What do we do with weather forecasts?

By: Peter Clark

As I sat in the Kia Oval in Kennington having taken a day off to watch the first One Day International between England and Pakistan, I had plenty of time to appreciate the accuracy and utility of weather forecasts. The afternoon proved to be a microcosm of both the successes of modern weather forecasting and issues surrounding the use of forecasts in more serious applications (though I may well join in with the cries of “there’s nothing more serious than Cricket!”).

Video  The view from the OCS stand and about 17:00 BST on 8th May 2019 as an intense hail shaft stopped play.

First question: to go to the match or not? When we bought the tickets 6 months ahead, we just had climatology to go on. Early May is a risk, but not very different from later in the season. By the time forecasts become available the question is then “is it worth turning up?” By the Friday five days before, there was a very strong consensus amongst computer forecasts that a cyclone would be tracking across England on the day, most likely during the first half of the day. In fact, the Met Office’s ‘deterministic’ forecast proved very accurate, with the continuous heavy rain passing through London by midday. However, behind the surface front close to the cyclone centre, cold air aloft was overrunning warmer air at the surface, which was given an additional boost as it came from the Atlantic and passed over land.  Warm (and moist) air beneath colder air leads to the likelihood of dreaded convective showers in the afternoon!

There have been real ‘revolutions’ in forecasting over the last few decades. At the centre lies the combination of vast improvements to computer power, more accurate computer models, vast increases in observations to ‘correct’ the data in the models, and development of much more powerful methods to use (or ‘assimilate’) those observations. An extratropical cyclone, or ‘low-pressure system’, is relatively large and long-lived. In this case, the system was at the small end of the scale and quite intense, roughly the size of England – say 500 km across with a life cycle of at least a day. 30 years ago, our computer models had to represent these systems with a grid of points not much better than 100 km apart (see the Met Office’s history of NWP, for example). Today our forecast models have little problem actually representing a cyclone. In practice, they are often predicted in forecast models even before there’s any clear sign of them in observations. While there will still be uncertainty in track and intensity, on the whole they are astonishingly well forecast several days ahead.

Here lies the problem. Showers are much smaller, say 10 km across with the core less than 1 km, and have a lifetime of an hour or so. These cannot even be directly represented in our global models. The most recent ‘revolution’ in forecasting has been the development of so-called ‘convection-permitting’ models (Clark et al. 2016). Regional models (with a grid spacing around 2 km) at last can represent showers, but not well. Something resembling showers can form and give us some very useful guidance on the probability that we’ll be affected by a shower. Such models are now helping produce more accurate flood forecasts, especially for smaller, faster reacting catchments (Dance et al. 2019). Within the ParaCon project we are working hard to find ways to improve the models.

Figure: Radar estimates of the surface rainfall rate at 17:00, 18:00 and 19:00 BST with inset showing the hail storm that hit the Kia Oval at 17:00 BST. (Courtesy of the Met Office). Showers are triggered along a ‘peninsular convergence’ line extending from Cornwall all the way to London that is present for several hours. Clearly, much depended on whether one was beneath or to one side of this.

The message was the same in the morning before the game. As the rain from the cyclone cleared, a high probability of seeing one or two showers or even thunderstorms during the afternoon – which is precisely what happened. We had a couple of flurries of not very intense rain, which did little to interrupt play, plus two hail storms; pea-sized hail fairly typical of a British summer shower. Each lasted about 5 minutes. The inset in the figure shows the hail storm that hit the oval around 17:00 BST. A mere speck on the scale of England, but locally extremely intense. A perfect forecast! However, a computer model run even a couple of hours before could not predict the precise shower hitting our precise location.

What more could we do? I spent the afternoon trying to look at the Met Office’s weather radar composites on my phone. A new rainfall picture is produced every 5 minutes. On the intermittent occasions when I could access data, the showers were very clearly tracked; interestingly they were forming along a broad ‘peninsular convergence’ line that could be tracked back to Land’s End. Along this line, air coming from either side of the south west peninsula meets and so is forced upwards, triggering showers (Golding et al, 2005). This is shown in the three radar images in the figure. Each is an hour apart, but this convergence line is very persistent. These lines were the topic of the COPE field campaign in 2013 (Leon, et al. 2016). This organisation by topography radically changed the overall predictability of the showers. The sharp-eyed reader might also notice an arc of showers moving east from central England into East Anglia, and it is probably no coincidence that the heaviest storm happened where this met the convergence line. Nevertheless, as we sat on the edge of this line, the best we could hope for several hours ahead was a realistic assessment of the probability of having a shower.

This example illustrates very well that the weather forecast is not the only piece in the jigsaw. First, and foremost, there is the investment in resilience; the Oval ground is very well prepared and drained, but there is a limit to what it can cope with. Similarly, investment in flood defences is often controversial, and the Environment Agency have recently announced that climate change is forcing a ‘new approach to flood and coastal resilience’ that may mean not investing in flood defences in some regions.

Second, there is preparedness. The available forecasts had prepared us well for the likelihood of showers. We equipped ourselves as well as we could. I kept a ‘weather eye’ on the radar, at least as far as technology allowed me. I could see the hail storms coming. In this case, the covers were deployed fast enough to protect the pitch and run-ups. Use of forecasts could enable the deployment of defences that take longer to deploy but ultimately save playing time. Currently, forecasts are used by the authorities to help emergency services prepare for likely (but rarely certain) flooding. How best to educate and prepare users including the public to respond to forecasts is one of the leading questions driving research, for example the World Meteorological Organisation’s ‘HIWeather Project’, which recognises the key importance of “better understanding by social scientists of the challenges to achieving effective use of forecasts and warnings” (HIWeather Impact plan). A key part of this is understanding the inevitability of false alarms. We have to be prepared to see play stopped because a forecast (in this case with a very short lead time) says there is a probability of a heavy shower. The price for not being pre-emptive may be the abandonment of the match. Which happened two and a half hours after the rain and hail stopped.

The modern challenge of forecasting is not just to improve the forecast (which may be an exercise in diminishing returns) but also to find ways to make sure that systems are in place to make full use of them and users are well-prepared to take action and understand the actions of others.

References:

Golding, B.W., Clark, P.A. and May, B., 2005, The Boscastle Flood: Meteorological Analysis of the Conditions Leading to Flooding on 16 August 2004, Weather60, 230-235,

Clark, P., Roberts, N., Lean, H., Ballard, S. P. and Charlton-Perez, C., 2016: Convection-permitting models: a step-change in rainfall forecasting. Meteorological Applications, 23 (2). 165-181. ISSN 1469-8080 doi: https://doi.org/10.1002/met.1538

Dance, S. L., Ballard, S. P., Bannister, R. N.Clark, P.Cloke, H. L., Darlington, T., Flack, D. L. A.Gray, S. L., Hawkness-Smith, L., Husnoo, N., Illingworth, A. J., Kelly, G. A., Lean, H. W., Li, D., Nichols, N. K.Nicol, J. C., Oxley, A., Plant, R. S., Roberts, N. M., Roulstone, I., Simonin, D., Thompson, R. J. and Waller, J. A., 2019: Improvements in forecasting intense rainfall: results from the FRANC (forecasting rainfall exploiting new data assimilation techniques and novel observations of convection) project. Atmosphere, 10 (3). 125. ISSN 2073-4433 doi: https://doi.org/10.3390/atmos10030125

Leon, D. C., French, J. R., Lasher-Trapp, S., Blyth, A. M., Abel, S. J., Ballard, S., Barrett, A., Bennett, L. J., Bower, K., Brooks, B., Brown, P., Charlton-Perez, C., Choularton, T., Clark, P., Collier, C., Crosier, J., Cui, Z., Dey, S., Dufton, D., Eagle, C., Flynn, M. J., Gallagher, M., Halliwell, C., Hanley, K., Hawkness-Smith, L., Huang, Y., Kelly, G., Kitchen, M., Korolev, A., Lean, H., Liu, Z., Marsham, J., Moser, D., Nicol, J., Norton, E. G., Plummer, D., Price, J., Ricketts, H., Roberts, N., Rosenberg, P. D., Simonin, D., Taylor, J. W., Warren, R., Williams, P. I. and Young, G., 2016: The COnvective Precipitation Experiment (COPE): investigating the origins of heavy precipitation in the southwestern UK. Bulletin of the American Meteorological Society, 97 (6). 1003-1020. ISSN 1520-0477 doi: https://doi.org/10.1175/BAMS-D-14-00157.1

Posted in Climate, Predictability, Weather, Weather forecasting | Leave a comment

Rescuing the Weather

By: Ed Hawkins

Over the past 12 months, thousands of volunteer ‘citizen scientists’ have been helping climate scientists rescue millions of lost weather observations. Why?

Figure 1: Data from Leighton Park School in Reading from February 1903.

If we are to inform decisions about adapting to a changing climate we need to better understand the risk from extreme weather events, and whether this risk is changing. This requires long and detailed records of the weather. In the UK we are fortunate that meteorologists have recorded the weather across the country for over 150 years. However, most of their observations are still only available as the original paper copies, stored in large archives (Figure 1).

Currently, the only way to transform these observations into useful data is to manually transcribe them from paper to computer. This is an enormous task and would be much easier if it was performed by thousands of people, rather than just a single PhD student.

The WeatherRescue.org website has been set up to enable anyone to help. The first phase of the project recovered 1.5 million observations that were taken on the summit of Ben Nevis and in the nearby town of Fort William between 1883 and 1904. The volunteers then transcribed 1.8 million observations from more than 50 locations across Europe taken between 1900 and 1910. They are now digitising observations taken in the 1860s and 1870s.

So, what can we do with all this data?

Figure 2: Map of pressure observations in the ISPD database for 27th February 1903, including from ships (yellow), with newly rescued data (black) and locations where we have images of the observation logbooks, but the data has not yet been digitised (red).

As a case study, there was a very intense storm on February 26th-27th 1903 which hit Ireland and northern England, uprooting thousands of trees, causing significant structural damage and several fatalities. Hundreds of pressure observations taken across the UK during this storm are not in our digital climate databases. Figure 2 shows the existing data (yellow), newly rescued data (black) and potential data still waiting to be rescued (red) for the period of the intense storm.

Figure 3: The 26th-27th February 1903 storm in the 20th Century Reanalysis (left) and an estimate of how it would look with the new observations (right). The black contours are isobars, and the green shading shows confidence in their position.

The new data allows us to better reconstruct the path and intensity of the storm. Figure 3 shows how the storm appears in the new 20th Century Reanalysis (left) – it is too weak to cause the damage that we know occurred, and the image appears fuzzy because there is much uncertainty about the storm’s location. The right-hand panel shows how the storm should appear with the newly rescued observations (black dots in figure above) – more intense and more certain, with strong winds over eastern Ireland and northern England where the damage occurred. The minimum central pressure is now simulated to be around 955mb.

Severe windstorms are relatively rare but cause significant damage. We need to learn as much about them as possible which means delving back into the past. Thousands of volunteers are helping us determine how the weather changed hour-by-hour over a century ago and to learn about such extreme events. Anyone can help at WeatherRescue.org.

 

Posted in Climate, data assimilation, Data processing, Historical climatology, Outreach, Weather | Leave a comment

Mapping bio-UV products from space

By: Michael Taylor

Solar radiation arriving at the Earth’s surface in the UV part of the spectrum modulates photosynthetically-sensitive life on the land and in the oceans. UV radiation also drives important chemical reaction pathways in the atmosphere that impact air quality. It can cause DNA-damage in the epithelial cells of our skin and is a key factor for tuning the rate of Vitamin D production in our metabolism.

Solar UV radiation may be measured in radiometric units or spectrally-weighted to account for biologically-effective UV radiation doses. The Commission Internationale de l’Éclairage (CIE) defines the reference action spectrum for the ability of UV radiation as a function of wavelength to produce just perceptible erythema (colour change from the Greek word “ερυθρός” for red) in human skin. The standard erythemal dose (Jm-2) is equivalent to an erythemal radiant exposure of 100 Jm-2 (ISO 17166:1999). According to the Bunsen-Roscoe law of reciprocity (Bunsen & Roscoe, 1859), a given biological effect due to UV radiant exposure is directly proportional to the total energy dose given by the product of irradiance (Wm-2) and exposure time (s).

Figure 1: TEMIS erythemal UV dose products (kJ m-2) from KNMI/ESA (Van Geffen et al, 2017): daily “Clear sky” UV from SCIAMACHY/GOME-2, daily “cloud-modified” UV from SCIAMACHY/GOME-2 revealing the impact of a weather system over Sicily, and the global climatological “clear sky” June mean from GOME showing the impact of desert dust as revealed (lower right) by the global climatology of aerosol mixtures (Taylor et al, 2015).

Satellites like GOME, GOME-2 and SCIAMACHY have operational processing algorithms that retrieve erythemal UV dose (kJ m-2) once daily from top of the atmosphere irradiance measurements which are strongly affected by both cloud and atmospheric aerosol (Fig. 1).

Recent studies performed in the context of solar energy (Kosmopoulos et al., 2017; 2018), have revealed that atmospheric aerosol, and desert dust in particular, strongly attenuates solar radiation and the UV component arriving at the ground. In the context of increasing our global capacity for renewable energy with solar power as a major component, this is important. Since atmospheric aerosols reduce solar radiation by absorbing and scattering light and reduce the strength of the direct beam from which solar power generation is most efficient, they also cause forecast uncertainty. As a result, electricity supplied to national grids from solar power must balance the demand by coping with these unexpected fluctuations.

Figure 2: (a) Window functions used by KNMI/ESA to derive Bio-UV dose products from UV spectral irradiances and (b) the back-propagation neural network used to perform ground-based validation. See Zempila et al (2017) for details.

Nevertheless, it is straightforward to obtain biological ultraviolet (“Bio-UV”) products from the UV spectra retrieved at the ground. By applying weighting functions to the ultraviolet part of the irradiance spectra over the range 285-400 nm, important Bio-UV products like Vitamin D dose, DNA-damage dose and photosynthetically active radiation can be calculated. Satellite Bio-UV products from TEMIS (KNMI/ESA) have been successfully validated with high temporal resolution (1-minute) ground-based measurements by Zempila et al. (2017). Fig. 2 shows how the weighting functions vary with wavelength together with the neural network model developed to convert combinations of UV irradiances (I) and solar zenith angle (sza) to Bio-UV products for the ground-based validation.

Figure 3: Zoom sequence showing how the surface solar radiance spectra (280-2500 nm) retrieved under cloudy conditions from space with fast neural network radiative transfer solvers (Taylor et al., 2016) can be used to extract erythemal UV spectral irradiances for calculation of Bio-UV products as per Zempila et al., (2017).

While polar orbiting satellites like GOME, GOME-2, SCIAMACHY and OMI allow global maps of Bio-UV products to be generated, geostationary satellites like Meteosat Second Generation (MSG) provide high spatial resolution images of the Earth disc (3 km x 3 km) every few minutes and allow us to dramatically increase the frequency of the data. In support of this, an operational algorithm capable of retrieving the UV part of the solar spectrum at the surface was recently developed (Taylor et al., 2016). This was achieved with a synergistic model that uses both machine learning with neural networks and a look-up table of radiative transfer simulations to help unravel the complexity of the atmosphere. The model includes the effects of clouds, aerosols, ozone, elevation and surface albedo (Taylor et al; 2016; Kosmopoulos et al., 2017) and provides the surface global horizontal irradiance (GHI), direct normal irradiance (DNI) and diffuse horizontal irradiance (DHI) spectrum over the broad wavelength range 285-2700 nm. Application of the weighting functions of Fig. 2 to the UV part of the solar radiation spectrum can then provide global maps of Bio-UV products at high frequency. Fig. 3 illustrates how various UV products can be obtained from surface solar radiation spectra retrieved from space.

One of the most exci­ti­ng appl­icati­ons of being able to map UV spectral information ­from space is the potential for creating mob­ile appli­cat­ions that pull data from surface UV spectra data cloud computing resources and combi­ne them wi­th users’ GPS i­nformati­on to produce real-ti­me UV alerts to the general publi­c with unprecedented precision. By improving our capacity to map UV impact on the quality of life in the global ecosystem from space at high frequency, we will be better placed to monitor progress towards achievement of the UN’s sustainable development goals as we proceed to a more climate resilient society.

References

Bunsen, R., Roscoe, H. E., 1859: Photochemische untersuchungen. Annalen der Physik  184(10), 193-273, DOI: 10.1002/andp.18591841002

Kosmopoulos, P., S. Kazadzis, H. El-Askary, M. Taylor, A. Gkikas, E. Proestakis, C. Kontoes, M. M. El-Khayat, 2018: Earth-Observation-Based Estimation and Forecasting of Particulate Matter Impact on Solar Energy in Egypt. Remote Sens. 10(12), 1870, DOI: https://doi.org/10.3390/rs10121870

Kosmopoulos, P. G., S. Kazadzis, M. Taylor, E. Athanasopoulou, O. Speyer, P. I. Raptis, E. Marinou, E. Proestakis, S. Solomos, E. Gerasopoulos, V. Amiridis, 2017: Dust impact on surface solar irradiance assessed with model simulations, satellite observations and ground-based measurements. Atmos. Meas. Tech., 10(7), 2435-2453, DOI: 10.5194/amt-10-2435-2017

Taylor, M., P. G. Kosmopoulos, S. Kazadzis, I. Keramitsoglou, C. T. Kiranoudis, 2016: Neural network radiative transfer solvers for the generation of high resolution solar irradiance spectra parameterized by cloud and aerosol parameters. J. Quant. Spectrosc. Radiat. Transfer, 168, 176–192, DOI: 10.1016/j.jqsrt.2015.08.018

Taylor, M., S. Kazadzis, V. Amiridis, R. A. Kahn, 2015: Global aerosol mixtures and their multiyear and seasonal characteristics. Atmos. Environ., 116, 112–129, DOI: 10.1016/j.atmosenv.2015.06.029

Van Geffen, J., Van Weele, M., Allaart, M. and Van der A, R., 2017: TEMIS UV index and UV dose operational data products, version 2, KNMI Dataset, DOI:
10.21944/temis-uv-oper-v2

Zempila, M. M., J. H. van Geffen, M. Taylor, I. Fountoulakis, M. E. Koukouli, M. van Weele, R. J. van der A, A. Bais, C. Meleti, D. Balis, 2017: TEMIS UV product validation using NILU-UV ground-based measurements in Thessaloniki, Greece. Atmos. Chem. Phys., 17(11), 7157–7174, DOI: 10.5194/acp-17-7157-2017

Acknowledgements

I am very grateful to colleagues from the Tropospheric Emission Monitoring Internet Service (TEMIS) at KNMI and ESA for kindly making available plots of UV radiation monitoring products generated from the v2 processing algorithm, and to colleagues from the Greek national network for the measurement of ultraviolet solar radiation (uvnet.gr) for permission to present results from Zempila et al., (2017) using their ground-based NILU-UV multi-filter radiometer measurement data and associated UV dose data obtained from a Brewer MKIII spectrophotometer. I would also like to acknowledge colleagues from solea.gr with whom I collaborated with to develop the solar radiation neural network modeling aspects presented.

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

The sky is the limit – How tall buildings affect wind and air quality

By: Denise Hertwig

Based on current UN estimates, by 2050 over 6.6 billion people (68% of the total population) will be living in cities. Across the world, tall (> 50 m height) and super-tall (> 300 m) buildings already define the skylines of many large cities and will become increasingly more common outside of city centres to accommodate growing urban populations, especially when horizontal urban sprawl is geographically limited. For London, 2019 was declared the “year of the tall building” (NLA London Tall Buildings Survey 2019). At the moment, 541 buildings over 20 storeys (approx. 60 m in height) are planned or already under construction in the UK capital. Tall buildings are currently being built in 22 out of the 33 London boroughs and 76 of them are expected to be completed this year.

Tall buildings, in isolation or as clusters, affect the urban micro-climate of the local surroundings and the neighbouring region. The impact on aerodynamics (e.g. local flow distortions, long-range wake effects), radiation budget (e.g. shadowing, radiative trapping) and components of the surface energy balance (e.g. storage of heat in building materials, anthropogenic heat emissions) can be large compared to low-rise buildings. Such modifications challenge current modelling frameworks for urban areas. Urban land-surface models used in numerical weather prediction, for example, typically do not account for building-height variations. They also rely on the concept that the flow within the urban canopy is sufficiently decoupled from the flow aloft, which is not the case if tall buildings protrude deep into the urban boundary layer.

Figure 1: Normalised pollutant concentrations (a,b) in an idealised building array. Pollutants are released from point sources located (a) in the street canyon behind a tall building, (b) in an intersection upwind of the tall building. (c) Mean-flow streamlines near the tall building with colours showing the mean vertical velocity. The black arrow indicates the upwind flow direction. Data are results from large-eddy simulations by Fuka et al. (2018) for the DIPLOS project.

Similarly, operational urban air quality and dispersion models do not usually account for tall-building effects (Hertwig et al. 2018). Tall buildings strongly change pedestrian-level winds in the surrounding streets and the flow field above the roofs of the low-lying buildings. This affects pollutant pathways and the overall ventilation potential of cities. Pollutants released near the ground in a street canyon on the leeward side of a tall building (Fig. 1a) can be rapidly lifted out of the building canopy by updrafts (Fig. 1c). Although the pollutants are emitted at the ground, the tall building causes a large proportion of the released mass to be transported above the roofs of the low-rise neighbourhoods, thereby reducing street-level pollution. A pollutant source located in an upwind intersection leads to drastically different results (Fig. 1b). The downdrafts on the windward side of the tall building result in strong horizontal flow out of the upwind street canyon (Fig. 1c). This outflow shifts the pollutants away from their release point in the intersection, creating a virtual source location in the adjacent street canyon and deteriorating air quality in the streets downwind.

Figure 2: (a) Building heights and (b) wind-tunnel model buildings of the neighbourhood between Waterloo station and Elephant & Castle in London (MAGIC project study area). Wind-tunnel measurements of the wake behind the central tall building (81 m height) in isolation and together with the low-rise building canopy shown in terms of (c) height profiles of flow speeds at several sites downwind of the tall building, (d) velocity differences to the ambient (undisturbed) flow with downwind distance at several heights. Details in Hertwig et al. (2019).

Flow interactions between tall and low-rise buildings also change the structure of the momentum deficit region (wake) that forms behind tall buildings. Wake models used for local air-quality predictions currently do not account for such interactions as they were derived for isolated buildings. Wind-tunnel experiments in a realistic scale model of the area between the Waterloo and Elephant & Castle stations in London (Fig. 2a,b) documented the strong impact of the canopy on tall-building wakes (Hertwig et al. 2019). Compared to tall buildings in isolation, the presence of a low-rise canopy displaces the wake vertically (Fig. 2c), so that flow speeds are reduced over longer distances downwind well-above the canopy (Fig. 2d). In the case shown, the wake extends over distances larger than 5 times the height of the tall building (i.e. > 400 m). The increasing spatial resolution (of the order of 100 m) of mesoscale and microscale atmospheric models means that tall-building wakes no longer are subgrid-scale phenomena, but have an impact at the grid-scale. Understanding and quantifying tall-building impacts on the boundary layer over cities is essential to identify needs for model refinements.

References

Fuka, V., Z.-T. Xie, I.P. Castro, P. Hayden, M. Carpentieri, A.G. Robins, 2018: Scalar fluxes near a tall building in an aligned array of rectangular buildings. Boundary-Layer Meteorology 167, 53–76, DOI: 10.1007/s10546-017-0308-4

Hertwig, D., L. Soulhac, V. Fuka, T. Auerswald, M. Carpentieri, P. Hayden, A. Robins, Z.-T. Xie and O. Coceal, 2018: Evaluation of fast atmospheric dispersion models in a regular street network. Environmental Fluid Mechanics 18, 1007–1044, DOI: 10.1007/s10652-018-9587-7

Hertwig, D., H.L. Gough, S. Grimmond, J.F. Barlow, C.W. Kent, W.E. Lin, A.G. Robins and P. Hayden, 2019: Wake characteristics of tall buildings in a realistic urban canopy. Boundary-Layer Meteorology, DOI: 10.1007/s10546-019-00450-7 (in press)

Posted in Boundary layer, Climate, Urban meteorology | Leave a comment

Balloon measurements at Stromboli suggest radioactivity contributes charge in volcanic plumes

By: Martin Airey

Volcanic lightning is an awe-inspiring and humbling display of nature’s power. It results from the breakdown of large electric fields that are generated within the volcanic plume. The processes that result in the accumulation of charge are varied and complex and by no means fully understood. Current knowledge of the key established mechanisms that are known to contribute to plume charging centre around the role played by ash. These mechanisms fall broadly into the categories of fractoemission and triboelectrification (Mather and Harrison, 2006). Fractoemission is the release of neutral and charged (electrons, positive ions, and photons) particles from fracture surfaces as magma fragments upon eruption (James et al., 2000); these particles may then interact with ash and aerosols to impart a net charge. Triboelectrification is a mechanism by which charge is transferred between the ash particles as they collide.

When charge has been produced, it must then be separated in order for an electric field to develop and a discharge to occur. The plume is a dynamic and chaotic environment, where primitive constituents of the magma, such as solid particles, gases, and metal species are mixed with atmospheric material as it is entrained by the plume. Above the initial jet region, thermal buoyancy-driven dynamics enable the plume to grow to an altitude at which neutral buoyancy is attained. Within this setting, charged aerosols and charged ash grains settle differently resulting in the separation of positively and negatively charged regions in the plume (Mather and Harrison, 2006), which can ultimately cause a discharge to occur.

But what if there are other additional mechanisms that contribute to either the charging or separation processes? As it is a complex, rapidly evolving, multiphase environment, there is the potential for many other chemical and physical interactions occurring within the plume that may currently be overlooked by this simplistic view. To test this, sensors and instrumentation developed at Reading over many years for deployment on weather balloons was combined through a NERC-funded project into a disposable modular payload called VOLCLAB (VOLCano LABoratory). The range of sensors that can be incorporated into the VOLCLAB package includes an optical backscatter droplet detector, a charge sensor, a sulphur dioxide sensor, an oscillating microbalance particle collector, and a turbulence sensor.


View from Stromboli’s summit into the vent complex showing the gas-rich plumes

In September 2017, a team of scientists from the University of Reading, Ludwig Maximillians Universität (Munich), and the University of Bath set off to Stromboli on fieldwork funded by National Geographic, equipped with VOLCLAB sensors, radiosondes, balloons, a thunderstorm detector, and lots of helium. Stromboli was an ideal choice for this expedition as it erupts frequently (several times an hour) and produces a wide range of plume types ranging from ash-rich to predominately gaseous. By launching these instruments directly into the plumes, in situ measurements may be acquired from all these plume types. The two-week long campaign required a daily hike to the summit at 900 m, often with very heavy kit. Many sensor-equipped balloons were launched from the summit with a range of success in encountering a plume, and VOLCLAB packages were deployed in fixed locations around the summit to continually record passing plumes.

   Martin Airey (holding VOLCLAB package) and Corrado Cimarelli

                 Keri Nicoll, Kuang Koh, and Martin Airey

Most interesting was the discovery of significant electric charge in plumes that contained negligible or no ash. This led to the investigation of what might be causing this unexpected charging mechanism. It is widely known that volcanoes emit a broad range of chemical products (Allard et al, 2000), one of which is radon, which is produced in high concentrations from all volcanoes. Radon is routinely monitored at many volcanoes, including Stromboli, which is known to constantly emit very large quantities through the soil near the vents, and even more during eruptions (Cigolini et al, 2009). As radon radioactively decays, it increases the charge present by ionising the air. This additional source of charge, inferred for the first time with these new direct measurements inside gaseous plumes, will inevitably contribute to the overall charge structure and may affect the likelihood of lightning strikes.

The original open access article, published in Geophysical Research Letters, may be found at: https://agupubs.onlinelibrary.wiley.com/doi/full/10.1029/2019GL082211

And additional press can be found in the following links:
Science: https://www.sciencemag.org/news/2019/03/volcanic-lightning-may-be-partially-fed-earth-s-natural-radioactivity
New York Times: https://www.nytimes.com/2019/03/29/science/volcanoes-lightning-radon-gas.html
Atlas Obscura: https://www.atlasobscura.com/articles/mount-stromboli-volcano-science
The VOLCLAB package covered in Meteorological Technology magazine: 
http://viewer.zmags.com/publication/235ac328#/235ac328/56
Some footage of the fieldwork was also included in the Arte documentary “Living with Volcanoes” from around 7 minutes: 
French: https://www.arte.tv/fr/videos/069786-010-A/des-volcans-et-des-hommes-iles-eoliennes/
German: https://www.arte.tv/de/videos/069786-010-A/leben-mit-vulkanen/

References:

Allard, P., Aiuppa, A., Loyer, H., Carrot, F., Gaudry, A., Pinte, G., et al. (2000). Acid gas and metal emission rates during long‐lived basalt degassing at Stromboli volcano. Geophysical Research Letters, 27(8), 1207–1210. https://doi.org/10.1029/1999GL008413

Cigolini, C., Poggi, P., Ripepe, M., Laiolo, M., Ciamberlini, C., Delle Donne, D., et al. (2009). Radon surveys and real‐time monitoring at Stromboli volcano: Influence of soil temperature, atmospheric pressure and tidal forces on 222Rn degassing. Journal of Volcanology and Geothermal Research, 184(3–4), 381–388. https://doi.org/10.1016/j.jvolgeores.2009.04.019

James, M. R., Lane, S. J., & Gilbert, J. S. (2000). Volcanic plume electrification—Experimental investigation of fracture charging mechanism. Journal of Geophysical Research, 105(B7), 16,641–16,649. https://doi.org/10.1029/2000JB900068

Mather, T. A., & Harrison, R. G. (2006). Electrification of volcanic plumes. Surveys in Geophysics, 27(4), 387–432. https://doi.org/10.1007/s10712‐006‐9007‐2

 

Posted in Climate, Convection, Measurements and instrumentation, Volcanoes | Leave a comment

Convective self-aggregation: growing storms in a virtual laboratory

By: Chris Holloway

Figure 1: An example of convective self-aggregation from an RCE simulation using the Met Office Unified Model at 4km grid length with 300 K SST.  Time mean precipitation in mm/day for (a) Day 2 (still scattered), and (b) Day 40 (aggregated).  Note that the lateral boundaries are bi-periodic, so the cluster in (b) is a single organised region.  Adapted from Holloway and Woolnough (2016).

Convective self-aggregation is the clumping together of isolated convective cells (rainstorms) into organised regions in idealised computer simulations.  This storm clustering may not seem all that unusual, but it is surprising because self-aggregation occurs in simulations of “radiative-convective equilibrium” (RCE) in which boundary conditions are homogeneous (sea surface temperature [SST] is constant in space and time), there is no imposed mean wind, and planetary rotation is set to zero (e.g., Figure 1).  In other words, there is no external cause of the clustering of convection in self-aggregation (hence the “self” prefix).  Instead, internal feedbacks such as cloud-radiation interactions and surface-flux feedbacks are key (Wing et al.2017 and references therein).

Figure 2: Satellite estimates of average fractional cover vs total Cold Cloud Area for a given domain-mean precipitation rate (R) range and for ranges of the “SCAI” aggregation index between 0.00 and 0.35 (red, aggregated), between 0.35 and 0.70 (black, intermediate), and between 0.70 and 1.50 (blue, disaggregated); t stands for optical thickness. Shaded regions indicate the 90% confidence interval. (a)–(c) thick anvil, (d)–(f) optically thin anvil.  Adapted from Stein et al. (2017).

While self-aggregation is intellectually interesting, many scientists are sceptical of the relevance of this phenomenon for real weather and climate.  After all, the real world has plenty of inhomogeneity in surface temperature as well as rotation and vertical wind shear.  However, organised tropical convection in real-world observations shows many similarities to self-aggregated convection in idealised simulations: for more aggregated conditions the mean state has lower relative humidity, outgoing longwave radiation (OLR) is larger, and anvil cloud amount is reduced (Holloway et al. 2017 and references therein).  For instance, work at the University of Reading using satellite observations has shown that optically thin anvil cloud cover decreases as convection becomes more aggregated, which could have implications for climate (Figure 2).    More realistic convective-scale simulations of organised tropical convection (with observed SSTs, rotational effects, and wind shear effects) also provide evidence that cloud-radiation feedbacks act to maintain organisation and reduce the mean relative humidity (Holloway 2017).   Other real-world forms of organised tropical convection, including the Madden-Julian Oscillation (MJO), tropical cyclones, and the Intertropical Convergence Zone (ITCZ) all show cloud-radiative feedbacks and moisture-convection feedbacks that resemble processes important for convective self-aggregation in idealised computer simulations.

The potential impact of convective aggregation on climate is an area of active research and debate.  Some idealised computer experiments show stronger self-aggregation with warmer SSTs sea surfaces, but others do not (Wing 2019).  If aggregation were to increase with increasing SST, this would likely be a slightly negative feedback for global warming, meaning it would allow for slightly less warming for a given increase in carbon dioxide concentrations, but this is also an active area of research and debate.  Aggregation tends to be weaker and more variable in simulations that include coupled ocean models (e.g. Hohenegger and Stevens 2016, Coppin and Bony 2017), so this is another area that needs more extensive research. 

Studying convective aggregation enables the scientific community to generate and test hypotheses and isolate mechanisms about fundamental processes that are potentially important for convective organisation, but which can be difficult to disentangle in more realistic settings.  Even if studies eventually demonstrate how self-aggregation is not an adequate framework for some climate problems, this will also be a form of important progress.  The Radiative-Convective Equilibrium Model Intercomparison Project (RCEMIP) is bringing scientific institutions together to compare self-aggregation at different model resolutions, domains and SSTs in order to facilitate further research into this exciting topic.  At Reading, Met Office Unified Model convection-permitting simulations have been performed and submitted to RCEMIP in association with the joint NERC-Met Office ParaCon project which seeks to greatly improve the representation of convection in weather and climate models.   RCEMIP and other research efforts will increasingly apply new concepts emerging from idealised simulations to the complex interactions between convection, moisture, clouds, radiation, surface fluxes, circulations and climate.

References:

Coppin, D., and S. Bony, 2017: Internal variability in a coupled general circulation model in radiative‐convective equilibrium, Geophysical Research Letters, 44, 10, 5142-5149, https://doi.org/10.1002/2017GL073658.

Hohenegger, C., and Stevens, B., 2016: Coupled radiative convective equilibrium simulations with explicit and parameterized convection. J. Adv. Model. Earth Syst., 8, 1468–1482, doi:10.1002/2016MS000666.

Holloway, C. E., 2017: Convective aggregation in realistic convective- scale simulations, J. Adv. Model. Earth Syst., 9, 1450–1472, doi:10.1002/ 2017MS000980.

Holloway, C. E., A. A. Wing, S. Bony, C. Muller, H. Masunaga, T. S. L’Ecuyer, D. D. Turner, and P. Zuidema, 2017: Observing convective aggregation.  Surveys in Geophysics, 38: 1199. doi:10.1007/s10712-017-9419-1. 

Holloway, C. E., and S. J. Woolnough, 2016: The sensitivity of convective aggregation to diabatic processes in idealized radiative-convective equilibrium simulations.  J. Adv. Model. Earth Syst., 8, 166–195, doi:10.1002/2015MS000511.

Stein, T. H. M., C. E. Holloway, I. Tobin, and S. Bony, 2017: Observed relationships between cloud vertical structure and convective aggregation over tropical ocean.  J. Climate, 30, 2187–2207. 

Wing, A. A., 2019: Self-Aggregation of Deep Convection and its Implications for Climate. Curr. Clim. Change Rep., 5: 1. https://doi.org/10.1007/s40641-019-00120-3.

Wing, A. A., K. Emanuel, C. E. Holloway, and C. Muller, 2017: Convective self-aggregation in numerical simulations: A review.  Surveys in Geophysics, 38: 1173. https://doi.org/10.1007/s10712-017-9408-4.

 

 

Posted in Climate, Climate modelling, Numerical modelling, Tropical convection | Leave a comment

Modelling Ice Sheets in the global Earth System

By: Robin Smith

As Till wrote recently, our national flagship climate model (UKESM1, the UK Earth System Model) has been officially released for the community to use, after more than six years in development by a team drawn from across the NERC research centres and the Met Office. The most unique capability of the UKESM effort isn’t included in that release however: UKESM1 can also be made to interactively simulate the evolution of the massive ice sheets of Greenland and Antarctica.

On millennial timescales, the growth and decay of ice sheets play one of the most fundamental roles in determining the climate of the Earth – think of the ice-age cycles of the last million years. But ice sheets aren’t just for the paleoclimate people. Loss of mass from ice sheets accounts for around a third of the currently observed global mean sea level rise and their contribution is expected to increase and dominate the sea level budget in the coming decades and centuries (Church et al. 2013). The climate change impact from ice sheets isn’t limited to sea level rise either, with ice melt input to the ocean linked to a range of wide climate change problems (Golledge et al. 2019).

Back near the start of this project I wrote in this blog about plans for the ice sheets in UKESM and some of the challenges that we were facing. Since then we’ve implemented a system of online climate downscaling over ice sheets in the Met Office Unified Model, taught the NEMO (Nucleus for European Modelling of the Ocean) ocean model to move its boundaries (a little) as it runs without becoming unstable and built a whole framework of Python code to transfer fields between the BISICLES (Berkley Ice Sheet Initiative for Climate Extremes) ice sheet model and the rest of the climate system. The whole system adds additional layers of complexity to what is already one of the most sophisticated climate models in the world. It’s all still a bit rough around the edges, which is why this isn’t included in the main UKESM1 release, but there is finally a functioning coupled climate-ice model that can do things that no other state of the art climate model can do.

 Figure 1: Surface Mass Balance (SMB) (the balance between accumulating and melting snow) estimated for Greenland in UKESM (blue and black lines) compared with the output from a specialised regional climate model (red line, Noel et al. 2018). UKESM captures the observed downturn in SMB at the end of the 1990s, often linked with decadal variability in the North Atlantic. The coupled ice sheet component additionally models the dynamic flow and calving rate of the ice sheet to give a complete estimate of how the ice mass will evolve.

So, has it all been worth it? What are we going to do with this model now we’ve made it? Whilst there are many open questions around the stability of the large ice sheets and how they interact with the climate around them, our first goal will be completion of a set of coupled climate—ice simulations for the Ice Sheet Model Intercomparison ISMIP6, an international model intercomparison that will provide projections of 21st century ice sheet mass loss. Early results from UKESM with our new online downscaling for the ice compare very well to regional climate model results (Figure 1) and suggest that additional surface melt of Greenland alone could be adding another millimetre to global-mean sea level every year by 2050.

Figure 2: Pine Island Glacier on Antarctica is observed to be thinning and retreating rapidly, most likely in response to ocean warming underneath its ice shelf. With the interactive ice in  UKESM we can model how the ocean melts the shelf away (top right), the flow of the glacier supplying new ice to the region (top left) and predict the overall rate of retreat. These are early results but show promise.

The most exciting new science we want to tackle with the ice in UKESM1 sits at the other end of the globe, however. There are many theories about how the floating ice shelves that fringe Antarctica respond to changes in ocean conditions, and how the flow of the grounded ice upstream will respond, but the uncertainties are enormous. Estimates of the resulting contribution to global mean sea level rise at 2100 range from centimetres to metres (see Edwards et al. 2019 for a recent perspective). This is an inherently coupled problem whose physics simply cannot be understood by modelling any one part of the system in isolation. Like all complex problems, it’s also going to be very hard – there are poorly-observed, crucial details in each component that can significantly alter the final outcomes – and we’re not pretending that one model is going to lead us straight to the answers. For one thing, there are long-standing biases in the climate simulation of the Met Office models in the high southern latitudes that will need to be improved before we are really simulating the processes against the right background.  But with UKESM1 we’re now getting our hands on some tools that can start to see the coupled atmosphere-ocean-ice physics evolving together for the first time (Figure 2), and that’s a very promising development.

References:

Church, J.A., and Coauthors, 2013: Sea Level Change. Climate Change 2013: The Physical Science Basis. Contribution of Working Group I to the Fifth Assessment Report of the Intergovernmental Panel on Climate Change, Stocker, T.F., D. Qin, G.-K. Plattner, M. Tignor, S.K. Allen, J. Boschung,  A. Nauels, Y. Xia, V. Bex and P.M. Midgley, Eds. Cambridge University Press, 1137-1216

Edwards, T.L., M.A. Brandon, G. Durand, N.R. Edwards, N.R. Golledge, P.B. Holden, I.J. Nias, A.J. Payne, C. Ritz & A. Wernecke, 2019: Revisiting Antarctic ice loss due to marine ice-cliff instability. Nature, 566, 58-64, https://doi.org/10.1038/s41586-019-0901-4

Golledge, N.R., E.D. Keller, N. Gomez, K.A. Naughten, J. Bernales, L.D. Trusel & T.L. Edwards, 2019: Global environmental consequences of twenty-first-century ice-sheet melt. Nature, 566, 65-72, https://doi.org/10.1038/s41586-019-0889-9

Noël, B. and Coauthors, 2018: Modelling the climate and surface mass balance of polar ice sheets using RACMO2 – Part 1: Greenland (1958–2016). The Cryosphere, 12, 811-831, https://doi.org/10.5194/tc-12-811-2018

Posted in antarctica, Arctic, Climate, Climate modelling, Cryosphere, Numerical modelling | Leave a comment

The Boundary Layer and Submesoscale Motions

By: Alan Grant

Science is an exciting career, although what you may consider to be exciting will depend on your field. Sometimes things get most exciting when what initially appears to be a frustrating problem turns into an interesting problem.

Our story starts with the decision to develop a new parametrization of the ocean surface boundary layer (OSBL). The main reason for this was to represent what is known as Langmuir turbulence (McWilliams et al, 1997), and as part of this, the depth of the base of the OSBL in the scheme would be determined using a prognostic equation. Simple models of the entraining boundary layer using a prognostic equation for the thickness of the boundary layer have been around for many years. However, a significant problem with the new scheme was to work out how to use a prognostic equation in a parametrization destined for use in a finite difference model. Eventually the new scheme was ready, and it was time to test it, initially using a one-dimensional model.

This scheme was developed as part of NERC’s Ocean Surface Mixing, Ocean Submesoscale Instabilities (OSMOSIS) project. As part of OSMOSIS, gliders were operated for a year in the area of the Porcupine Abyssal Plain (PAP, named after HMS Porcupine) observatory, off the continental shelf, to the south west of Ireland The scheme was assessed against this year of data.

Figure 1 : Comparison between mixed layer depth determined from the gliders (symbols) and the boundary layer depth from the model (red curve) during the Autumn. The start date of the simulation is 23rd September.

The first test looked at the deepening of the boundary layer which occurs during the autumn. Figure 1 shows that there is good agreement between the boundary layer depth from the model and the mixed-layer depth obtained from the gliders. Over the seventy days of the simulation the thickness of the boundary layer increases from about 50m to 150m, due to entrainment.

Figure 2 : Same as Figure 1 but for the winter. The start date of the simulation is 25th December

Encouraged by these results, we moved on to look at the performance of the model during the winter. Figure 2 shows the results of this comparison, and the reason for the frustration should be clear – during the winter the base of the boundary layer in the model gets much deeper than the mixed layer depth from the gliders. This deepening of the model boundary layer is due to entrainment, and there is little scope to simply tune the scheme to improve agreement, particularly given how well the scheme performs during the autumn.

This is where submesoscale motions enter in to our story. Submesoscale motions have scales between 100m to 10km, are generally associated with baroclinic instabilities, and are thought to restratify the upper ocean. While parametrizations of submesoscale motions are still crude, one developed by Fox-Kemper et al. (2008) is of particular interest. In this parametrization, the effect of the submesoscale motions is formulated in terms of a buoyancy flux. This makes it possible to directly couple the submesoscale parametrization to the equation for the thickness of the boundary layer in the new boundary layer scheme.

And so, our frustrating problem becomes an interesting problem, since we now need to think about how to couple parametrizations of two distinct processes. Preliminary tests of one way of doing this coupling are encouraging.

I forgot to mention, the new boundary layer parametrization has been implemented in the ocean model Nucleus for European Modelling of the Ocean (NEMO) where it gives overly deep boundary layers during the winter. The hope is that submesoscale motions will provide the solution to this problem.

References :

Fox-Kemper B., R. Ferrari, and R.W.Hallberg, 2008 : Parameterization of mixed layer eddies. I. Theory and diagnosis. J. Phys. Ocean., 38, 1145–1165, doi:10.1175/2007JPO3792.1

McWilliams, J.C., 2016: Submesoscale currents in the ocean. Proc. R. Soc. A., 472, http://doi.org/10.1098/rspa.2016.0117

McWilliams, J. C., P. P. Sullivan, and C.-H. Moeng, 1997: Langmuir turbulence in the ocean. J. Fluid Mech.334, 1-30, https://doi.org/10.1017/S0022112096004375

Posted in Boundary layer, Climate, Numerical modelling, Oceans | Leave a comment

Multi-fluids Modelling of Convection

By: Hilary Weller

Atmospheric convection – the dynamics behind clouds and precipitation – is one of the biggest challenges of weather and climate modelling. Convection is the driver of atmospheric circulation, but most clouds are smaller than the grid size and so cannot be represented accurately or at all. Consequently, all but the highest resolution models use convection parameterisations – statistical representations which estimate the heat and moisture that would be produced and transported by clouds if they were resolved. These parameterisations have become sophisticated, estimating the mass that is transported upwards and how this will influence the momentum, temperature, moisture and precipitation. Without these parameterisations climate models dramatically fail. However, there are still big problems with these parameterisations; they tend to produce unrealistic hot columns of air and the main regions of convection in the tropics are usually misplaced (see Figure 1). These are large heat sources for the global atmosphere and so errors in these locations have knock-on effects across the globe.

Figure 1: Fig 11 from Neale et al, (2013) showing “Annually averaged (a) observed precipitation from GPCP (1979–2003) and model precipitation biases (mm day 21 ) in AMIP-type experiments for (b) CAM4 at 18, (c) CAM4 at 28, and (d) CAM3 at T85, and in fully coupled experiments for (e) CCSM4 at 18, (f) CCSM4 at 28, and (g) CCSM3 at T85.”

There are two assumptions made in convection parameterisations that could be to blame for their poor performance. Firstly, there is no memory of the properties of convection from one time-step to the next. The convection properties are calculated each time step from scratch as if there had been no convection in the previous time step. Secondly, although there is a class of convection scheme called “mass flux”, convection schemes do not actually transport air upwards. They transport the heat, moisture and momentum but the distribution of mass in the vertical is not changed by the convection scheme. Removing these assumptions is quite tricky. You realise that what you need to do is solve the same equations of motion in clouds and outside clouds, but these need to be modelled separately because the clouds are such a small fraction of the total volume. This is the multi-fluid approach. Separate equations for velocity, temperature, moisture, and volume fraction are solved for the air in clouds and the environmental air, outside clouds. As the clouds and the environment are interwoven, we assume that they share the same pressure.

John Thuburn and I are working on making this approach work as part of the NERC/Met Office Paracon project to develop big changes to the way that convection is parameterised in order to remove some of the less realistic assumptions (Thuburn et al, 2018, Weller and McIntyre, 2019). As ever, it is proving more difficult than we expected. We knew that we would need to solve simultaneous equations for the properties in and outside the clouds and that these would need to be simultaneous with the single pressure. I naively thought that this would be sufficient and that my expertise in this area would make it possible whereas convection modellers are usually less familiar with this simultaneous solution procedure. However, it turns out that the multi-fluid equations can be unstable. They are easy to stabilise but the stabilisation can have the effect of making the two fluids behave as one which defeats the purpose. We need to make the two fluids move through each other. John has made some good progress on this (Thuburn et al, 2019).

The multi-fluid equations need transfer terms to transfer air in and out of the clouds. These are not a mystery. Traditional parameterisations predict these transfers, and these have been thoroughly validated and tested. However, the multi-fluid equations are sensitive to these transfer terms and do not behave in the same way as the traditional parameterisations. We also want to base the transfers on the sub-grid scale variability both inside and outside the cloud so that a cloud is formed if some of the air in a grid box is ready to rise and condense out water. There is plenty to do.

While the multi-fluids team from the Paracon project have been working on simultaneous solutions of equations for in and outside clouds, ECMWF thought that they would try a more direct approach – simply adding a term to the continuity equation based on the mass flux predicted by their traditional parameterisation (Malardel and Bechtold, 2019). This was also the approach taken by Kuell and Bott (2008). The problem with this approach is that it will be unstable if a large fraction of a grid box is cloudy. However, ECMWF and Kuell and Bott (2008) have not reported any stability problems, although the ECMWF approach did ensure that only a small fraction of each grid box is transported by the convection scheme. The results so far seem promising. However, to be able to increase the resolution and run with sufficiently long time steps so that the model is competitive, we will need the multi-fluid approach.

References:

Kuell, V., and A. Bott, 2008: A hybrid convection scheme for use in non-hydrostatic numerical weather prediction models. Meteorol. Z. 17 (6), 775-783, https://doi.org/10.1127/0941-2948/2008/0342

Malardel, S. and Bechtold, P. (2019), The coupling of deep convection with the resolved flow via the divergence of mass flux in the IFS. Q J R Meteorol Soc. https://doi:10.1002/qj.3528

Neale R.B., J. Richter, S. Park, P.H. Lauritzen, S.J. Vavrus, P.J. Rasch, M. Zhang, 2013: The Mean Climate of the Community Atmosphere Model (CAM4) in Forced SST and Fully Coupled Experiments. J. Climate., 26, 5150-5168, https://doi.org/10.1175/JCLI-D-12-00236.1

Thuburn, J., G.A. Efstathiou, R.J. Beare, 2019: A two‐fluid single‐column model of the dry, shear‐free, convective boundary layer. Quart. J. Roy. Meteor. Soc., https://doi.org/10.1002/qj.3510

Thuburn, J., H. Weller, G.K. Vallis, R.J. Beare, and M. Whitall, 2018: A framework for convection and boundary layer parameterization derived from conditional filtering.  J. Atmos. Sci., 75 (3), 965-981, https://doi.org/10.1175/JAS-D-17-0130.1

Weller, H., W. McIntyre, 2019: Numerical Solution of the Conditionally Averaged Equations for Representing Net Mass Flux due to Convection. Quart. J. Roy. Meteor. Soc., https://doi.org/10.1002/qj.3490

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

SuPy: An urban land surface model for Pythonista

By: Ting Sun

Python is now extensively employed by the atmospheric sciences community for data analyses and numerical modelling thanks to its simplicity and the large scientific Python ecosystem (e.g., PyData community). Although I cherish Mathematica as my native programming language (like Mandarin as my mother tongue), I can see I have coded much more in Python than in Mathematica over the past year for my urban climate research.

One of the core tasks in urban climate research is to build climate resilience, where climate information at various spatiotemporal scales is an essential prerequisite. To obtain such climate information, accurate and agile modelling capacity of the urban climate is essential. Urban land surface models (ULSM) are widely used to simulate urban-atmospheric interactions by quantifying the energy, water and mass fluxes between the surface and urban atmosphere.

One widely used and tested ULSM, the Surface Urban Energy and Water balance Scheme (SUEWS) developed by Micromet@UoR, requires basic meteorological data and surface information to characterise essential urban features (i.e., urban surface heterogeneity and anthropogenic dynamics). SUEWS enables long-term urban climate simulations without specialised computing facilities (Järvi et al., 2011; 2014a; Ward et al., 2016). SUEWS is regularly enhanced and tested in cities under a range of climates worldwide.

Figure 1: SUEWS-Centred workflow for urban climate simulations

The typical workflow of conducting a SUEWS simulation may consist of a few steps (Figure 1), where several pre- and post-processing procedures can in fact be easily done by Python. However, one inevitable step that often bothers me is that the change of a single parameter may lead to another loop of the above workflow, which is somewhat annoying and tedious as you need to switch back and forth between numerous applications again and again.

Given the glue-like ability of Python, I started the project SuPy (SUEWS in Python) to use Python as a central tool to build a SUEWS-back-ended urban land surface model since the development of SUEWS v2017b. After several months of development and testing, I’m very pleased to release SuPy (Sun 2019; Sun and Grimmond 2019) via PyPI that allows the easy installation with the following one-liner for all desktop/server platforms (i.e., Linux, macOS and Windows) with Python 3.6+:

python3 -m pip install -U supy

Figure 2: SuPy-aided workflow for urban climate simulations

Figure 3: SuPy simulation results of surface energy balance.

And the whole workflow in Figure 1 can now be done in a much simpler way (Figure 2) with the following code in Python for one stop to quickly perform a simulation and produce a plot of its results (Figure 3):

import supy as sp
 
#load sample data
df_state_init, df_forcing = sp.load_SampleData()
grid = df_state_init.index[0]
 
#run supy/SUEWS simulation
df_output, df_state_end = sp.run_supy(df_forcing, df_state_init)
 
#plot results
res_plot = df_output.loc[grid,’SUEWS’].loc[‘2012 6 4′:’2012 6 6’, [‘QN’, ‘QF’, ‘QS’, ‘QE’, ‘QH’]].plot()

Along with the software, we also have a dedicated documentation site to provide more information of SuPy (e.g, installation, usage, API, etc). In particular, to familiarise users with SuPy urban climate modelling and to demonstrate the functionality of SuPy, we provide three tutorials in Jupyter notebooks. They can run in browsers (desktop, mobile) either by easy local configuration or on remote servers with pre-set environments (e.g., Google Colaboratory, Microsoft Azure Notebooks). Those impatient tasters of SuPy can even try out the package in an online Jupyter environment without any configuration by clicking here.

The SuPy package represents a significant enhancement that supports existing and new model applications, reproducibility, and functionality. We expect SuPy will help guide future development of SUEWS (and similar urban climate models) and enable new applications of the model. Moreover, the improvement in SUEWS model structure and deployment process introduced by the development of SuPy paved the way to a more robust workflow of SUEWS for its sustainable success. To foster the sustainable development of SuPy as an open source tool, we welcome all kinds of contributions – for example, incorporation of new feature (pull requests), submission of issues, and development of new tutorials.

References:

Järvi, L., C. S. B. Grimmond, and A. Christen, 2011: The Surface Urban Energy and Water Balance Scheme (SUEWS): Evaluation in Los Angeles and Vancouver, J. Hydrol., 411(3-4), 219–237, doi:10.1016/j.jhydrol.2011.10.001

Järvi, L., C. S. B. Grimmond, M. Taka, A. Nordbo, H. Setälä and I. B. Strachan, 2014: Development of the Surface Urban Energy and Water Balance Scheme (SUEWS) for cold climate cities, Geosci Model Dev, 7(4), 1691–1711, doi:10.5194/gmd-7-1691-2014

Sun, T.: sunt05/SuPy: 2019.2 Release, doi:10.5281/zenodo.2574405, 2019.

Sun, T. and Grimmond, S.: A Python-enhanced urban land surface model SuPy (SUEWS in Python, v2019.2): development, deployment and demonstration, Geosci. Model Dev. Discuss., https://doi.org/10.5194/gmd-2019-39, in review, 2019.

Ward, H. C., S. Kotthaus, L. Järvi, and C. S. B. Grimmond, 2016: Surface Urban Energy and Water Balance Scheme (SUEWS): Development and evaluation at two UK sites, Urban Climate, 18, 1–32, doi:10.1016/j.uclim.2016.05.001

 

Posted in Boundary layer, Climate, Climate modelling, Urban meteorology | Leave a comment