Understanding Met Office Climate Data

 A selection of Met Office weather and climate data is available in this portal. It consists of:

1.       Past weather observations

The Met Office has weather observations which have been collected from thousands of locations around the world over long periods of time. Gridded weather observations covering two 30-year periods are available to access via this portal.

2.       Future climate projections

The Met Office uses climate models to understand what large-scale changes in the global climate we could expect to see in the future. These are called future climate projections. This portal contains future climate projections for several emission scenarios or future warming levels. These climate projections cover either the UK or the entire globe and are available for various future time periods.

In this portal, both observational data and climate projection data are provided on a grid; each point on the grid is called a cell. The grid cells are three-dimensional and have been created by dividing the world up into boxes (see image below). In different models these grid cells are different sizes. This is described as the resolution: the higher the resolution, the smaller the grid cell. Higher resolution models provide more regional detail but take a longer amount of time to execute and require more computer power.  

The resolution of the datasets in the portal vary. Those with a global coverage are around 60km resolution. Datasets with just UK coverage have a higher resolution of 12km; some other datasets for the UK have an even higher resolution of 2km.

The following observational data are available on this portal:

• Temperature
• Precipitation

The following climate projection data are available on this portal:

• Temperature
• Precipitation
• Sea level

Additionally, UK socioeconomic data is available in the portal, divided by local authority or groups of authorities. This data helps to explain how the economy and society might evolve during the remainder of the 21st Century.

Gridded weather observations covering two 30-year periods, 1981-2010 and 1991-2020, are available in this portal. Climatologists refer to data summarised over 30 years as a 'climate reference period' and consider this to be of sufficient length of time to capture the natural year-to-year fluctuations in weather conditions, for example, warm years, cool years, and average years. If too long a period is selected, e.g., 60 years, the harder it is to see any signal of a changing climate.

• Global data for  1981-2010 , from the CRU TS v4.06 dataset.

• UK datasets for  1991-2020 , from the HadUK-Grid dataset.

•  CRU TS v4.06 

CRU TS is a monthly high-resolution global climate dataset containing information from 1901-2019 on a 0.5° x 0.5° grid (approximately 60 km x 60 km grid close to the equator). It covers all land surfaces apart from Antarctica and is produced by the UK’s National Centre for Atmospheric Science (NCAS) at the University of East Anglia’s Climatic Research Unit (CRU). CRU TS consists of ten variables, all based on near-surface measurements: temperature (mean, minimum, maximum and diurnal range), precipitation (total and rain day counts), humidity (as vapour pressure), frost day counts, cloud cover, and potential evapotranspiration. A selection of these variables are available on this portal.

•  HadUK-Grid 

HadUK-Grid provides climate variables from the network of UK land surface observations which have been interpolated onto a uniform grid. The data sets cover the UK and are available at a range of resolutions from 1km x 1km resolution to 60km x 60km resolution to enable users to compare observational data with climate projections. The HadUK-Grid data have also been aggregated across UK countries, administrative regions and river basins.

Gridded data have been produced on daily, monthly, seasonal and annual time scales as well as long term averages for a set of 30-year climatological reference periods. Variables include air temperature (maximum, minimum and mean), precipitation, sunshine, mean sea level pressure, wind speed, relative humidity, vapour pressure, days of snow lying, and days of ground frost. A selection of these variables are available on this portal.

It is not recommended that single grid cell values are used in isolation. This is because within the individual grid boxes, detail is lost (e.g., a mountain and a valley within the same grid cell will be assigned the same value). This can be seen in the image below. Additionally, whilst the 2km grid resolution shown in the image below will have more detail, the data will have a higher interpolation uncertainty. Interpolation uncertainty occurs because gridded observation data is generated from a network of weather stations which are not uniformly spread out over an area. In places where there is no weather station, the dataset is completed by calculating approximate conditions relative to neighbouring weather stations. In higher resolution datasets, a larger number of grid cells rely on this interpolated data, increasing the uncertainty.

We recommend the use of multiple grid cells or an average of grid cells around a point of interest to help users get a sense of the variability in the area and potential extremes. This will provide a more robust set of values for informing decisions based on the observations or future climate projections.

Climate projections are simulations of the Earth’s future climate which demonstrate how the climate may change in the future. They are created by running complex numerical computer models (called “climate models”) of the Earth’s climate on  supercomputers  (such as those at the Met Office). The projections are based on assumed futures or ‘scenarios’ which include projections of the concentrations of greenhouse gases, aerosols and other atmospheric components that affect the amount of radiation that the Earth receives. The amount of radiation the Earth receives is described as the planet's ‘radiative balance’, which is the sum of the planet’s incoming and outgoing radiation and controls the temperature of the Earth. Global warming occurs when Earth receives more radiation than it emits back to space.

We recommend the use of multiple grid cells or an average of grid cells around a point of interest to help users get a sense of the variability in the area and potential extremes. This will provide a more robust set of values for informing decisions based on the observations or future climate projections.

The climate projection datasets in the portal fall into two categories:

1. Emission scenarios
2. Warming levels

You can find more information on these two categories in the sections below.

The future climate projection data in the portal are from the  UK Climate Projections (UKCP18).  UKCP is a set of tools and data that show how the UK and global climate may change in the future. The data covers different domains (global or UK only), resolutions and emission scenarios.

More information on the data availability and formats can be found  here .

The future climate projections in this portal draw upon the following UKCP products:

In order to predict what the climate might be in the future, it is necessary to make assumptions about the economic, social and physical changes to our environment that will influence how the climate might change.

Representative Concentration Pathways (RCPs) are a method to capture these assumptions within a set of scenarios. RCPs specify concentrations of greenhouse gases that result in the total radiative forcing (difference between incoming and outgoing radiation at the top of the atmosphere) increasing by a target amount by 2100 relative to pre-industrial levels.

Radiative forcing targets for 2100 used in the Intergovernmental Panel on Climate Change (IPCC) 5th Assessment report (AR5) were 2.6, 4.5, 6.0 and 8.5 Wm ^-2  to span a wide range of equally plausible future emission scenarios. These targets are incorporated into the names of the RCPs; RCP2.6, RCP4.5, RCP6.0 and RCP8.5. Each pathway results in a different range of global mean temperature increases over the 21st century. The graph below shows the projected warming to 2100 (for the four RCP scenarios) and to 2300 (for three of the RCP scenarios).

The table below lists each RCP and the projected increase in global mean surface temperature averaged over 2081–2100 compared to the pre-industrial period (1850 – 1900). The table also includes a high-level overview of each scenario to help you select the most appropriate RCP for your question.

The RCP pathways represent a broad range of plausible climate outcomes and are neither forecasts nor policy recommendations. They include a wide range of assumptions regarding population growth, economic development, technological innovation and attitudes to social and environmental sustainability. Each pathway can be met by a combination of different socioeconomic assumptions. More information on RCPs is available  here .

The emission scenario datasets available in the portal are:

• UK datasets for 2050-2079

• Global datasets for 2040-2069 and 2070-2099

Future sea level projections

Sea level rise is the dominant driver of coastal flood risk in the UK. UK sea level rise is expected to increase with higher emission scenarios. However, different areas around the UK will experience different amounts of sea level rise.

Future sea level projections have been produced for four different emission scenarios at annual time scales relative to a baseline period (1981-2000) as part of UKCP18. The projected sea level rise is available for a range of percentiles, e.g., low end (5th percentile), middle of the range (50th percentile) and high end (95th percentile). The portal contains a selection of this data for three emission scenarios for the start of each decade for three percentiles.

The datasets are:

•  Time-mean sea level projections for 2010-2100 
•  Exploratory time-mean sea level projections for 2010-2300 

These datasets are averaged values of time-mean sea level change (5th, 50th and 95th percentile) relative to a 1981-2000 baseline for RCP2.6 (i.e., aggressive mitigation), RCP4.5 (medium stabilisation) and RCP8.5 (high emissions) pathways at the start of each decade (sea level rise at 2010, 2020, 2030…). The sea level data are on an approximately 12 km (WGS84) grid and cover the entire coastline of the British Isles.

Please note that sea level datasets for RCP6.0 are not currently available in the portal.

Instead of considering future climate change during specific time periods (e.g. decades), some of the datasets on the portal are calculated at various levels of global warming relative to the pre-industrial (1850-1900) period. The world has already warmed by around 1.1°C (between 1850–1900 and 2011–2020). The warming level datasets allow for the exploration of greater levels of warming.  

The global warming levels available in this dataset are 1.5°C, 2°C, 2.5°C, 3°C and 4°C. The data at each warming level was calculated using a 21 year period. These 21 year periods are calculated by taking 10 years either side of the first year at which the global warming level is reached. This time will be different for different model ensemble members.

We cannot provide a precise likelihood for particular emission scenarios being followed in the real world future. However, we do note that RCP8.5 corresponds to emissions considerably above those expected with current international policy agreements. The results are expressed for several global warming levels because we do not yet know which level will be reached in the real climate as it will depend on future greenhouse emission choices and the sensitivity of the climate system, which is uncertain. Estimates based on the assumption of current international agreements on greenhouse gas emissions suggest a median warming level in the region of 2.4-2.8°C, but it could either be higher or lower than this level.

The warming level datasets on the portal are grouped into three categories:

1.       Annual counts of days above a threshold temperature for key variables.

2.       Annual degree days

3.       Changes in precipitation and surface temperature compared to a baseline.

For each category, the counts or changes are calculated from the  UKCP18 regional projections  (12km resolution) using the RCP8.5 high emission scenario. Three values ('upper', 'median', and 'lower') have been calculated for each grid cell for each warming level.

Examples of some of the warming level datasets available in the portal:

• Tropical nights, summer days, icing days and frost days have units of days per year and measure how many times the threshold is exceeded (not by how much). A maximum daily temperature of 27°C contributes the same amount to the Annual Count of Summer Days as a day at 35°C.

• Cooling degree days, growing degree days and heating degree days have units in degree days. They measure the annual sum of the number of degrees above or below a temperature threshold. For example, if the daily average temperature is 6°C, this would contribute 0.5 Growing Degree Days, whereas if the daily average temperature is 10.5°C, this would correspond to 5 Growing Degree Days. So for one year, the number of growing degree days, heating degree days or cooling degree days can exceed 365.

• The drought severity index is not threshold based. Instead, it is calculated with 12-month rainfall deficits provided as a percentage of the mean annual climatological total rainfall (1981–2000) for that location. It is therefore a measure of drought severity, not frequency, and higher values indicate more severe drought. Twelve-month accumulations have been selected as this is likely to indicate hydrological drought - water scarcity over a long period of time. These droughts can heavily deplete water resources on a large scale as opposed to meteorological or agricultural drought, which generally occur on shorter timescales of 3-12 months. However, this categorisation is not fixed, because rainfall deficits accumulated over 12 months could lead to different types of drought and drought impacts, depending on the level of vulnerability to reduced rainfall in a region.

Climate models are numerical representations of the climate system. To capture uncertainty in projections for the future, an ensemble, or group, of climate models are run. Each ensemble member has slightly different starting conditions or model set-ups. Considering all the model outcomes gives users a range of plausible conditions which could occur in the future.

In the portal the future climate projections use either 'lower, median and upper' or percentiles to characterise the uncertainity in the projections for a particular scenario or warming level.

What does 'lower, median and upper' mean?

In the portal, the metrics of interest (e.g., Summer Days) are calculated for each ensemble member which helps to capture the uncertainty. The ensemble members are then ranked in order from lowest to highest. For example, the UKCP18 regional (12km) and global (60km) climate projections have 12 and 15 ensemble members respectively. For both, the lower, median and upper values are:

• The 'lower' field is the second lowest ranked ensemble member
• The 'median' field is the central value of the ensemble
• The 'upper' field is the second highest ranked ensemble member

This spread of outputs can be used to infer the uncertainty in the projections. The larger the difference between the lower and higher fields, the greater the uncertainty.

Using the median value may not suit all users. Users who are risk averse may choose to use the upper value, and those who are risk-tolerant may choose to use the lower value. You may also consider using a combination of the three values depending on your use case. To help decide which is the most appropriate data to use you can click on any grid box to see all the values.

What is a percentile?

Instead of ranking the ensemble members to assess uncertainty, time-mean sea level projections use the 5 ^th , 50 ^th , and 95 ^th  percentiles.  Each ensemble member is used to construct a distribution of possible outcomes called a model distribution. From this distribution, percentiles are used to characterise the uncertainty in the model projections.

For example, the 50 ^th  percentile value means 50% of the model distribution is below that value. It represents the central estimate (median) amongst the model projections. The 95 ^th  percentile value means 95% of the model distribution is below that value and similarly the 5 ^th  percentile value means 5% of the model distribution is below that value. The range between the 5 ^th  to 95 ^th  percentiles represent the projection range amongst models and corresponds to the IPCC AR5 “likely range”. It should be noted that, by definition, there may be a greater than 10% chance that the real-world sea level rise lies outside this range

A new hospital is going to be built and the developer wants to check the projected temperature for the location to ascertain the level of air conditioning required. The example map below shows the projected July average temperatures (2050-2079 median for the high emissions scenario, RCP8.5) for Devon and Cornwall with existing hospital locations overlaid as circles. Red circles are NHS sector hospitals and blue circles represent independent hospitals.

The UK Shared Socioeconomic Pathways (SSPs) are five different storylines of projected future socioeconomic scenarios specifically for the UK between the years 2020 and 2100. They were developed by the  UK-SSP project  to be consistent with the global SSPs used by the IPCC community which explain how the global economy and society might change over that period. These societal elements are crucial for determining challenges related to mitigation and adaptation. The UK-SSPs are created independent of climate change and climate change policy and so don't consider the potential impact climate change has on societal and economic choices.

A  separate StoryMap explores UK SSPs  in more detail and includes examples of ways that the data can be used, and it is recommended that this is consulted prior to utilising the data.

Decadal estimates for each of the  socioeconomic indicators  listed in the table below are available for the period 2020-2100 in the UK-SSP datasets. These datasets are available as GIS polygon shapefiles split by local authority or groups of local authorities, except for  Population  which is available as a gridded dataset.

Demography, Life Expectancy and Rail Infrastructure use the Office for National Statistics Local Authority District (ONS LAD) boundaries. Whereas, Inequality and Social Cohesion use the Office for National Statistics Nomenclature des Unités territoriales statistiques (which translates to Nomenclature of territorial units for statistics) level 3 boundaries (ONS NUTS3). You can find more information on NUTS here.

Results presented for emission scenarios and global warming levels both have their advantages and disadvantages, depending on the application.

Emission scenarios

Emission scenarios are based on narratives of how global carbon emissions will evolve in the future as drivers such as population and GDP vary. Highlighting the potential future climate under different emission pathways is relevant for determining how ambitious Nationally Determined Contributions (NDCs) should be. They are particularly relevant for understanding how ambitious mitigation targets need to be to limit warming to 1.5°C and well below 2°C in the Paris Agreement. Additionally, for adaptation planning that evolves over time, such as in dynamic adaptation pathways, the emission scenario approach can be a more suitable choice.

Warming levels

You may wish to choose warming levels when you want to see how various temperature and precipitation variables are affected by a defined global temperature increase above a pre-industrial level, e.g., 1.5°C, 2°C, 2.5°C, 3°C and 4°C but do not wish to make a judgement on which emission scenario might be followed. Warming levels can have significant policy relevance as they allow us to examine the climate metrics at different ‘levels’ of global average temperature change. We have already experienced global warming of around 1.1°C above pre-industrial levels (between 1850–1900 and 2011–2020) and it is important to assess the different climate metrics as further changes are expected in the future.

Yes, some UK and global observation datasets are directly compatible with climate projection datasets as they have the same grid resolution and map projection:

1.  Monthly and annual averages of precipitation and surface air temperature weather observations for the period 1991-2020  can be directly used with  RCP8.5 climate projections for 2050-2079  for the UK on a 12 km BNG projection grid.

2.  Monthly averages of precipitation and mean, maximum and minimum surface air temperature observations for the period 1981-2010  can be used with  RCP2.6 climate projections for 2040-2069 and 2070-2099 . The data is global (roughly 60 km grid). Projections are given as lower, median, or upper.

Surface air temperature refers to the temperature of the air at a height between 1.5 and 2 metres above the surface.

Surface air temperature in a climate model refers to the temperature of the air at a height of 1.5m above ground level and shielded from direct solar energy and precipitation. It is subtly different to the definition used at weather observation sites where surface air temperature is measured between 1.25 and 2 metres above the surface.

Precipitation is defined as any liquid or frozen water that forms in the atmosphere and falls back to the Earth. It comes in many forms such as rain, hail, sleet and snow. Rainfall is defined as the total liquid product of precipitation or condensation from the atmosphere.

Note that the terms rainfall and precipitation are used interchangeably in the portal.

There are two time-mean sea level rise datasets in the portal. One which has  projections out to 2100  and one which is more  exploratory with projections out to 2300 .

Both datasets have the following limitations:

• We cannot rule out substantial additional sea-level rise associated with ice sheet instability processes that are not represented in the UKCP18 projections, as discussed in the recent  IPCC Sixth Assessment Report (AR6) .
• For any given geographic location, there may be other vertical land motion processes at play that are important when carrying out coastal risk/impacts assessments.

The  projections out to 2300  are included because current levels of global warming have already committed us to certain amounts of sea level rise beyond 2100 (i.e., multi-century). As the ocean is slow to respond to changes in the atmosphere, even if global average air temperature stops rising (i.e., global emissions are reduced), sea level will continue to rise well beyond the time changes in global average air temperature level off or decline.

However, these sea level projections out to 2300 have a greater degree of uncertainty than the projections out to 2100 and should therefore be treated as illustrative of potential future changes. For this reason, they have been designed to be used alongside the  21st century projections  for users that are interested in exploring post-2100 changes. They might be suitable for long term planning; however, the limitations above and the larger degree in uncertainty should be taken into account.

There are several ways to visualise global data on a 2D map, depending on the application. In the portal, two map projections (also known as coordinate reference systems) are used.

The World Geodetic System 1984 (WGS84) is a 3-dimensional coordinate reference system for establishing latitude, longitude and heights. WGS84 represents the best global geodetic (i.e., relates to coordinates on a curved surface) reference system for the Earth available at this time for practical applications of mapping, charting and navigation.

The Ordnance Survey British National Grid (BNG) is a Cartesian coordinate system. It represents coordinates as distances on a flat (planar) surface from a fixed point, called the origin. BNG is commonly used for mapping and surveying purposes across the United Kingdom.

Note: When viewing the data in the portal in map view, the dataset will be automatically reprojected to be displayed on the WGS 1984 Web Mercator (auxiliary sphere) map projection. This is slightly different to WGS84; WGS 1984 Web Mercator is used as this is what the mapping in ArcGIS Online is configured to use. However, if the data are downloaded, they will be in either WGS84 or BNG. Note there is not an option to select between downloading the data on a WGS84 projection or a BNG projection. Only one option will be available per dataset. This is an important consideration for which datasets can be used together.