Improving Hydrological Predictions Of Land System Models

Posted on 7 December, 2020 by danaallen

By: Thibault Hallouin

Given the existence of feedbacks between the Earth’s atmosphere and the Earth’s surface, hydrological knowledge (e.g. soil moisture and open water available for evaporation and plant transpiration) is as critical to atmospheric scientists, as meteorological knowledge is to hydrologists. In the context of a changing climate, both communities need to work together, so that we can model the impacts of changes in atmospheric conditions on hydrological conditions and vice versa. This is crucial to answering key societal questions regarding the future availability of water resources, and the intensity of extremes such as floods and droughts.

Historically, land system models were developed as a lower boundary condition to atmospheric models. To this day, this still has implications on how land system models represent hydrological processes. When a land system model is coupled to an atmospheric model, it typically inherits the spatial resolution of the atmospheric model. However, the spatial resolution of any model must be appropriately chosen with respect to the spatial scale of the dynamics being modelled, and the spatial scales of atmospheric and hydrological processes can differ. Land system models typically use a tiling scheme to consider any sub-grid heterogeneity beyond the atmospheric grid resolution: the resolution of the land system is split into a collection of tiles, each considering different functional types (e.g. vegetation, bare soil, urban fabric, ice sheet, etc.). From a hydrological point of view, this approach is insufficient because it does not consider lateral movement of water on or below the surface, and it does not necessarily resolve the hydrological dynamics at their adequate spatial scale.

In order to overcome these limitations, the NERC National Capability 5-year programme Hydro-JULES is developing a framework for models of the terrestrial water cycle (see Figure 1). In collaboration with the UK Centre for Ecology & Hydrology, Bryan Lawrence, Grenville Lister, and I actively contribute to the implementation of this framework, given the expertise of the Department of Meteorology in complex model development and model integration.

Figure 1: The Hydro-JULES framework subdividing the terrestrial water cycle into three inter-connected components.

This framework represents the hydrological processes in the terrestrial water cycle as three interconnected components: surface layer, subsurface, and open water. Each component runs at its own resolution. This supports lateral movement of water and water-borne contaminants through the landscape and allows for the community of hydrological modellers to contribute alternative components that can be compared with existing ones, to improve hydrological predictions.

In collaboration with the UK Met Office, a key objective of Hydro-JULES is to refactor the Joint UK Land Environment Simulator (JULES) so it complies with this new framework. JULES is the land component currently used in the Met Office Unified Model, as in, for example, the UK Earth System Model configuration. JULES is also used on its own to model the land system alone. Therefore, advances in the representation of hydrological processes in JULES will not only benefit the hydrological community but will also benefit atmospheric modellers. Accounting for the lateral movement of water will also benefit ocean and Earth system modellers because water-borne contaminants such as nutrients and sediment are drained to the oceans.

Posted in Climate, Hydrology, Urban meteorology | Leave a comment

Does working from home help to reduce climate change?

Posted on 23 November, 2020 by danaallen

By: Helen Dacre

During a recent conversation about working from home during the COVID-19 pandemic a friend asked me ‘Given that I no longer commute to work in my car every day, will that help to reduce climate change?’  The answer to this question is probably yes, due to a reduction in CO2 emissions (ignoring increased heating and electricity use).  But by how much?  This got me thinking about the overall impact of COVID-19 emissions reductions on CO2 concentrations in the atmosphere and therefore on climate change. 

Since the start of 2020, COVID-19 restrictions have significantly reduced power generation, industrial activity and transport volume. Current estimates suggest that in April global CO2 emissions were down by 12-25% compared to the previous year and that over 2020 CO2 emissions may be down by 7-8%.  A decline in annual CO2 emissions of this size would exceed any decline since the end of WWII.  So, can we detect this decline in the atmospheric CO2 concentrations?

Background CO2 measurements are taken at several sites around the UK.  One of these sites is located at Ridge Hill in the West Midlands (O’Doherty et al. 2019). Figure 1(a) (blue line) shows daily CO2 concentrations at Ridge Hill between 1 January 2020 and 31 May 2020.  Interestingly CO2 concentrations start to reduce from April onwards, after the UK went into lockdown (23 March).  However, if we zoom out and look at CO2 concentrations over a longer time period we find that there is a reduction in CO2 concentrations every year in the Spring (figure 1b, blue line).  This is because in the Spring plants begin to photosynthesize and consume CO2 from the atmosphere to use for growth. Therefore, the decrease in CO2 concentrations is not likely to be due to COVID-19 related CO2 emissions reductions.

Given the dramatic reduction in CO2 emissions, why isn’t there an obvious reduction in CO2 concentrations?  There are two explanations for this.  Firstly, the long atmospheric lifetime of CO2 (50-200 years) makes any perturbation in emissions small compared to the reservoir of CO2 currently present in the atmosphere.  This is a bit like tipping a bucket of water every month for the last 100 years into a swimming pool with a pin sized hole in it.  Would you notice a change in the water level if you stopped adding water for a few months? Probably not. Secondly, the large daily and annual variability of CO2 concentrations makes changes in CO2 concentrations difficult to detect. This is similar to trying to detect the change in water level during a gale.   The waves created by the wind make it difficult to measure the water level and detect any changes.

So how long would we need to wait to detect a change in daily CO2 concentrations due to COVID-19 magnitude CO2 emission reductions? To answer this question, I built a simple statistical model to predict CO2 concentrations using only meteorological data.  This model doesn’t capture the decadal-timescale interactions included in complex climate models, but it does allow me to determine how long it would take for COVID-19 magnitude emission reductions to be detected in daily CO2 concentration measurements over short timescales (2-5 years).

Figure 1: (Top) Ridge Hill daily CO2 concentrations from January  2020 – May 2020, observed (blue) and predicted (red). Lockdown on 23 March 2020 (black dashed). (Bottom) Ridge Hill daily, monthly and yearly averaged CO2 concentrations from January 2015 – May 2020, observed (blue, cyan and black respectively), predicted (red, orange and grey respectively).

I used 5 years of meteorological data (2015-2019) plus the date as explanatory variables in my statistical model.  The observed (blue line) and predicted (red line) daily CO2 concentrations are shown in figures 1(a) and (b).  The predicted CO2 concentrations match the observed daily CO2 concentrations pretty well (capturing 76% of the observed variability). The observed 2ppm/year increase (trend) in CO2 concentrations is explained by inclusion of the date in the model.  The observed seasonal cycle in CO2 concentrations (due to photosynthesis) is explained by inclusion of monthly averaged temperature in the model.  Finally, the observed day-to-day variability in CO2 concentrations (due to the weather) is explained by including wind speed, wind direction and boundary layer depth in the model.

Since the model compares reasonably well with the observations, I can perform simple emission scenario simulations with my model by varying the trend whilst maintaining the seasonal and daily variability. Setting the trend to 0ppm/year is equivalent to a net-zero emissions scenario.  Similarly setting the trend to 1.8ppm/year is equivalent to a 10% reduction in CO2 emissions. The difference between CO2 concentrations modelled with the observed 2ppm/year trend and those modelled using a reduced trend can be compared to the daily variability in observed CO2 concentrations. If the difference is larger than the daily variability then we can detect a change in CO2 concentrations due to a change in emissions. For COVID-19 magnitude emissions reductions of 10% this occurs after 36-54 months.  Thus, we would expect to detect a reduction in trend in the daily CO2 concentrations only after 4 years of sustained reduced emissions. Of course, if we average the data further to calculate monthly CO2 concentrations then we smooth out the daily variability.  This means that the variability in observed CO2 concentrations reduces and we can detect a reduction in the monthly CO2 concentration trend earlier, after about 12 months. Therefore, if current global lockdown restrictions continue we might detect a reduction in CO2 trend some time in 2021.

So, does working from home reduce climate change? Unfortunately, while the recent reductions in CO2 emissions are substantial, they do not immediately equate to similar reductions in the trend in atmospheric CO2.  COVID-19 magnitude reductions would only result in significantly reduced daily CO2 trend if they were sustained for many years.  This is bad news for climate change as it means that emissions reduction policies need to be both large and sustained to reverse the upward trend in CO2 concentrations. The current COVID-19 CO2 emissions reductions are similar in magnitude to those stated by the Paris agreement as necessary to keep global temperatures below 2oC. However, the methods employed to control the pandemic are not sustainable long-term. The COVID-19 crisis offers a real wake-up call to highlight the substantial changes in behaviour and infrastructure that are necessary if we are to achieve CO2 reduction targets set out by the Paris agreement.

References:

O’Doherty, S.; Say, D.; Stanley, K. (2019): Deriving Emissions related to Climate Change Network: N20, SF6, CO, H2 and other trace gas species measurements from Ridge Hill Tall Tower, Herefordshire. Centre for Environmental Data Analysis, accessed 1 November 2020. https://catalogue.ceda.ac.uk/uuid/4370dacd17544eb781aa1e51cc4dc633

Posted in Climate modelling, Covid-19, Greenhouse gases, IPCC | Leave a comment

Sources of rainfall over East Asia

Posted on 9 November, 2020 by danaallen

By: Liang Guo

The East Asian monsoon causes intense rainfall over China, Japan and the Koreas every summer, and a cold, dry season every winter. This is driven by thermal and dynamical contrasts between the vast Pacific Ocean and the elevated Tibetan Plateau. Over 80% of East Asian rainfall is carried in from tropical oceans and mid-latitude land and oceans(Guo, Klingaman et al. 2018).

The strength of the monsoon varies each year, with stronger monsoon seasons associated with a higher total rainfall. However, the intensity of individual rainfall events is decided by smaller-scale perturbations in the weather, such as a vortex moving down from Tibet or a tropical cyclone from the Pacific.

As extreme rainfall events are damaging to both people and property, and occurring more often(Chevuturi, Klingaman et al. 2018), we wanted to know how much of the rainfall in East Asia is caused by the mean monsoon flow and how much is caused by perturbations in the weather.

To investigate this, we used a moisture tracking tool: the Water Accounting Model (WAM). This deconvolves atmospheric circulation and moisture using a spatial filter, and assumes that the atmosphere maintains the same hydrological balance at each grid point as occurred in reality (i.e. evaporation and precipitation are unchanged) except over East Asia.Figure 1: Annual cycles of rainfall over five East Asian subregions tracked using decomposed moisture fluxes (coloured lines). Sums of each components (solid grey line) are compared to precipitation of ERA-Interim (dotted grey line).

Figure 1 shows the results of applying this tool to five different regions of East Asia: southeastern China (region 1), Tibetan Plateau (region 2), central eastern China (region 3), northwestern China (region 4) and northeastern China (region 5). This shows the amount of rainfall which occurred due to the mean flow (black); the mean humidity gradient (red); the eddy flow (yellow); the eddy humidity gradient (blue) over the course of a year. The contributions from each of these components added together show the total rainfall for the region (solid grey line). This matches up well to the total precipitation in the ERA-Intermin dataset (dotted grey line), which can be used as a reference.

We found that the mean monsoon flow and mean monsoon gradient are the largest causes of rainfall, particularly in southeastern China and Tibet.

References:

Chevuturi, A., et al. (2018). “Projected Changes in the Asian‐Australian Monsoon Region in 1.5°C and 2.0°C Global‐Warming Scenarios.” Earth’s Future 6(3): 339-358. https://agupubs.onlinelibrary.wiley.com/doi/full/10.1002/2017EF000734 

Guo, L., et al. (2018). “The contributions of local and remote atmospheric moisture fluxes to East Asian precipitation and its variability.” Climate Dynamics. https://journals.ametsoc.org/jhm/article/20/4/657/344214/Moisture-Sources-for-East-Asian-Precipitation-Mean

Posted in Climate, Monsoons | Tagged | Leave a comment

Tropical rainfall and sea surface temperature link could improve forecasts

Posted on 2 November, 2020 by danaallen

By: Chris Holloway  and the University of Reading Press Office. 

Tropical rainfall, averaged on seasonal time scales, is influenced far more strongly by nearby sea temperatures in the real world than in almost all climate simulations, scientists have found, paving the way for more accurate global weather forecasts.

A team led by Dr Peter Good at the Met Office and including Dr Chris Holloway at the University of Reading Meteorology Department studied the effect of tropical sea surface temperatures and resulting wind patterns on seasonal rainfall in the region by filtering out other influences, revealing a stronger relationship in the real world than that simulated by 43 of 47 climate models studied.

The study, published in Nature, used new analysis of satellite observations and other meteorological data and also found that important low-altitude wind patterns in the wider tropical region were stronger than most models simulate.  Deficiencies in the simulation of low-altitude cloud cover were highlighted as a potential cause of these discrepancies.

Weather and climate patterns in the tropics are known to have a knock-on effect on weather thousands of miles away. This means that the findings could help improve seasonal weather forecasts for Europe, as well as longer-term climate predictions.

Dr Chris Holloway said: “Global weather and climate models continue to have errors in simulating and predicting tropical rainfall patterns. Limited quality measurements of tropical rainfall and atmospheric circulation have also made it difficult to understand and rectify these errors.

“In our study, we were able to filter out interactions between remote regions to focus on the relationship between sea surface temperatures, rainfall and tropical winds within a particular region, showing us how much the models are underestimating the increase of rainfall that accompanies a warmer sea surface temperature.

“Resolving these discrepancies between models and the real world can make a big difference to how accurately we can predict the weather or how much confidence we have in projections of regional climate change in the future.”

Reference:

Good, P., Chadwick, R., Holloway, C.E., Kennedy, J., Lowe, J. A., Roehrig, R. and Rushley, S. S.  High sensitivity of tropical precipitation to local sea-surface temperature. Nature (2020). https://doi.org/10.1038/s41586-020-2887-3     (free read-only PDF)

 

Posted in Climate, Oceans, Weather forecasting | Tagged | Leave a comment

The Core-cloak Convection Model

Posted on 26 October, 2020 by danaallen

By: Jian-Feng Gu

Moist convection plays a fundamental role in large-scale circulations and climate, ranging from cumulus clouds smaller than 100m to organized weather systems of several thousands of kilometers. Limited by their grid spacing, numerical models are not able to fully resolve moist convection across its broad range of scales, therefore, we represent moist convection in numerical models using a simplified process, called parameterization.

Figure 1:. A schematic diagram of bulk mass flux approximation. The left panel shows many individual different clouds. The right panel shows how the clouds are represented using the top-hat assumption.

In current climate models, convection is mostly parameterized using bulk plume models, which estimate the vertical transport of heat, moisture and momentum of a large group of clouds. To understand the general idea, imagine a grid box of around (100×100) km2 – that’s about half the size of Wales. Imagine a large number of different clouds randomly scattered across the domain (see the left panel of figure 1). Representing each cloud individually is very computationally expensive and hence simplification is necessary. The simplest idea to describe the overall behavior of these clouds is to consider them as a single entity, that is, a bulk cloud (see the right panel of figure 1).

A further simplification is to assume that each property of the bulk cloud is the averaged value of all the individual clouds, and that these values are distributed evenly within the bulk cloud. This is called the top-hat assumption (Randall et al. 1992). This means that the overall vertical transport by these clouds can be described as the transport of mean properties by the bulk cloud, which is called the bulk mass flux approximation. It approximates the sub-grid vertical flux of a quantity as being the product of the convective mass flux with the departure from the grid-box average of the transported quantity. However, this approximation can underestimate the vertical fluxes by 30-50% (Yano et al. 2004), depending on the variables considered and the resolution of the model. Therefore, a parameterization of the neglected contributions to the vertical flux is necessary. How might this be achieved without sacrificing computational efficiency?

The vertical flux underestimation arises from the physical assumptions we make in the bulk plume model. In Gu et al. (2020), we show both the mean properties of the clouds, and the departures from these mean properties contribute toward the total vertical flux. By representing many clouds as a single bulk cloud, we are removing differences between individual clouds. By assuming a top-hat distribution, we are neglecting the inhomogeneity within each cloud. These two neglected variabilities are called inter-object and intra-object variability, respectively. As a result, the bulk mass flux approximation underestimates the total vertical transport by moist convection. However, it can be improved by relaxing these assumptions. For example, a spectral model that deals with clouds of different sizes is able to minimize the inter-object variability because it takes into account the differences of mean properties between different types of clouds. But it still underestimates the vertical heat fluxes because of neglecting the intra-object variability.

Figure 2: A schematic of the core-cloak representation of convection. Both updrafts and downdrafts are represented as the combination of a strong core surrounded by a weak cloak.

To improve the representation of both the vertical heat and water fluxes, we proposed the “core-cloak” conceptual model (Figure 2, Gu et al. 2020). In this model, we decompose the flow into different types of drafts depending on the strength of vertical motions. More specifically, we collect the strong updrafts together as the updraft “core” and the weak updrafts together as the updraft “cloak”. This core-cloak structure can also be applied to downdrafts. This flow decomposition partly includes the inter-object and intra-object variability and therefore better represents the vertical heat and water transport.

To evaluate our conceptual model, we performed simulations of shallow convection (dx=25 m, 50 m, 100 m) and deep convection (dx=100 m, 200 m, 400 m), and cloud resolving simulations of organized deep convection (dx=1 km). Our results show that the “core-cloak” conceptual model can significantly improve the representation of vertical heat and moisture fluxes, compared to the bulk mass flux approximation. The improvement can be seen in both shallow and deep convection, and even in organized convection.

We also found that the clouds which have a “core-cloak” structure contribute most of the vertical fluxes. Therefore, the “core-cloak” conceptual model provides simply a possible decomposition of the flow that gives a reasonable and efficient description of turbulent fluxes using a mass flux approximation. Parameterization of this core-cloak model would need careful treatment of exchanges between the different types of drafts. We intend to pursue the practical implications of this conceptual model within the future development of a convection parameterization.

References:

Gu, J.-F., Plant, R. S., Holloway, C. E., Jones, T. R., Stirling, A., Clark, P. A., Woolnough, S. J. and Webb, T. L., 2020: Evaluation of the bulk mass flux formulation using large eddy simulations. J. Atmos. Sci., 76, 2297–2324, doi: https://doi.org/10.1175/JAS-D-19-0224.1

Randall, D. A., Q. Shao, and C.-H. Moeng, 1992: A second-order bulk boundary-layer model. J. Atmos. Sci., 49, 1903-1923, doi: https://doi.org/10.1175/1520-0469(1992)049<1903:ASOBBL>2.0.CO;2

Yano, J.-I., F. Guichard, J.-P. Lafore, J.-L. Redelsperger, and P. Bechtold, 2004:  Estimations of mass fluxes for cumulus parameterizations from high-resolution spatial data. J. Atmos. Sci., 61, 829–842, doi: https://doi.org/10.1175/1520-0469(2004)061<0829:EOMFFC>2.0.CO;2.

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

People are in the cities – how could we provide the weather climate information they need?

Posted on 19 October, 2020 by danaallen

By: Sue Grimmond

Climate services provide climate information to help individuals and organizations make climate smart decisions. Climate services work by integrating high quality meteorological data (temperature, rainfall, wind, soil moisture and ocean conditions); as well as maps, risk and vulnerability analyses, assessments, and long-term projections and scenarios; with socio-economic variables and non-meteorological data (such as agricultural production, health trends, human settlement in high-risk areas, road and infrastructure maps for the delivery of goods). The data and information collected is transformed into customized products such as projections, trends, economic analyses and services. The aim is to equip decision makers in climate-sensitive sectors with better information to help society adapt to climate variability and change.  At the core are users and their needs.

The demands for such services are wide ranging – for public health; for disaster risk reduction and response; the prediction of energy and water demands or food production; or the operation of climate sensitive infrastructure. These services though need to go further, they need to be developed for a wider range of time scales and conditions (for weather and climate) and to integrate other elements of the environment and human behaviours and responses.

Nowhere is this more important than in cities. Increasingly dense, complex and interdependent urban systems leave cities particularly vulnerable to extreme weather and to changes in climate. Through domino effects, a single extreme event can lead to a wide-scale breakdown of a city’s infrastructure (Figure 1).

Figure 1:  The ‘domino effect’ partially shown for a typhoon or hurricane event, which produces multiple hydro-meteorological hazards (blue) that have immediate effects (green) and follow-on impacts (purple) that can be both short- and long-term. Source: Grimmond et al. 2020)

Integrated Urban Systems (IUS) are needed (Baklanov et al. 2018, Grimmond et al. 2020), yet globally there are few (if any) fully operational systems (e.g. Baklanov et al. 2020). Those systems that do exist often were developed because of a major event (e.g. Olympics; Pan-Am games, Expos) or a significant weather-related disaster (hurricane Sandy. New York City; heat waves in Paris and London in 2003). Recognising this need, the World Meteorological Organization (WMO) is advocating for the development of Integrated Urban Weather, Environment and Climate Services (IUS) for safe, healthy and resilient cities. The concept and methodology for developing these systems was adopted by the 70th WMO Executive Council in 2018 (WMO 2019). The need for Demonstration Cities was adopted by the 71st WMO Executive Council in 2019.

The research community, working collaboratively with others, has an important role to play across the many activities required. This includes contributing to and identifying critical research challenges, developing impact forecasts and warnings, promoting and delivering IUS internationally, and supporting national and local communities in their implementation (Figure 2).

Figure 2: An Integrated Urban Hydrometeorological, Climate and Environmental Service (IUS) System has several components.  Here a generic framework is shown for impact-based prediction systems. Integration may occur in the various boxes with a mature IUS aiming for integration in all components. Source: WMO (2019)

At Reading, we have been working with a range of collaborators and stakeholders to develop UMEP (Urban Multi-scale Environmental Predictor), a city-based climate service tool, that combines models and tools essential for climate simulations (Lindberg et al. 2018). Alongside climate information transformed for the urban context, detailed information about the materials and morphology of the city are integrated through GIS, with information about residents and their behaviour through agent-based models and tools (e.g. Capel- Timms et al. 2020). UMEP has been used to identify heat waves and cold waves; the impact of green infrastructure on runoff; the effects of buildings on human thermal stress; solar energy production; and the impact of human activities on heat emissions. It has been applied in many cities across the world, from London to Shanghai (Google Citations 2020).

UMEP includes tools to enable users to input atmospheric and surface data from multiple sources, to characterise the urban environment, to prepare meteorological data for use in cities, to undertake simulations and consider scenarios, and to compare and visualise different combinations of climate indicators. An open-source tool, UMEP is designed to be easily and widely used. This summer we offered the international urbisphere UMEP workshop, that was planned to be in person in Reading, online.

Much more work is needed in this realm. IUS need to be developed to meet the special needs of cities through a combination of dense observation networks, high-resolution forecasts (weather, climate, air quality, hydrological), multi-hazard early warning systems, disaster management plans and climate services. Such an approach will give cities the tools they need to reduce emissions, build thriving and resilient communities and implement the UN Sustainable Development Goals. This focus on urban environments is particularly important given the large and ever increasing fraction of the world’s population that live in cities (more than 3.5 billion) and the importance of cities, not only regionally but nationally and globally, to the world’s economy.

References: 

Baklanov A, CSB Grimmond, D Carlson, D Terblanche, X Tang, V Bouchet, B Lee, G Langendijk, RK Kolli, A Hovsepyan 2018: From urban meteorology, climate and environment research to Integrated City services, Urban Clim., 23, 330-341,  10.1016/j.uclim.2017.05.004

Baklanov A, B Cárdenas, T Lee, S Leroyer, V Masson, L Molina, T Müller, C Ren, FR Vogel, J Voogt 2019: Integrated urban services: Experience from four cities on different continents, Urban Clim., 32, 10.1016/j.uclim.2020.100610

Capel-Timms I, ST Smith, T Sun, S Grimmond 2020: Dynamic Anthropogenic activities impacting Heat emissions (DASH v1.0): Development and evaluation Geosci. Model Dev. https://doi.org/10.5194/gmd-2020-52

Grimmond S, V Bouchet, L Molina, A Baklanov, J Tan , H Schluenzen, G Mills, B Golding, V Masson, C Ren, J Voogt, S Miao, H Lean, B Heusinkveld, A Hovespyan, G Terrug, P Parrish, P Joe 2020: Integrated Urban Hydrometeorological, Climate and Environmental Services: Concept, Methodology and Key Messages Urban Climate https://doi.org/10.1016/j.uclim.2020.100623

Lindberg F, CSB Grimmond, A Gabey, B Huang, CW Kent, T Sun, NE Theeuwes, L Järvi, H Ward, I Capel-Timms, YY Chang, P Jonsson, N Krave, DW Liu, D Meyer, KFG Olofson, JG Tan, D Wästberg, L Xue, Z Zhang 2018: Urban multiscale environmental predictor (UMEP) – An integrated tool for city-based climate services Environmental Modelling and Software, 99, 70–87 https://doi.org/10.1016/j.envsoft.2017.09.020

WMO 2019: Guidance on Integrated Urban Hydrometeorological, Climate and Environmental Services Volume I: Concept and Methodology WMO-No. 1234 https://library.wmo.int/doc_num.php?explnum_id=9903

Posted in Climate, Urban meteorology, Weather forecasting | Leave a comment

Large and irreversible future decline of the Greenland ice-sheet

Posted on 12 October, 2020 by danaallen

By: Jonathan Gregory

Sea-level rise is one of the most serious consequences of global warming. By the end of this century, if emissions of greenhouse gases continue to increase (mostly carbon dioxide, from burning oil, natural gas and coal), global mean sea level could be more than a metre higher than now. About a quarter of a billion people currently occupy land less than a metre above present sea level (Kulp and Strauss, 2019). By the end of this century, most coastal locations around the world will annually experience extreme sea levels which have historically occurred as a result of violent storms only about once in 100 years (Ocean, cryosphere and climate change, Royal Society briefing, 2019).

Whereas global warming itself and some consequences of climate change could be mostly halted within in a few decades by ceasing greenhouse-gas emissions (although that would be hard enough to achieve), sea level would continue to rise for centuries or millennia under any climate as warm as or warmer than present. By 2300 sea level is projected to rise by 0.6-1.1 metres even if climate is stabilised in coming decades, and by 2.3-5.4 metres if emissions of greenhouse gases are large. It is hard to envisage the seriousness of this for some areas of the world. For instance, two-thirds of Bangladesh is less than 5 metres above sea level, and the highest point in the Maldives is 2.4 metres above sea level.

Large and irreversible future decline of the Greenland ice-sheet 

Figure 1: An image of the Greenland ice-sheet (CPOM/UCL/ESA).

There are several contributions to sea-level rise. In recent decades, a third to a half of the total has been due to the expansion of sea water as it gets warmer (Chambers et al., 2017), and this will continue to be an important effect. The future of the Antarctic ice-sheet is the largest uncertainty in projections of the current century. Although it is smaller, the Greenland ice-sheet (figure 1) is presently contributing more than the Antarctic, and more than all the world’s mountain glaciers together (Special report on the ocean and cryosphere in a changing climate, IPCC i.e. Intergovernmental Panel on Climate Change, 2019). As the climate gets warmer, both surface melting and snowfall on Greenland increase. (Precipitation generally increases in a warmer climate, and on Greenland most precipitation is snow.) The melting increases more rapidly with global warming, so there is a net loss of ice, which ends up as water in the ocean. If the ice-sheet was completely eliminated, global mean sea level would be 7.4 metres higher.

My colleagues Steve George, Robin Smith and I have recently studied the future of the Greenland ice-sheet under a range of climates (work under open review), illustrative of those expected in the late 21st century under various scenarios. In our experiments, the climates were constant. We wanted to see what happens if you maintain a warm climate indefinitely. We used an atmosphere general circulation climate model and an ice-sheet model coupled together. The climate model is like those used for IPCC projections but with less geographical detail because it has to run faster, since we needed to carry out experiments simulating tens of millennia. We ran about 50 experiments, at about 2000 simulated years per day of computer time. This is the first time the future of Greenland has been investigated with a model of such complexity; it has many unavoidable approximations and inaccuracies, but it’s more physically realistic and complete than previous models.

Large and irreversible future decline of the Greenland ice-sheet

Figure 2: Contribution of the Greenland ice-sheet to global-mean sea-level rise in our experiments. The coloured lines are the results of the first set of experiments under constant climates. The colours indicate the global warming with respect to the climate of the late 20th century, from blue (little warming) to red (5 degrees Celsius warmer). The black lines show the second set of experiments. The ice-sheet at the point along a coloured line where each black line begins was instantaneously transplanted into a late 20th-century climate. The solid black lines are from states below the threshold, in which the ice-sheet regrows to around its present size; the dashed black lines are those which began above the threshold, in which it never fully regrows.

Under all climates like the present or warmer the ice-sheet loses mass and contributes positively to sea level (figure 2). It takes tens of thousands of years to reach a new constant state: the warmer the climate, the smaller the final ice-sheet, and the larger the sea-level rise. Unlike in some previous studies, there is no sharp threshold dividing scenarios in which the ice-sheet suffers little reduction from those in which it is mostly lost. Rather, there is a broad range of outcomes. In the warmest climate we consider (about 5 degrees Celsius warmer than recent, which is similar to the most extreme scenarios for 2100), the ice-sheet is reduced over about 10,000 years to a small ice-cap with 1.5% of its present volume. Initially it contributes about 3 millimetres per year to sea level, which is similar to the current observed rate of rise due to all effects. On the other hand, in climates resulting from strongly mitigated emissions during this century (roughly consistent with the Paris target), the final contribution to sea-level rise is less than 1.5 metres.

In a second set of experiments, we took some of the reduced states of the ice-sheet and put them back in a steady climate like the late 20th century, to see if it would regrow, meaning that sea level would consequently fall. The ice-sheet gained mass in all cases, taking even longer than it did in a warm climate to reach a constant state, because snowfall adds mass more slowly than melting can remove it. We found that the final states constitute two groups. If the sea-level rise under the warm climate remained below about 3.5 metres, the ice-sheet eventually regrew to around its present size. If sea level passed this threshold, the ice-sheet did not fully regrow. In this case, about 2 metres of the sea-level rise was irreversible under recent climate. (The full ice-sheet could probably be regenerated in an ice-age climate.) The reason for the irreversibility is that the ice-sheet is a large object which affects its own local climate, like a high cold mountain. Without the ice-sheet, but with present-day temperatures of the surrounding seas, Greenland would a warmer place. In our model, the ice-sheet cannot readvance into the northern part of the island once ice-free, because the snowfall is less than at present.

While our result requires corroboration by other workers with their own models, it illustrates the importance of the coupling of the ice-sheet and its climate. Each affects the other and modelling them independently may lead to unrealistic projections. The experiments also underline the global practical importance of mitigating global warming. Precautionary action to mitigate the threat of irreversible damage is a principle of the Framework Convention of Climate Change, even when there is not full scientific certainty. According to our results, in order to avoid partially irreversible loss of the ice-sheet, climate change must be reversed (not just stabilised in a warmer state, but put back to how it was) before the ice-sheet has declined to the threshold mass, which would be reached in about 600 years at the highest rate of mass-loss for this century in the IPCC assessment.

References:

Kulp, S. A. and B. H. Strauss, 2019, New elevation data triple estimates of global vulnerability to sea-level rise and coastal flooding, Nat. Commun, 10, 4844, https://doi.org/10.1038/s41467-019-12808-z

Chambers, D. P, A. Cazenave, N. Champollion, H. Dieng, W. Llovel, R. Forsberg, K. V. Schuckmann and Y. Wada, 2017, Evaluation of the global mean sea level budget between 1993 and 2014, Surv. Geophys. 38, 309-327, https://doi.org/10.1007/s10712-016-9381-3

 

Posted in Climate, Climate change, Climate modelling, IPCC, Polar | Leave a comment

Recent progress in simulating North Atlantic weather regimes

Posted on 5 October, 2020 by danaallen

By: Alex Baker

Weather is chaotic. Low-pressure weather systems bring rainfall; areas of high pressure block the passage of these weather systems. Take this year so far, for instance. February and March were much wetter than average, and April and May much drier. May, in particular, was England’s driest—and the U.K.’s sunniest—on record, with similar conditions enjoyed across much of Western and Central Europe. Such swings in weather are down to where low- or high-pressure conditions prevail, and for how long these synoptic situations persist.

One way to make sense of this variability is by identifying so-called weather regimes, reoccurring patterns of high and low pressure across the central and eastern North Atlantic and Europe. Conventionally, meteorologists recognise four Euro-Atlantic regimes: the positive and negative phases of the North Atlantic Oscillation (NAO+ and NAO–, respectively), Scandinavian blocking (SB), and a North Atlantic ridge pattern (AR)—more on each presently. How often these regimes occur not only dictates regional weather, but also plays a role in whether a season becomes wetter or drier over time, and is important on longer, climatological timescales too.

Weather regimes exhibit characteristic spatial patterns of high- and low-pressure centres, visualised here using geopotential height data from which the climatological mean seasonal cycle was removed (see Figure 1). (Geopotential height is a common variable used to infer atmospheric circulation patterns; it tells us altitude above mean sea level, accounting for gravitational variations over Earth’s surface.) The North Atlantic Oscillation’s positive and negative phases describe variability between the Icelandic Low and the Azores High. During NAO+, high-pressure conditions prevail over much of Central Europe and the Mediterranean. During NAO–, the high-pressure anomaly sits over Greenland and low pressure spans much of continental Europe. Scandinavian blocking is the occurrence of a high-pressure anomaly over western Scandinavia and the North Sea. The Atlantic Ridge pattern is characterised by high pressure over the central North Atlantic at a latitude of about 55°N. Each regime roughly corresponds to a preferred position of the North Atlantic jet stream.

Figure 1: The Euro-Atlantic weather regimes, based on daily geopotential height data at the 500-mb isobaric level from the ERA40 and ERA-Interim reanalyses. The four regimes are the positive and negative phases of the North Atlantic Oscillation (NAO+ and NAO–, respectively), Scandinavian blocking (SB), and a North Atlantic ridge pattern (AR). Regimes patterns visualised following removal of the climatological mean seasonal cycle. Figure adapted from Fabiano et al., 2020.

In a recent paper published in Climate Dynamics, led by Federico Fabiano of the Institute of Atmospheric Sciences and Climate, Consiglio Nazionale delle Ricerche, we examined how well six current-generation, fully coupled global climate models are able to represent Euro-Atlantic weather regimes—their spatial patterns, their persistence, and how realistically distinct the regimes are. Here, I focus on regime patterns, and how well those simulated by low- and high-resolution climate models compare with reference datasets: the ERA-40 and ERA-Interim reanalyses.

Establishing whether or not global climate models can reproduce each regime’s characteristic spatial pattern is important because these patterns are related to where westerly storm systems track and make landfall downstream over Europe—and where these storms’ impacts will be felt. Do high-resolution models reproduce real-world regime patterns better than standard, low-resolution models? To assess this, we calculated pattern correlations between the models and reanalyses. Overall, we found that the NAO+, SB and AR regime patterns are better represented at high resolution, but the NAO– regime is not (see Figure 2). Why NAO– is something of an outlier here will be the subject of future research. Additionally, the AR regime shows greater variance than the other regimes. We also found that simulated regimes are more realistically distinct from one another (to use the jargon, more tightly ‘clustered’) at high resolution and better match the reanalyses.

Figure 2: Pattern correlations between low- (teal) or high-resolution (red) models and reanalyses (black) for each weather regime. Perfect model representation of a regime’s observed spatial pattern is indicated by a pattern correlation coefficient of 1. From distributions of 30-yr bootstrapping for each model, the ensemble mean (dot), median (horizontal line), interquartile range (boxes), 10th and 90th percentiles (bars), and minimum and maximum values across all available ensemble members (triangles) are shown. Figure adapted from Fabiano et al., 2020.

However, increasing models’ resolution had little impact on the frequency and duration of weather regimes. The evidence suggests that these errors are due to biases in simulated sea-surface temperatures and the mean geopotential height field. Simulating realistic regime persistence in models is important because prolonged wet and dry periods, like those seen across Europe earlier this year, are very often related to the persistence of a single regime. This research suggests that increasing model resolution alone is not enough; developments in model physics and dynamics are needed to better simulate North Atlantic weather regimes.

Author’s note

The climate models in this study participate in the sixth phase of the World Climate Research Programme’s Coupled Model Intercomparison Project (CMIP6), the modelling framework underpinning the Intergovernmental Panel on Climate Change’s Assessment Reports that are indispensable for global climate policy-making. These model simulations (hist-1950) were supported by the European Commission-funded PRIMAVERA project, the European contribution to HighResMIP, a CMIP6-endorsed and coordinated assessment of the impact of increasing model resolution, which is documented by Haarsma et al., 2016.

References

Fabiano, F. et al., 2020. Euro-Atlantic Weather Regimes in the PRIMAVERA coupled climate simulations: impact of resolution and mean state biases on model performance. Climate Dynamics 54, 5031–5048. . https://doi.org/10.1007/s00382-020-05271-w

Haarsma, R. J. et al., 2016. High Resolution Model Intercomparison Project (HighResMIP v1.0) for CMIP6. Geoscientific Model Development 9, 4185–4208.  https://doi.org/10.5194/gmd-9-4185-2016 

Posted in Atlantic, North Atlantic, Weather Regimes | Leave a comment

Do urban heat islands provide thunderstorm predictability?

Posted on 28 September, 2020 by danaallen

By: Suzanne Gray 

The UK and the rest of western Europe experienced a heatwave in the middle of August 2020 with temperatures exceeding 30oC in Reading. Fortunately for us this was broken by a heavy downpour on the afternoon of Wednesday 12th August (Figure 1(a)) yielding over 4 mm of rainfall at the Reading University Atmospheric Observatory over a period of only about 20 minutes. While watching the rain cascading down the road towards my front door (we live at the bottom of a small, but very steep, hill) I naturally searched up the latest radar images and analysis charts to see the system causing the rainfall and to try to predict how long it might last.

a)     b) 

Figure 1: (a) Temperature observations from the Reading University Atmospheric Observatory for the 11th and 12th August 2020 and (b) cutout from a Met Office analysis with mean sea level pressure contours and marked fronts from 12 UTC on 12th August 2020 (copyright Met Office).

The Met Office had issued a yellow warning for thunderstorms and associated flooding for most of the UK but, as is usual with thunderstorm warnings, was not able to tell us exactly where and when the thunderstorms would occur. According to the Met Office’s synoptic analysis (Figure 1(b)), the mean sea level pressure gradient over the UK was very weak and this was associated with weak easterly winds (~2-3 ms-1 at our Observatory). A small-scale weak low-pressure system to the west of the UK was associated with an upper-level trough directly above, with an extension of the trough axis towards northern Spain. This synoptic situation transported the warm plume of air up from Africa, across Spain and into southern England giving rise to the heatwave. This is not an unusual occurrence and even has a name: “The Spanish plume”. More precisely, this situation compares well with the modified Spanish plume synoptic situation described in Lewis and Gray (2010).    

Spanish plumes often lead to large thunderstorms over the UK, either initiated locally or imported from France. When thunderstorms are organised into a large single cloud system we call it a mesoscale convective system. The challenge is knowing when and where these storms will initiate because that initiation often depends on small-scale variations in the environment (such as due to local hills) that are not well captured by the numerical models used to generate weather forecasts. The storm that broke the heatwave in Reading was a mesoscale convective system that initiated over west London and tracked westwards as clearly shown by satellite imagery (Figure 2) and associated radar (Figure 3).

a)b) c)

Figure 2: Sequence of Meteosat Infrared Satellite imagery for (a) 14, (b) 15 and (c) 16 UTC on 12th August 2020 (copyright EUMETSAT).

a)      b)

Figure 3: Radar imagery for 12th August 2020 at (a) 1425 UTC and (b) 1615 UTC captured from www.netweather.tv/live-weather/radar. Note that the times in the panels are in BST, so one hour ahead of UTC.

So, that got me wondering whether the existence of London as a major urban area could have had a role in initiating this event. Large urban areas are known to affect their local environment in many ways including generating locally enhanced temperatures known as urban heat islands. A bit of research led me published papers examining the relationship of urban heat islands to thunderstorms. For example, a review by Han et al. in 2013 found that updraughts produced by heat islands initiate clouds, and rainfall can be enhanced by high aerosol levels due to pollution; the enhanced surface roughness associated with cities doesn’t play a major role in thunderstorm initiation, though it may affect systems passing over them. Of course, more analysis would be required to tell whether London’s urban characteristics were important in initiating this storm, and so potentially provided some predictability, in this case or whether the storm initiated over London for some other reason. Whatever the cause though, I appreciated the consequent temperature crash and excuse to do some meteorological-based web surfing.  

References:    

Han, J., Baik, J. and Lee, H. (2014) Urban impacts on precipitation. Asia-Pacific J Atmos Sci 50, 17–30. https://doi.org/10.1007/s13143-014-0016-7

Lewis, M. W. and Gray, S. L. (2010) Categorisation of synoptic environments associated with mesoscale convective systems over the UK. Atmospheric Research, 97, 194-213. https://doi.org/10.1016/j.atmosres.2010.04.001

 

Posted in Climate, Predictability, Thunder Storms, Urban meteorology | Leave a comment

Deep Water Formation In The Mediterranean Sea

Posted on 21 September, 2020 by danaallen

By: Giorgio Graffino

“The Mediterranean Sea is a small-scale ocean”, as my old teacher used to tell me. All right, that was probably a bit exaggerated. Still, it’s true that the Mediterranean Sea provides an almost unique environment to study important ocean processes, such as general circulation, air-sea interactions, and climate change, in a fairly small basin. In particular, the Mediterranean Sea is home to some of the few deep water formation regions of the World Ocean.

Deep water formation is when surface waters sink down below 2000m of depth. This occurs in small ocean regions, usually found at high latitudes. The most famous are in the Labrador Sea (between Canada and Greenland), and in the Weddell Sea (close to Antarctica). These regions are particularly important for the global climate, for they are the sinking branches of the thermohaline circulation. Key ingredients of deep water formation are: weak stratification of the surface waters, and strong air sea-interactions. These conditions are usually met during winter, when large surface heat fluxes cause a large buoyancy loss of the surface waters. The effects of this are felt at great depths (even below 2000 m). For this reason, deep water formation is also called open-ocean convection.

It is not easy to explain how open-ocean convection works, but let’s give it a try. Convection is triggered locally, in plumes with widths less than 1 km. This happens when the surface water is forced to stay in contact with low atmospheric temperatures and strong winds. There is little mass exchange involved in open-ocean convection. Rather, convective plumes help to mix temperature and salinity over the water column. This causes buoyancy loss over great depths, right down to the abyss (the region of ocean between 2000 m and 6000 m of depth).

How can these small structures mix water properties across such great vertical extent? Why do lateral exchanges (like entrainment) not dissipate these plumes away? Because of the Earth’s rotation! In fact, rotation makes convective plumes more “rigid”, which inhibits entrainment. Figure 1 shows an example of convective plumes created through an experiment using a rotating tank. Two fans at the edge of the tank force the water into the centre, which then sinks to the bottom of the tank. This process is called Ekman pumping. A dye tracer was added to the water to show the downward convection occurring in the centre, in the form of small columns. The small shapes visible in the centre of the tank in Figure 1 are the plumes seen from above.

Figure 1: Convective plumes (seen from above) in a rotating tank.

Let’s now move to our area of interest. Although the Mediterranean Sea is not among the most famous deep water formation sites, there are still four regions where deep water formation occurs (Figure 2). For the sake of simplicity, I will focus on one region. I chose the Gulf of Lion for my master’s thesis, and I will do the same here. The formation process has three phases.

  1. Preconditioning: A cyclonic gyre, formed by the action of the wind on the sea surface and by the sea floor structure, causes subsurface waters to rise to the surface
  2. Violent mixing: The Mistral and the Tramontane (cold and dry northerly winds) cause violent mixing in small convective plumes where the surface waters are weakly stratified
  3. Lateral exchange: Lateral exchange occurs between the mixed patch and the surrounding water, which restores the initial seawater conditions

As usual, timing is fundamental. Convection can only occur if the buoyancy loss is suitably intense. This means that the buoyancy losses must be concentrated in few strong events, rather than be evenly distributed over the whole winter season. Usually the preconditioning takes place during December and January, and the violent mixing phase occurs during February and March.  Lateral exchange occurs either simultaneously with the mixing phase, or shortly thereafter.

Figure 2: Deep water formation regions in the Mediterranean Sea (Pinardi et al. 2015).

How is deep water formation measured? As the vertical mass exchange is negligible, it is difficult to measure the velocity of this mass. Instead, we assess the seawater density. The convective plumes create a uniform-density patch of water. As all water masses have a characteristic density range, we can compute the amount of water formed in that density class over time, i.e. the water mass formation rate.

Unfortunately, extensive seawater density measurements are not easy to achieve in both space and time, but we can “fill the gaps” in observations with ocean reanalysis products, such as the data provided by the Mediterranean Forecasting System (MFS). Figure 3 shows the water mass formation rate in the Gulf of Lion from 1987 to 2012, calculated using MFS data. The black line shows monthly averages, which peak during winter. The red line shows the winter average (November to April) each year, and the green line shows the average February and March water mass formation rate in each year. There are big variations from year to year. That depends on how much the preconditioning phase is weakening the surface stratification, and on how intense is the atmospheric forcing.

Figure 3: Water mass formation rate (1 Sverdrup = 106 m3 s-1) in the Gulf of Lion, computed from MFS reanalysis data. The black line is the monthly average rate, the red line is the seasonal average rate and the green line is the February+March average rate. Adapted from Graffino (2015).

Deep water formation is just one example of the many processes observed in the Mediterranean Sea. To have such a rich and diverse environment right in our backyard is quite a stroke of luck for ocean and climate scientists in Europe (yes, the UK is still part of Europe). The Mediterranean Sea is currently experiencing big changes due to climate change, and its ecosystems are undergoing increasing pressure. Understanding its importance helps to protect it. So, let’s do it!

References:

 Graffino, G. (2015). A study of air-sea interaction processes on water mass formation and upwelling in the Mediterranean Sea. Master’s thesis, University of Bologna, https://amslaurea.unibo.it/8337/.

Marshall, J., and F. Schott, (1999). Open‐ocean convection: Observations, theory, and models, Rev. Geophys., 37(1), 1-64, https://doi.org/10.1029/98RG02739

Pinardi, N. and Coauthors, (2015). Mediterranean Sea Large-Scale Low-Frequency Ocean Variability and Water Mass Formation Rates from 1987 to 2007: a retrospective analysis. Prog. Oceanogr. 132, 318-332. https://doi.org/10.1016/j.pocean.2013.11.003

 

Posted in Climate, Fluid-dynamics, Oceanography, Oceans | Leave a comment