and climate. Although reducing particle pollution can temporarily accelerate climate change, moving rapidly to a cleaner, low carbon future by reducing greenhouse gas emissions and other air pollution is essential in avoiding dangerous climate change and improving quality of life. And how will populations cope when the next pandemic strikes if countries are crippled by the effects of climate change?


Allan RP, M Barlow, MP Byrne, A Cherchi, H Douville, HJ Fowler, TY Gan, AG Pendergrass, D Rosenfeld, ALS Swann, LJ Wilcox and O Zolina (2020) Advances in understanding large-scale responses of the water cycle to climate change, Annals of the New York Academy of Sciences, in press, doi: 10.1111/nyas.14337

Luo, F., Wilcox, L. J., Dong, B., Su, Q., Chen, W., Dunstone, N., Li, S. and Gao, Y. (2020) Projected near-term changes of temperature extremes in Europe and China under different aerosol emissions. Environmental Research Letters, 15 (3). 034013. doi: 10.1088/1748-9326/ab6b34

Samset, B. H., Lund, M. T., Bollasina, M. A., Myhre, G. and Wilcox, L. (2019) Emerging Asian aerosol patterns. Nature Geoscience, 12. pp. 582-584. doi: 10.1038/s41561-019-0424-5

Schwarz M, D Folini, S Yang, RP Allan, M Wild (2020) Changes in atmospheric shortwave absorption as important driver of dimming and brightening, Nature Geoscience, 13, 110-115, doi: 10.1038/s41561-019-0528-y

Wilcox, L. J., Liu, Z., Samset, B. H., Hawkins, E., Lund, M. T., Nordling, K., Undorf, S., Bollasina, M., Ekman, A. M. L., Krishnan, S., Merikanto, J. and Turner, A. G. (2020) Accelerated increases in global and Asian summer monsoon precipitation from future aerosol reductions. Atmospheric Chemistry and Physics Discussions. ISSN 1680-7375 doi: 10.5194/acp-2019-1188

Posted in Aerosols, Air quality, Climate, Climate change | Leave a comment

Weather in Cameroon and Adaptation Plan for Climate Change

By: Chimene Daleu

The Republic of Cameroon is a country in central Africa. The country extends from the coast of the Gulf of Guinea and the Atlantic Ocean through west-central Africa to Lake Chad. The majority of Cameroon has a tropical climate with two seasons: the rainy and dry seasons.

Figure 1: Mean monthly precipitation across Cameroon (1961-2001). Plot taken from Ernest Molua, 2006

The geography of Cameroon is highly diverse. The country contains highlands in central and western regions, plains in the north, and tropical forests in the south and along the coast. Its topographic features superimpose climatic variations on this north‐south gradient. The low‐lying coastal plain rises rapidly to the inland regions of high plateaus and mountain ranges. The Cameroon mountain range stretches along the country’s northern border with Nigeria, with peaks in excess of 13,000 ft. This diversity in the geography of Cameroon causes its weather to vary markedly from one region to the other. Overall, the southern regions of Cameroon are generally humid and equatorial, while the climate becomes semi‐arid toward the northern regions. The rainy season is from May to November with most precipitation falling on coastal regions (e.g., Douala, see Figure 1). The wettest months – which should be avoided if you are not a fan of rain – are between July and October (see Figure 1). The climate is pleasant in the dry season between November and February; the “Harmattan”-season between December and February is characterized by the dry and dusty north-easterly trade wind, which blows from the Sahara Desert over West and Central Africa into the Gulf of Guinea.

Figure 2: Mean monthly temperature across Cameroon (1961-2001). Plot taken from Ernest Molua, 2006

The semi‐arid north of Cameroon (Maroua, see Figure 2) is the hottest and driest part of the country. That region experiences average temperatures between 22‐27°C in the cooler seasons (SON, DJF), and 26‐31°C in the warmer seasons (MAM, JJA). On the other hand, temperatures in the southern regions are largely dependent on altitude, ranging between 20-25°C, and varying little with season. Annual rainfall is highest in the coastal and mountainous regions of Cameroon (e.g., Douala; see Figure 1). The driest season is in December, January and February. The main wet season lasts between May and November for most of the country, when the West African Monsoon winds blow from the southwest, bringing moist air from the ocean. Rainfall can exceed  600 mm per month in the wettest regions (e.g., Douala and Kribi; see Figure 1), while in the semi‐arid northern regions of Cameroon the peak monthly rainfall does not exceed 400 mm (e.g., Maroua; see Figure 1). The southern plateau region has two shorter rainy seasons, occurring between May and June and between September and October (e.g., Yaounde).

When it comes to climate change impacts, Cameroon is particularly exposed because of its territories in the Sahelian zone, which are hit hard by desertification, and its territories in coastal areas that are threatened by rising sea levels. Due to the great geographic diversity in Cameroon, the nature of climate change and its impacts vary widely from one region to another. However, all agroecological zones will be affected in one way or another as well as all the sectors. The Cameroonian people must, therefore, face an important challenge, as their economic and social well-being are largely dependent on the viability of the main development sectors.

Cameroon is already facing consequences of climate change, including an abnormal recurrence of extreme weather phenomena such as high temperatures, violent winds, and heavy rainfall, which endanger communities’ ecosystems and the services they provide. The most recent extreme weather phenomena occurred over the first two weeks of October 2019, when heavy rains poured ceaselessly, destroying farms, crops and houses. This left 70,000 homeless in the northern part of Cameroon and 30,000 homeless across the border in Chad, where the Logone River broke its banks on October 1st. It was the worst flooding in northern Cameroon since 2012 when heavy rainfall persisted in the area for over a month and claimed 60 lives.

Photo by Raphael Mwadime

The economy of Cameroon depends strongly on agriculture. Therefore, the effects of global warming and climate change are likely to threaten both the welfare of the population and the economic development. For instance, the aforementioned heavy rainfall and flooding have undermined Cameroon’s efforts to reduce poverty and develop a strong, diversified, and competitive economy. Agricultural policy should, therefore, prepare for changing climate hazards. As a result, the National Adaptation Plan for Climate Change (NAPCC) was created to assist the Cameroonian people in facing such important challenges in accordance with the United Nations Framework Convention on Climate Change (UNFCCC). In 2016, within the framework of strengthening their partnership for effective fundraising for the NAPCC implementation, the Ministry of Environment, Nature Protection and Sustainable Development (MINEPDED) and Global Weather Partnership (GWP) Cameroon initiated the process of the development of a National Investment Plan for Adaptation on Climate Change (NIPACC).

The approach, taken for preparing the NAPCC, was participatory, multidisciplinary, and systematic. The NAPCC was built on 4 strategic axes:

  • improving knowledge on climate change
  • educating the population to climate change adaptation
  • reduce the country’s vulnerability to the impacts of climate change and strengthen its capacity for adaptation and resilience
  • facilitate the coherent integration of climate change adaptation in relevant policies and programs, especially in development planning processes and strategies


Some of this information was found in Wikipedia. Figures 1 and 2 were found in Molua, 2006, and Figure 3 was found online.

Ernest L. Molua (2006) Climatic trends in Cameroon: implications for agricultural management. Clim. Res., 30, 255-262.

Banseka H. and L. C. T. Levesque (2018) Cameroon: Preparing the National Adaptation Plan for Climate Change (NAPCC) and its Investment Strategy. 

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