Coupled atmosphere-ocean data assimilation re-interpreted

Coupled atmosphere-ocean data assimilation re-interpreted

by Polly Smith

So my original plan for this blog was to write something about my research on coupled atmosphere-ocean data assimilation but then my PI Amos Lawless beat me to it with his recent post. I was pondering on how I might put a new spin on things when I remembered someone telling me about the Up-Goer Five challenge. The idea is to try to describe a complicated subject using only the thousand most common English words. A quick google search shows that as usual I am rather late to the party – plenty of people have already had a go and there are some brilliant descriptions out there.

This is what I came up with …

People want to know if it will be hot or cold, dry or raining today, tomorrow, next month, next year and further. People who work with large pretend worlds on big computers want to get better at telling everyone what might happen to help them plan ahead; to do this we need to look at what both the near-sky and big blue water are doing now and what they might do in time to come. To get the best guess going forward we have to start from the right point. My work is trying to help the big computer people move forward with this problem. This is hard because there are lots of things we do not know about how the real world near-sky and big blue water work and about the conversations they have together; they are joined to each other but not in an easy way. Seeing what the near-sky is doing can help us understand the big blue water and seeing what the big blue water is doing can help us understand the near-sky, but we can’t see everything all the time. Our approach is to use the things we already understand about how the near-sky and big blue water work together to make a small pretend world on a computer, and then add some numbers seen from the real world about what the near-sky and big blue water are doing close to now and what they were doing in the near past. We also give the computer a starting guess and tell it how good we think this first guess is and how good we think the seen world is – this is hard to know for sure but we are working on different ways to do this as best we can. The computer takes our directions and puts all this stuff together in an amazing way to give us an even better starting point from which we can build a picture of how things may look going forward. Things can change very quickly in the real world so finding exactly the right answer is not possible, but we can make a very good guess and we expect to get even better as we continue to work hard to learn more about the world.

This was far from easy, but a really useful exercise in describing science in a non-technical way; a great idea if you want to try and communicate your science to a wider audience. Why not give it a go?

“Weather” by Randall Munroe (https://xkcd.com/1324/)

 

 

Improving Aircraft Observations using Data Assimilation

Improving Aircraft Observations using Data Assimilation

By Jeremy Holzke

I am half way through my six week Summer research placement which is funded by the EPSRC DARE project. As a second year Robotics student at the University of Reading, I am interested in collecting data from various sources and processing it so it can then be used for a robot to interact with its environment. I am undertaking this project as it has a very similar goal to a robot sensing its environment apart that the processed data will be used for better estimates of temperature in our atmosphere. I am also taking part in this Summer placement to see if I would be interested to do research in the future as my career. I will be investigating how data from aircraft and data from a numerical weather prediction (NWP) model can be combined to give the best estimate of the true temperature at the location of the observation.

Collecting observations from aircraft for meteorological purposes is most definitely not a new concept; in fact, it has been around since World War 1. The number of observations collected has grown ever since, especially with the wide range of applications weather now/forecasting provides. Some of these include military applications, agriculture and in particular for air traffic management. A main advantage of using aircraft derived observations in the 21st century, is that there around 13-16 thousand planes around the world at any time that can transmit valuable meteorological data.

Most commercial airplanes transmit a report called Mode Selective Enhanced Surveillance (Mode-S EHS) which contains data such as the speed, direction and altitude of the plane, as well as the Mach number which can be used to derive temperature and horizontal wind observations. The advantage of using Mode-S EHS reports is that they are transmitted at a high frequency but because of the short nature of the reports, data precision is reduced. Hence, large errors can appear in the derived temperature.

The aim of this project is to take aircraft-derived observations and combine them with modelled weather data from the Met Office UKV (UK variable resolution model), to get a better estimate of the temperature observation. A technique known as Optimal Interpolation, that takes account of the relative uncertainties in the two data sources was implemented in MATLAB. I have carried out some initial tests of the method using observation data from the National Centre for Atmospheric Science’s research plane; the Facility for Atmospheric Airborne Measurements (FAAM).

References:

A.K. Mirza, S.P. Ballard, S.L. Dance, P. Maisey, G.G. Rooney, and E.K. Stone, “Comparison of aircraft-derived observations with in situ research aircraft measurements,” Quarterly Journal of the Royal Meteorological Society, vol. 142, no. 701, pp. 2949–2967, 2016, issn: 1477-870X. doi: 10.1002/qj.2864 [Online]. Available from Royal Meteorological Society

 

Can observations of the ocean help predict the weather?

Can observations of the ocean help predict the weather?

by Dr Amos Lawless

It has long been recognized that there are strong interactions between the atmosphere and the ocean. For example, the sea surface temperature affects what happens in the lower boundary of the atmosphere, while heat, momentum and moisture fluxes from the atmosphere help determine the ocean state. Such two-way interactions are made use of in forecasting on seasonal or climate time scales, with computational simulations of the coupled atmosphere-ocean system being routinely used. More recently operational forecasting centres have started to move towards representing the coupled system on shorter time scales, with the idea that even for a weather forecast of a few hours or days ahead, knowledge of the ocean can provide useful information.

A big challenge in performing coupled atmosphere-ocean simulations on short time scales is to determine the current state of both the atmosphere and ocean from which to make a forecast. In standard atmospheric or oceanic prediction the current state is determined by combining observations (for example, from satellites) with computational simulations, using techniques known as data assimilation. Data assimilation aims to produce the optimal combination of the available information, taking into account the statistics of the errors in the data and the physics of the problem. This is a well-established science in forecasting for the atmosphere or ocean separately, but determining the coupled atmospheric and oceanic states together is more difficult. In particular, the atmosphere and ocean evolve on very different space and time scales, which is not very well handled by current methods of data assimilation. Furthermore, it is important that the estimated atmospheric and oceanic states are consistent with each other, otherwise unrealistic features may appear in the forecast at the air-sea boundary (a phenomenon known as initialization shock).

However, testing new methods of data assimilation on simulations of the full atmosphere-ocean system is non-trivial, since each simulation uses a lot of computational resources. In recent projects sponsored by the European Space Agency and the Natural Environment Research Council we have developed an idealised system on which to develop new ideas. Our system consists of just one single column of the atmosphere (based on the system used at the European Centre for Medium-range Weather Forecasts, ECMWF) coupled to a single column of the ocean, as illustrated in Figure 1.  Using this system we have been able to compare current data assimilation methods with new, intermediate methods currently being developed at ECMWF and the Met Office, as well as with more advanced methods that are not yet technically possible to implement in the operational systems. Results indicate that even with the intermediate methods it is possible to gain useful information about the atmospheric state from observations of the ocean. However, there is potentially more benefit to be gained in moving towards advanced data assimilation methods over the coming years. We can certainly expect that in years to come observations of the ocean will provide valuable information for our daily weather forecasts.

Figure 1

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

References

Smith, P.J., Fowler, A.M. and Lawless, A.S. (2015), Exploring strategies for coupled 4D-Var data assimilation using an idealised atmosphere-ocean model. Tellus A, 67, 27025, http://dx.doi.org/10.3402/tellusa.v67.27025.

Fowler, A.M. and Lawless, A.S. (2016), An idealized study of coupled atmosphere-ocean 4D-Var in the presence of model error. Monthly Weather Review, 144, 4007-4030, https://doi.org/10.1175/MWR-D-15-0420.1

First recording of surface flooding in London using CCTV cameras

On Friday 2nd of June 2017 Met Office issued a yellow warning of heavy rain with possible hail and lightning over London. Also Environmental Agency issued a number of flood alerts for London for the same period of time. This allowed us to test our newly setup system for recording open data CCTV images from London Transport Cameras (aka JamCams).

Following the flood alerts we setup to record all Transport for London (TFL) cameras which where within the main flood alert areas, these were 4 areas in London.

Figure 1. Areas selected for recording TFL CCTV camera images on 2nd of June 2017 corresponding to flood alerts from Environmental Agency.

This resulted in downloading images from just over 110 CCTV cameras accross from  the marked areas in Figure 1. Dowload started on many cameras at 2:30pm on 2nd of June 2017 and continued for 24h with an image downloaded every 5min.

Many of these images showed heavy rain as it passed over London on the afternoon of the 2nd June 2017; some cameras even captured images of lightning which was seen over North London but we didn’t capture any images of flooding in the four coloured areas in Figure 1.

Figure 2. Image of heavy rain on A23 Brixton Rd/Vassell Rd as seen by one of the CCTV cameras in London on 2nd July 2017 at 5:19pm
Figure 3. Image of lightning on captured on London CCTV camera at A12 East Cross Route on 2nd of June 2017 at 4:17pm

However, following the flooding allert on London for Transport site allowed us to capture surface flooding that happened on the North Circular road between 4-7pm resulting in traffic jams in the area.

Figure 4. Map of the surface flooding on the North Circular on 2nd of June 2017

The surface flooding was very localised and only one camera captured it, the one just below the blue circle in the Figure 4. We recorded both still and video images from this camera.

We are currently setting up similar systems to download live traffic CCTV images from Leeds, Bristol, Exeter, Newcastle, Glasgow, and Tewkesbury.

Data Assimilation in the Snow

by Sarah Dance

Snowbird mountains

I’ve just got back from attending the Society of Industrial and Applied Mathematics (SIAM) Conference on Dynamical Systems in the beautiful mountains of Snowbird, Utah, USA.  I was invited to attend the meeting to give part of a Mini-Tutorial on Data Assimilation (available here) with Elaine Spiller and Eric Kostelich.

Even though my undergraduate degree and Ph.D. were in Applied Mathematics, I don’t tend to go to many Mathematics conferences. I often meet with fellow data assimilation practitioners at Meteorology conferences instead.  So it was great to see people proving data assimilation related theorems, applying data assimilation in different applications like neuroscience and cancer treatment, and of course, to get some new ideas from dynamical systems approaches that have potential to be applied in different ways.  I particularly enjoyed Mary Silber’s talk on using Landsat data to understand vegetation pattern formation in the drylands of Africa

A slow march through the desert
Gowda/Silber’s work on African drylands. This image shows shrublands in Somalia from high above. Two images – from 1952 (purple) and 2006 (green) – are overlaid here for comparison. The colors highlight the large communities of shrubs and grasses which grow in bands along this sloping landscape. Over the fifty years shown here, all the vegetation has moved uphill – the green bands of modern plant growth are further up the hillside than the purple bands from 1952.

Tales from the Alice Holt Forest: carbon fluxes, data assimilation and fieldwork

by Ewan Pinnington

Forests play an important role in the global carbon cycle, removing large amounts of CO2 from the atmosphere and thus helping to mitigate the effect of human-induced climate change. The state of the global carbon cycle in the IPCC AR5 suggests that the land surface is the most uncertain component of the global carbon cycle. The response of ecosystem carbon uptake to land use change and disturbance (e.g. fire, felling, insect outbreak) is a large component of this uncertainty. Additionally, there is much disagreement on whether forests and terrestrial ecosystems will continue to remove the same proportion of CO2 from the atmosphere under future climate regimes. It is therefore important to improve our understanding of ecosystem carbon cycle processes in the context of a changing climate.

Here we focus on the effect on ecosystem carbon dynamics of disturbance from selective felling (thinning) at the Alice Holt research forest in Hampshire, UK. Thinning is a management practice used to improve ecosystem services or the quality of a final tree crop and is globally widespread. At Alice Holt a program of thinning was carried out in 2014 where one side of the forest was thinned and the other side left unmanaged. During thinning approximately 46% of trees were removed from the area of interest.

Figure 1: At the top of Alice Holt flux tower.

 

Using the technique of eddy-covariance at flux tower sites we can produce direct measurements of the carbon fluxes in a forest ecosystem. T

he flux tower at Alice Holt has been producing measurements since 1999 (Wilkinson et al., 2012), a view from the flux tower is shown in Figure 1. These measurements represent the Net Ecosystem Exchange of CO2 (NEE). The NEE is composed of both photosynthesis and respiration fluxes. The total amount of carbon removed from the atmosphere through photosynthesis is termed the Gross Primary Productivity (GPP). The Total Ecosystem Respiration (TER) is made up of autotrophic respiration (Ra) from plants and heterotrophic respiration (Rh) from soil microbes and other organisms incapable of photosynthesis. We then have, NEE = -GPP + TER, so that a negative NEE value represents removal of carbon from the atmosphere and a positive NEE value represents an input of carbon to the atmosphere. A schematic of these fluxes is shown in Figure 2.

                                                           Figure 2: Fluxes of carbon around a forest ecosystem.

 

The flux tower at Alice Holt is on the boundary between the thinned and unthinned forest. This allows us to partition the NEE observations between the two areas of forest using a flux footprint model (Wilkinson et al., 2016). We also conducted an extensive fieldwork campaign in 2015 to estimate the difference in structure between the thinned and unthinned forest. However, these observations are not enough alone to understand the effect of disturbance. We therefore also use mathematical models describing the carbon balance of our ecosystem, here we use the DALEC2 model of ecosystem carbon balance (Bloom and Williams, 2015). In order to find the best estimate for our system we use the mathematical technique of data assimilation in order to combine all our available observations with our prior model predictions. More infomation on the novel data assimilation techniques developed can be found in Pinnington et al., 2016. These techniques allow us to find two distinct parameter sets for the DALEC2 model corresponding to the thinned and unthinned forest. We can then inspect the model output for both areas of forest and attempt to further understand the effect of selective felling on ecosystem carbon dynamics.

Figure 3: Model predicted cumulative fluxes for 2015 after data assimilatiom. Solid line: NEE, dotted line: TER, dashed line: GPP. Orange: model prediction for thinned forest, blue: model prediction for unthinned forest. Shaded region: model uncertainty after assimilation (± 1 standard deviation).

 

In Figure 3 we show the cumulative fluxes for both the thinned and unthinned forest after disturbance in 2015. We would probably assume that removing 46% of the trees from the thinned section would reduce the amount of carbon uptake in comparison to the unthinned section. However, we can see that both forests removed a total of approximately 425 g C m-2 in 2015, despite the thinned forest having 46% of its trees removed in the previous year. From our best modelled predictions this unchanged carbon uptake is possible due to significant reductions in TER. So, even though the thinned forest has lower GPP, its net carbon uptake is similar to the unthinned forest. Our model suggests that GPP is a main driver for TER, therefore removing a large amount of trees has significantly reduced ecosystem respiration. This result is supported by other ecological studies (Heinemeyer et al., 2012, Högberg et al., 2001, Janssens et al., 2001). This has implications for future predictions of land surface carbon uptake and whether forests will continue to sequester atmospheric CO2 at similar rates, or if they will be limited by increased GPP leading to increased respiration. For more information on this work please see Pinnington et al., 2017.

 

References

Wilkinson, M. et al., 2012: Inter-annual variation of carbon uptake by a plantation oak woodland in south-eastern England. Biogeosciences, 9 (12), 5373–5389.

 

Wilkinson, M., et al., 2016: Effects of management thinning on CO2 exchange by a plantation oak woodland in south-eastern England. Biogeosciences, 13 (8), 2367–2378, doi: 10.5194/bg-13-2367-2016.

 

Bloom, A. A. and M. Williams, 2015: Constraining ecosystem carbon dynamics in a data-limited world: integrating ecological “common sense” in a model data fusion framework. Biogeosciences, 12 (5), 1299–1315, doi: 10.5194/bg-12-1299-2015.

 

Pinnington, E. M., et al., 2016: Investigating the role of prior and observation error correlations in improving a model forecast of forest carbon balance using four-dimensional variational data assimilation. Agricultural and Forest Meteorology, 228229, 299 – 314, doi: http://dx.doi.org/10.1016/j.agrformet.2016.07.006.

 

Pinnington, E. M., et al., 2017: Understanding the effect of disturbance from selective felling on the carbon dynamics of a managed woodland by combining observations with model predictions, J. Geophys. Res. Biogeosci., 122, doi:10.1002/2017JG003760.

 

Heinemeyer, A., et al., 2012: Exploring the “overflow tap” theory: linking forest soil co2 fluxes and individual mycorrhizo- sphere components to photosynthesis. Biogeosciences, 9 (1), 79–95.

 

Högberg, P., et al., 2001: Large-scale forest girdling shows that current photosynthesis drives soil respiration. Nature, 411 (6839), 789–792.

 

Janssens, I. A., et al., 2001: Productivity overshadows temperature in determining soil and ecosystem respiration across european forests. Global Change Biology, 7 (3), 269–278, doi: 10.1046/j.1365-2486.2001.00412.x.

2017 Annual European Geosciences Union (EGU) Conference

by Liz Cooper

The 2017 Annual European Geosciences Union (EGU) conference was held at the International Centre in Vienna from 23rd to 28th April.  During that time over 14,000 scientists from 107 countries shared ideas and results in the form of talks, posters and PICOs .The PICO (Presenting Interactive COntent) format is a relatively new idea for presenting work, where participants prepare an interactive presentation. In each PICO session the presenters first take turns to give a 2 minutes summary of their work for a large audience. The PICOS are then each displayed on an interactive touch screen and conference delegates can chat to the presenters and get further details on the research, with the PICO for illustration. This format has features of both traditional poster and oral presentations and provides a great scope for audience participation. I saw several which took advantage of this, including a very popular flood forecasting adventure game by a fellow Reading Phd student Louise Arnal.

I was delighted to be able to present some of my own recent results at EGU, in a talk titled ‘The effect of domain length and parameter estimation on observation impact in data assimilation for inundation forecasting.’ (see photo)

Presenting at an international conference was a really valuable and enjoyable experience, if a little daunting beforehand. I found it a really useful opportunity to get feedback from experts in the field and find out more about work by people with related interests.

The EGU conference has many participants and covers a huge range of topics from atmospheric and space science to soil science and geomorphology. My research deals with data assimilation for inundation forecasting, so I was most interested in sessions within the Hydrological Sciences and Nonlinear Processes in Science programmes. Even within those disciplines there was a huge breadth of research on display and I saw some really interesting work on synchronization in data assimilation, approaches to detection of floods from satellite data and various methods for measuring and characterizing floods.

As well as subject-specific programmes, there was also a very good Early Career Scientist (ECS) programme at EGU, with networking events, discussion sessions and a dedicated ECS lounge with much appreciated free coffee!

EGU was a hugely enjoyable experience and Vienna is a beautiful city with excellent transport links. With so many parallel sessions it’s really essential to plan which talks and posters are a priority in advance but I would heartily recommend it to anyone involved in geosciences research.

7th Japanese Data Assimilation Workshop

By Joanne A. Waller

For decades data assimilation (DA) has played a crucial role in numerical weather prediction (NWP) where it is used to provide initial conditions for weather forecasts. These ‘initial conditions’ describe the current atmospheric state and are estimated using data assimilation by blending previous forecasts with atmospheric observations, weighted by their respected uncertainties. However data assimilation is not only applicable to NWP and in recent years it has been applied widely to different applications where numerical simulations and observations are available.

At the end of February 2017 over 100 scientists from around the globe arrived at the Japanese RIKEN Advanced Institute for Computational Science (AICS)  for the 7th Japanese Data Assimilation Workshop. The aim of the symposium was to bring together scientist from from numerous different disciplines, such as neuroscience, cardiology, molecular dynamics, cosmology, nanoscale materials science, terrestrial magnetism, paleoclimate, oceanography, atmospheric chemistry and of course NWP, to discuss the data assimilation issues shared  across these broad applications.

Presentations and posters covered a wide variety of topics including: how data assimilation combined with advanced intelligence can help improve numerical models; how high performance computing can be used to deal with the new era of ‘Big Data’; how non-Gaussianity and non-linearity can be handled in data assimilation; ideas on how assimilate data into multi component models (i.e. systems that connect multiple models such as atmospheric, land and ocean models) and many more.

The conference provided a perfect platform for many cross-disciplinary discussions and this highlighted that much can be learnt in general about data assimilation by considering the issues that arise across different scientific areas.

(Photo from http://www.data-assimilation.riken.jp/risda2017/)

Mathematics of Planet Earth Jamboree

by Jemima Tabeart

On 20th-22nd March the Mathematics of Planet Earth Centre for Doctoral Training (MPE CDT) held its third annual Jamboree event. This is a celebration of the work of the staff and students of the CDT and includes seminars from industrial and academic speakers, as well as the opportunity for students to present their research. For the first time this year, the first two days of the Jamboree were used to host an Industrial study group. Representatives from EDF Energy and AIR Worldwide (catastrophe modelling for the insurance and re-insurance industry) posed real-world problems to cross-cohort groups of students, who then attempted to provide some new mathematical insight into possible solutions.

 

Our group was given a task by EDF Energy to investigate the interaction of extreme wind and rain events in the UK. EDF Energy’s assets in the UK include nuclear and other types of power plants, so understanding of extreme events is important in order to they can take appropriate safety measures. Currently extreme rain and wind events are considered separately, and we were asked to consider ways of determining how to define and deal with extreme wind-rain events. We were given hourly reanalysis data from the last 40 years, on a coarse 1 degree grid over the UK. The group split into two parts: one looking at more conceptual ideas about how extreme events can be caused by an interaction of factors, and the other considering the data provided.

 

Our part of the group identified some known extreme weather events, and focused on the data for these time periods. We looked at which events had both extreme wind and extreme rain, and mapped these to geographical locations to see where extreme wind-rain occurs most frequently. We also tried to see if there was a time lag between rain and wind events in the same location. Initial plots indicated that the most likely lag time was 0 hours, although this might be due to the relatively coarse resolutions. Other members of the group also suggested a method for combining the threshold values for extreme wind and rain to create a combined parameter. As well as a presentation of the main ideas that took place on the day to industry representatives, written reports will be sent to the respective companies so that they can take the suggestions further.

 

The study group was a great opportunity for cross-cohort work that brought together students with contrasting research interests. The challenge of producing something in a short amount of time is very different to what we normally expect as PhD students, and the ideas of getting stuck in straight away and not spending hours agonising over every decision is something that will be useful going forward. I really enjoyed working with real-world data and on problems outside my usual subject area – applying techniques I’ve learned during my PhD to other applications is very satisfying, and gives me the confidence that I am developing my transferable skills through my research!

DARE Board Meeting

On 22 February 2017 we held  our first advisory board meeting with stakeholders.

The presentations from the meeting may be downloaded here: