Impact case study database

The impact relates to the publication and collection of official population statistics and has four strands: (i) implementation of the AZTool team’s methods and tools by UK national statistical organisations to create official statistics for widespread use, (ii) exploitation of the team’s concepts and methods by other organisations to create complementary data products, (iii) utilisation of these data products including by central and local government and business, and (iv) use of the team’s methods and tools to plan and manage the workloads of field staff in the collection of official statistics, notably for the 2021 Census.

(i) The team’s methods and tools have been implemented by UK national statistical organisations. ONS previously employed the AZTool software and automated zone design methods to generate a completely new set of 53,578 small areas, called Workplace Zones (WZs), covering England and Wales (released Jan 2013). As a direct result of these more suitable geographical zones, in May 2014 ONS were able to publish 21 detailed tables of demographic and employment data for workers and workplaces in each of the WZs, compared to just four such tables released for Output Areas (OAs) in 2001 due to confidentiality concerns [ 5.1]. Reviewing these outputs in 2019, ONS noted that “One of the successes… was the advent of workplace zones to help provide a more appropriate geography for workplace statistics, where ONS SDC [Statistical Disclosure Control] worked closely with ONS Geography and University of Southampton to address disclosure risk” which “ allowed more to be provided publicly in all areas across the country, and this particularly benefited some of the origin-destination outputs” [ 5.2].

Following positive user feedback for WZs and strong demand for UK-wide statistics, National Records of Scotland (NRS) and Northern Ireland Statistics and Research Agency (NISRA) requested the creation of WZ boundaries for Scotland and Northern Ireland. Produced by ONS in collaboration with the AZTool team, these were published in 2016, together with associated data for WZs in Scotland. Building on the team’s prototype geodemographic classification of workplace zones for England and Wales (COWZ-EW), a full UK version (COWZ-UK) was produced and published by ONS in 2018 [ 5.3].

(ii) The reach of this impact has continued to grow, as other organisations have adopted the concepts of OAs/SOAs and WZs/COWZ and produced related data products. UK Travel to Work Areas were built from SOAs by Newcastle University/ONS in 2015 and are used in policy and planning by business, central and local government. For their 2016 Census, the Central Statistical Office (CSO) created WZs and released associated data for the Republic of Ireland [ 5.4]. In 2017, the Greater London Authority (GLA) commissioned a bespoke London Workplace Zones Classification [ 5.5]. GLA summarised the power of these new data, which can be used by organisations across London “to help with economic, transport and other planning and to identify appropriate areas for siting projects” [ 5.5]. The Health and Safety Executive (HSE) aggregated potentially disclosive business-level data from ONS’ Inter-Departmental Business Register to WZs, thus enhancing their National Population Database (NPD). “ The use of WZs has allowed HSE to undertake analysis based on comprehensive and annually updated business data which have been statistically controlled to protect data sensitivity”. This has meant that “ de-sensitised data can be shared with external partners for wider research and analysis. Outside of HSE the data have been used for natural hazard risk assessment, emergency planning, and transport planning” [ 5.6]. WZs and COWZ are embedded within Public Health England’s online Strategic Health Asset Planning and Evaluation (SHAPE) Atlas, used by NHS and Local Authority professionals to plan service delivery in health and social care.

(iii) While the most direct impact relates to the production of official statistics, by far the largest group of beneficiaries are the users of small area statistical data, who now have access to detailed information and mapping on populations both at locations of residence and workplace. Since 2014 ONS has provided an Open Geography Portal service from which users can download OA, SOA and WZ digital boundaries and lookup tables, of which there have been more than 80,000 downloads to date [ 5.1].

The new series of workplace-related data products has provided users with greatly enhanced insights into the distribution and characteristics of workers and workplaces at the small area level and is now underpinning planning and decision-making in a range of sectors. Representative examples from local government and business are included here. The City of London Corporation describes WZs as “a valuable tool for analysing workplace data at a local level and being able to clearly identify the spatial patterns and characteristics of the workforce” [ 5.7] . Hampshire County Council has grouped together WZs to permit new insights into the demographic, socio-economic, occupational and travel-to-work characteristics of workers in local ‘employment centres’ in Hampshire [ 5.8]. Greater London Authority has employed WZs and COWZ-EW to explore changes in the number of employees in WZs in London between 2009 and 2015 and to provide in-depth insights into different types of employment clusters across London at the local level [ 5.9]. Hackney Council based its 2015-2025 Transport Strategy on evidence gained from analysis of its resident and workplace populations [ 5.10]. CACI, a key commercial data provider, has generated a unique segmentation of the UK workforce based on WZs (termed Workforce Acorn) for exploitation by clients in a range of sectors [ 5.11].

New and ongoing impact of the OA hierarchy is evidenced by its adoption as a common non-disclosive geography for data aggregation and sharing between different agencies. A key example is the 2015 and 2019 Indices of Deprivation, constructed by the Ministry for Housing, Communities and Local Government as the official measure of relative deprivation in England, based on the 2011 SOAs. These “ are used by national and local organisations to identify places for prioritising resources and more effective targeting of funding” [ 5.12]

Most recently, ONS and the Joint Biosecurity Centre (JBC) combined the unique workplace and residential perspectives provided by the WZ and SOA geographies to inform national response to the COVID-19 pandemic [ 5.13]: (a) The WZ geography has provided the ideal basis for aggregation of potentially disclosive data on businesses and employees from the Inter-Departmental Business Register, permitting the estimation of industry-related COVID-19 risk. This workplace data has been combined with SOA-level residential characteristics to inform decisions about local lockdowns. (b) The SOA geography has been employed for integration of data from a wide variety of sources (including the Indices of Deprivation and industrial risk) to construct a COVID-19 risk index. These data have been linked to wastewater catchment areas to identify areas at greatest risk. This information has been used to inform targeted sampling to detect virus RNA in wastewater to provide early warnings of outbreaks.

(iv) The team’s most recent enhancements to AZTool enable the design of efficient and effective workload areas for the collection of data in the field. This functionality is now firmly embedded in ONS’ 2021 Census fieldforce management processes and is being explored for use in survey design [ 5.1]. AZTool was used to design Coordinator Areas for managing workloads of field staff for the 2019 Census Rehearsal and Interviewer Areas for the 2019 Census Coverage Survey Rehearsal. These were the precursors to the full 2021 Census and 2021 Census Coverage Survey. The ability of AZTool to rapidly and repeatedly design zones using different criteria was exploited by ONS to evaluate different sample designs and associated cost implications for the 2021 Census Coverage Survey, informing the decision to retain a two-stage sampling method based on OAs and postcodes.

Significance and long-term sustainability of impact. This is evidenced through ONS’ confirmed plans for 2021 Census design. ONS’ user consultations revealed strong demand for OAs, SOAs and WZs, and their associated data products, to be continued as 2021 census outputs. “Engagement with users on the value of the OA and WZ geographies continues to demonstrate a great deal of support for the principles of stability, comparability over time and continuity with outputs from previous censuses” and “the OA hierarchy (OAs, [SOAs] , WZs) for which census outputs will be presented will remain largely unchanged to enable comparability with both 2001 and 2011 Census results” [ 5.14]. These outputs will be produced using the AZTool team’s automated design and maintenance methods, which are now well-established in-house within ONS [ 5.1].

Beyond UK official statistics, the AZTool team has provided expert advice on automated zone design to analysts in a range of sectors, many of whom have employed AZTool for analysis. These include UK local boundary reform, utilities and sports management, cancer surveillance in the US, and design of new census output geographies for South Africa and Costa Rica.

The impact of this research was recognised in 2015 through the Royal Geographical Society’s Back Award to Martin “for influencing policy with respect to the Census and its applications”. One award is made annually for applied or scientific geographical studies which make an outstanding contribution to the development of national or international public policy [ 5.15].

The development and implementation of methods for population mapping in the absence of adequate census data has resulted in a range of impacts:

-Service delivery and public health planning in Afghanistan: Collaboration with UNFPA and the government of Afghanistan in 2015-17 produced new spatially detailed population estimates. These were presented by Tatem in Kabul to President Ghani, his cabinet, heads of UN agencies and multiple diplomats and ambassadors in late 2017. WorldPop’s estimates have since been adopted by the UN and many other international organizations in the country and are undergoing cabinet approval to become the new official government population statistics, replacing projections from the last census in 1979 [ 5.1]. The estimates have since been used continuously in planning and implementing polio vaccination in the country by the government and the WHO’s polio eradication initiative, as well as by the government and World Bank in designing new household surveys [ 5.1]. In both cases the modelled estimates have replaced 1979 projections to produce more reliable, precise and efficient approaches.

- Public health and education planning by Nigeria’s government: Similar population modelling in Nigeria has been used to plan polio elimination efforts since 2014, after the 2006 census data proved to be too inaccurate. The switch to the modelled estimates for assessment of needs, planning vaccination strategies and operational implementation contributed to successful delivery of vaccination and the elimination of polio in the country in 2015. Evidence from post-campaign coverage surveys showed that where the modelled estimates were used to plan vaccination in northern Nigeria, no areas were found to have unvaccinated children, compared to the south where children in at least ten of the sampled settlements were found to have been missed [ 5.2]. As the Nigerian National Primary Health Care Development Agency notes [ 5.2] “Outbreaks fell to zero and have stayed there. These innovations did more than help eradicate polio in northern Nigeria”. The data are currently used in the national vaccination tracking system [ 5.3], have fed into the National Surgical, Obstetrics, Anaesthesia & Nursing Plan (NSOANP) for Nigeria, and are used by the Ministry of Education (UBEC) to assess school coverage [ 5.4].

-Uptake by UNFPA and for service delivery in Colombia: The Afghanistan and Nigeria work resulted in WorldPop’s modelling approaches being adopted in the UNFPA census strategy in July 2019 [ 5.5], and the establishment of a new $40M programme, GRID3, where the WorldPop group are funded to support governments in sub-Saharan Africa with modelling and capacity strengthening. As of December 2020, this involves active support to 10 governments, covering 55% of the population of sub-Saharan Africa [ 5.6]. Moreover, through partnerships established with UNFPA, additional modelling efforts continue in collaboration with national governments. This includes the government of Colombia, where the national statistics office, DANE, was unable to complete full enumeration in their 2018 census [ 5.1], depriving many regions of the accurate data required for effective governance and resource allocation. The WorldPop group worked with DANE to develop modelling methods to impute population estimates for unenumerated areas. These were released as national statistics in 2019 [ 5.7] . In Colombia and elsewhere, the data saw “ widespread usage within governments for …health system planning, school placement and census planning” [ 5.1; Dr. Juran, UNFPA].

Following production of the high resolution gridded population datasets produced through disaggregation of administrative unit-based census or official estimates, these were made open access through the WorldPop website, and a range of collaborations and uses followed:

- Disaster relief by UNITAR, FAO and via UN humanitarian data hub: From November 2013 to present, WorldPop’s maps have consistently formed the standard dataset for the Operational Satellite Applications Programme (UNOSAT) of the UN Institute for Training and Research (UNITAR) for assessment of populations impacted by disasters and other events (e.g. the 13.5 million people affected by Cyclone IRMA-17 [ 5.8]), where no other detailed population data exist. They provide the numbers used by response and aid agencies to assess needs and scope budgets. The data also form the core population input to the FAO’s geospatial platform [ 5.8], used to plan food security and agricultural operations. From 2017, collaboration with the UN's Humanitarian Data Exchange enabled population total estimates to be produced globally for UN-recognised boundaries. In 2020, the datasets were downloaded 11,883 times by 4,779 individual users to support humanitarian operations such as aid and healthcare delivery, with 155,000 further downloads via the WorldPop web site [ 5.9].

-Training via UNFPA: In April 2016, the population mapping work was presented by Tatem to heads and representatives of all national statistical offices of the world as the keynote talk at the UN’s Commission on Population and Development [ 5.10]. WorldPop delivered training on data use at UNFPA headquarters in New York, via multiple UN regional workshops for heads of statistical offices, and multiple country capacity strengthening sessions, with cohorts of university staff within GRID3 focus countries trained to deliver future training to governments. This has included co-leading with UNFPA sessions on population modelling for over 50 representatives from the UN and governments in the Arab States region in 2018 and a similar number for the West and Central Africa region in 2019. In 2020, more than 500 government, local university and UN staff attended training run through GRID3. Dr. Juran of UNFPA notes ‘the collaborations with WorldPop have helped UNFPA build important new areas to fulfil one of its key missions in supporting population and housing censuses in all low- and middle-income countries’ [ 5.1].

- International health metrics: The datasets are widely used in the health metrics field, forming the denominator used for many African countries in the World Malaria Reports by WHO in 2016-2019 [ 5.11], the ongoing Local Burden of Disease work [ 5.12] and from 2014, UNFPA's State of the World's Midwifery series [ 5.13]. Such metrics are used to set international funding allocations, strategic priorities and for global advocacy, and reliable and consistent denominators at subnational scales are vital for their production.

-Pandemic response: WorldPop datasets have been widely used in COVID-19 pandemic response. This has included use as the demographic basis for the highly publicised Imperial College and IHME COVID transmission models [ 5.14], which led to the implementation of lockdown measures by the UK and US governments in March 2020. Moreover the datasets form the demographic basis of UNFPA’s COVID-19 Vulnerability platform and UN-OCHA’s COVID-19 Map Explorer [ 5.15]. Via Grid3, WorldPop datasets have been used by at least three African governments in their response efforts [ 5.16].

The research’s impact was recognised in 2020 through the Royal Geographical Society’s Back Award to Prof Tatem “ for leading the development of geospatial and demographic data to assist the work of public policy around the globe”. One award is made annually for geographical studies making an outstanding contribution to national or international public policy [ 5.17].

Shaping new infectious disease control and elimination strategies

In Namibia, limited information meant that the government classified its entire northern region as high risk for malaria and focussed on delivering interventions accordingly to the 1.2 million people living there. However, mapping work led by Tatem from 2013 onwards that integrated satellite, survey and mobile network data, highlighted areas of much higher risk within this northern region for precise targeting [ 5.1]. Namibia’s Mobile Telecommunications Ltd provided the Ministry of Health and Tatem’s WorldPop group with anonymous phone records for more than 2 million users – 70% of the country’s total population – between 2010 and 2014. WorldPop analysed the phone records to track call frequencies and locations, and integrated these with surveillance data to map sources and sinks of malaria parasite movements. Nine billion communications were examined to track aggregated movement patterns in regions where malaria is endemic. With support from the Global Fund and CHAI, in Oct-Dec 2013 the Ministry of Health’s National Vector Borne Disease Control Programme used WorldPop’s risk maps to provide insecticide-treated bednets to the most 80,000 at-risk individuals: ** ‘without the maps produced by WorldPop, the program would simply have reverted to standard practice’* [ 5.2]. The work therefore allowed Namibia’s government to more efficiently use limited funds by targeting interventions to the most vulnerable. Through the partnership, Namibia was able to target its distribution of insecticide-treated bed nets and allocation of community health workers to the Omusati, Kavango and Zambezi regions from October 2013. Until 2004, about 600,000 new malaria cases were reported every year. By 2016, cases had fallen 98% to 14,400 per year as the result of several malaria elimination schemes, including this one [ 5.1]. Elsewhere, subsequent human and parasite mobility models based on mobile data have been used since 2016 in Meso-America (e.g. Nicaragua) and southern Africa (e.g. Mozambique) [ 5.3]. They ‘provided an important basis for [CHAI’s] government support’ [ 5.2] via ongoing analyses of malaria importation and the mapping of sources and sinks to guide cross-border strategy .

Following the analysis and use of mobile network data in sub-Saharan Africa and the Americas, multiple major international policy and strategy documents based their recommendations around the WorldPop team’s methods and examples. These include the Aspiration2Action document co-authored by the Bill and Melinda Gates Foundation and UN in 2015 [ 5.4], which underpinned multi-billion investments in international malaria strategies, and in October 2018, the WHO Expert Reference Group Manual on Measuring Receptivity and Vulnerability for Malaria Elimination [ 5.5], which guides country governments aiming for malaria elimination. Aspiration2Action noted ‘combining parasite positivity data with modelling of human movement patterns based on census data and mobile phone call data records [CDRs] …enables malaria interventions to be coordinated for maximum impact and efficiency’. Moreover, in 2015 Tatem presented to the Global Fund in Geneva WorldPop’s analysis on southern Africa malaria connectivity, showing how strongly the region is connected by parasite flows and therefore needs to coordinate malaria elimination activities. Following this, the Global Fund provided its first multi-country regional grant to eight countries in southern Africa to accelerate progress towards malaria elimination [ 5.6]. WorldPop’s adaptation of mobile phone-based mobility methods to COVID-19 applications informed public debate over the pandemic. For example in July 2020, BBC’s Panorama used Tatem, N. Ruktanonchai and Lai’s analysis of population mobility in China in their investigation of the pandemic’s origins, and in May 2020 Chinese State Media reported on the work [ 5.7]. Moreover, the China analyses fed into the design of COVID-19 interventions used by the China CDC. Similar analyses undertaken by Tatem, N. Ruktanonchai and Lai for Europe were used in European Centre for Disease Prevention and Control guidelines provided to all European Union member states and the UK [ 5.8].

Supporting disaster relief operations

In disaster situations, displaced populations tend to be those most vulnerable and in need of aid, but obtaining data on their numbers, locations, and trends remains a major challenge in chaotic post-disaster settings. Flowminder [ 5.9] is a non-profit foundation focussed on the use of mobile network data for international development, of which Tatem is a director and in which he has been instrumental in shaping the direction of the organisation. For many years Nepal has been one of the countries with greatest potential for a devastating earthquake. Given this high potential risk, WorldPop/Flowminder and Ncell (the largest mobile operator in Nepal) agreed a collaboration in December 2014 to be able to respond to potential future earthquakes and to support long-term development objectives in Nepal. The anticipatory project was initiated and rapid response capacity set up in Kathmandu, just one week before the major M7.8 Gorkha earthquake (epicentre ~80km NW of Kathmandu) occurred in April 2015, killing 8,964, injuring 21,952 and making 3.5 million more homeless.

Following the initial earthquake, within two days the WorldPop team produced updated population density maps, including gender and age distributions for the whole of Nepal. These data were used by key relief agencies such as the World Food Program [ 5.10] and the United Nations Office for the Coordination of Humanitarian Affairs (UN OCHA) [ 5.11]. The WorldPop/Flowminder team then combined these population data with the Ncell anonymised data from 12 million mobile phones in Nepal, to quantify the earthquake’s impact on population displacement, producing first reports within four days of the initial quake, thereby providing “ vital information for directing strategies and resource allocation” [ 5.11]. These fed into the multi-stakeholder situation reports used by disaster response agencies to guide their delivery of aid to the displaced populations [ 5.10; 5.11]. As highlighted by Kurt Burja of the UN’s World Food Program [ 5.12], "When the first… report came out after the Nepal earthquake, we used it right away in our national assessment of food security. Displaced people are often the most food insecure. Getting national and district level numbers on displaced populations was thus an important component in our assessment of where to focus support after the earthquake. The (WorldPop/Flowminder) analyses were widely read and circulated in the humanitarian community during the Nepal earthquake disaster response operations." Kimberly Lietz of UN OCHA notes “This data is regularly used in producing analysis for Flash Appeals and Humanitarian Needs Overview documents as part of the Humanitarian Programme Cycle” [ 5.11] . The work was awarded the Global Mobile Award for 'Mobile in Emergency or Humanitarian Situations' in February 2016 [ 5.13], cited as “A brilliant example of how the application of big data analysis to mobile technologies can be used to accelerate emergency aid, and provide intelligence to help prepare for future disasters.” Similar analyses for Hurricane Matthew in Haiti in 2016 in partnership with the World Food Program [ 5.10] were used to guide response efforts, enabling assessment of displaced populations in need of food aid and more effective planning of delivery.

In recognition of his pioneering use of mobile data to track population mobility, the Royal Geographical Society presented its 2020 Back Award to Prof Tatem “for leading the development of geospatial and demographic data to assist the work of public policy around the globe” [ 5.14] . This annual award is for outstanding contributions to national or international policy.

Research at the University of Southampton made a key contribution to the European Copernicus programme’s capability to monitor global vegetation by (i) providing near real-time data on global vegetation chlorophyll content that are routinely distributed to users by the ESA, and (ii) establishing an operational validation procedure for vegetation monitoring products. Both of these contributions are vital to meeting the objectives of the Copernicus land monitoring services programme to provide geographical information on the status of vegetation cover to a broad range of users in the field of terrestrial environmental applications, including those in agriculture, food security and forest management.

ESA used Dash’s algorithm to develop a new vegetation data product, the OLCI Terrestrial Chlorophyll Index (OTCI). OTCI is based on the Sentinel-3 satellite’s Ocean and Land Imaging spectrometer (OLCI) instrument, and replaced MTCI after MERIS ceased operation in 2012. User demand for a replacement product, coupled with evidence of the Southampton algorithm’s robustness, led ESA to include the ‘red edge’ detection capability offered by OTCI within the mission requirements document for the Copernicus Sentinel-3 satellite programme. Since Sentinel 3 was launched in October 2016, ESA has used OTCI as a near real-time, global data product on vegetation chlorophyll content. Dash’s group authored the user guide and algorithm description [ 5.2]. The OTCI is currently (February 2021) the only near-real time, global operational product estimating terrestrial vegetation chlorophyll content. An operational data product is one that ESA has committed to deliver long-term and in near-real time, guaranteeing users its continued and timely production. It is distributed to users free of charge through the Copernicus programme. As ESA states ‘the unique capability of the OTCI algorithm to exploit the red edge position, thus providing accurate information on very high level of chlorophyll contents which cannot be performed by other sensors, put Europe and the European Space Agency and Copernicus Programme in a unique and leading position for monitoring the state of vegetation all over the world” [ 5.1].

During 2017 alone, 2.44 pebibytes (> 1000 terabytes) of Sentinel 3 OTCI data were downloaded from Copernicus’ dissemination system [ 5.3]. Copernicus programme users have deployed OTCI to monitor and manage terrestrial ecosystems, including estimating forest and crop health. As ESA stated “ The innovative product (OTCI) that was developed by Dash’s team now contributes to an overall observation system that allows us to better understand how our planet is evolving and how it is responding to climate change and at the same time allows us to better understand climate change itself” [ 5.1] .

To ensure the quality of their operational data products meets the expected user requirements, in 2016 the European Commission established a new service called the Ground-Based Observations for Validation (GBOV). Southampton’s research on the ground-based validation of satellite-derived vegetation data products [ 3.3, 3.4] led to Dash’s group writing an Algorithm Theoretical Basis Document for the European Commission specifying validation procedures for three other vegetation products – leaf area index, LAI; fraction of vegetation cover, FVC; and fraction of absorbed photosynthetically active radiation, FAPAR [ 5.4]. Dash’s group led the validation of the vegetation component [ 5.5] and from 2017, conducted one of the largest and most comprehensive ground validation exercises globally, processing ground data across the world to determine the products’ accuracy as part of ESA’s Sentinel-3 Validation Team. In 2017, ESA noted ‘Feedback from the Sentinel-3 Validation Team meeting provided essential information to ESA and EUMETSAT to progress with the evolution and improvement of Sentinel-3's core data products’ [ 5.5]. The validation data are now being routinely used by the European Commission (EC) to ensure the operational products meet user requirements. The EC has noted that they ‘… reproduce well the changes due to drought and wet conditions in Europe, as well as the realism of the temporal variation over specific events around the globe’ [ 5.6]. Since 2017, the EC has delivered these three products – LAI, FVC, FAPAR – through the Copernicus programme to users worldwide. They are used by more than 6,000 organisations and users worldwide for a wide range of applications ranging from monitoring biodiversity to drought surveillance [ 5.7]. For example, the EC’s Joint Research Centre uses Copernicus FAPAR and LAI product as key inputs to the Europe-wide Monitoring Agriculture ResourceS (MARS) crop early warning programme [ 5.8], which supports the Common Agricultural Policy and food security assessments worldwide.

In addition to ensuring the quality of the three satellite-based vegetation products [ 5.4], the Copernicus programme has distributed the ground-based GBOV validation data set produced by Dash’s group since October 2018. The GBOV data are the largest ground data set for in-depth quality assessment of any satellite-derived vegetation product. Since their release in 2019, they have been downloaded by more than 300 users worldwide [ 5.9]. Through such downloads, users have applied Southampton’s ground data to accuracy assessment of other satellite-derived vegetation data products (e.g. those from NASA’s Moderate Resolution Imaging Spectroradiometer (MODIS) sensor) alongside the data products distributed through the Copernicus programme [ 5.10].

Meanwhile, users have also benefitted significantly from implementing Dash’s chlorophyll algorithms independently of ESA. For example, in 2018 the World Bank reviewed the performance of algorithms for their prediction of crop yields from satellite data and recommend use of MTCI for countrywide maize crop yield estimates in Uganda, noting ‘the superior performance of MTCI is noteworthy’ [p. 17] and ‘The large boost in performance when using MTCI with Sentinel2 therefore more than outweighed any loss in accuracy from using coarser resolution’ [ 5.11].

The MTCI and OTCI algorithms developed at the University of Southampton thus formed the basis for innovative and robust data products for global vegetation monitoring which help the European Commission to deliver part of its flagship Copernicus programme. These data products provide accurate and timely information for users to monitor and manage both crops and our terrestrial biosphere.