Improving Availability, Access and Use of Climate Information

by Tufa Dinku1, Kidane Asefa2, Kinfe Hilemariam2, David Grimes3 and Stephen Connor4

Climate variability and change are serious challenges to sustainable development in Africa. The current famine crisis in the Horn of Africa is yet another reminder of how fluctuations in the climate can destroy lives and livelihoods. Ethiopia, one of the countries impacted by the current drought, has been suffering from climate fluctuations for decades. Climate variability has been one of the major hindrances to development in Ethiopia. Droughts and floods have reduced Ethiopia’s annual growth potential by more than one-third (Grey et al., 2006).

Figure 1 — Comparison of GDP growth rate and normalized June–August rainfall averaged over the country. Most of the severe droughts resulted in significant fall in GDP growth. There is a one-year lag between the occurrence of drought and fall in GDP growth. The relationship is not always that simple as many factors are in play.

The 1984–1985 drought reduced Ethiopia’s agricultural production by about 21 per cent with a 9.7 per cent fall in the Gross Domestic Product (GDP) (World Bank, 2006). The effect of climate variability is, of course, felt more by poor households (Stern, 2007). The 1998–2000 drought cost each household over north-eastern parts of the country more than 75 per cent of its average annual income in crop and livestock losses (Carter et al., 2004).

Credible information

Building resilience against the negative impacts of climate and maximizing the benefits from favourable conditions will require the design and implementation of effective climate risk management strategies. This cannot be accomplished without the availability of decision-relevant climate information. Credible information about the past climate, recent trends and swings, likely future trajectories and associated impacts is very important for climate risk management. Long-term climate time series are critical in many applications including:

  • Putting observed and anticipated climate into context;
  • Assessment of climate-related risks under current conditions;
  • Understanding and modelling of climate impacts on different socio-economic activities; and
  • Improving predictions at different time and spatial scales.

Climate information has been used in Ethiopia for decades, particularly for drought monitoring and early warning (Hellmut et al., 2007). Yet, the avail­ability, access and use of climate data are far from ideal. The main source of climate data is the network of weather stations managed by Ethiopia’s National Meteorology Agency (NMA). Even though the number of stations is reasonably good, and has been increasing (Figure 2), station distribution is very uneven. There are very few stations over the lowlands. Most of the stations are located in cities and towns along the main roads. This limits the availability of climate information and services for rural communities. Where records do exist, they frequently suffer from data gaps and poor quality and are often not easily accessible. This, in turn, has limited the use of the available climate data.

Figure 2 — Trends in the number of raingauge stations over the years 1981–2008. Only stations with data for 10 years or more are represented here. The decrease in the number of stations during 1991–1992 is because observations were interrupted during government change. The decrease during the later years is due to delayed reports from regional meteorological offices.

If the problems of data availability, access and use could be overcome it would enable the effective and efficient use of climate information in Ethiopia. The problem of data availability could be significantly improved by combining station observations with globally available products such as satellite proxies and model reanalysis data. The main advantage of the global products is the excellent spatial convergence. These data are available over most parts of the world at increasingly improved spatial and temporal resolutions. Satellite rainfall estimates now go back 30 years. The combination of ground-based observations with satellite and/or model information should therefore help to overcome the spatial and temporal gaps in station data while improving the accuracy of the global products. This will alleviate the inadequacy of climate data, particularly for rural Ethiopia.

Data availability may not necessarily lead to data access, so improved access to climate information and services should be provided. One approach to this would be to make data, analysis tools and targeted products available through the Internet. However, even good access tools and high quality data may not guarantee effective use of climate information.


Figure 3 – Rainfall products for the second 10-day period in April 1996. The top-left panel is the raingauge data for that 10-day period, while the top-right panel shows satellite estimate. The lower-left panel is interpolated raingauge, and the lower-right panel is combined raingauge and satellite data. The interpolated gauge follows the overall spatial structure of rainfall as depicted by gauge data, but with significant smoothing.

The main problem with the gridded data is that values over the north-eastern and north-western lowlands, where there are few or no gauges, are significantly overestimated or underestimated. The satellite estimate depicts the spatial structure of the rainfall very well, but significantly underestimates high rainfall values. The combined product overcomes, to a degree, the shortcomings of both the interpolated gauge and satellite estimates.

Figure 4 – Maximum temperature for second 10-day period in April 2000. The top-left panel is station data, while the top-right panel is interpolated station data. The bottom-left panel is station data combined with 10-day period averages of MODIS LST and elevation. The bottom-right panel has included topography for reference.

The station-only product significantly underestimates temperature over the data-sparse lowlands because of the influence of data from the neighbouring highland areas, while the combined product does not have this problem.

Appropriate use requires knowledge within the user community of what information is available and how it might be used. A facilitated dialogue between meteorologists and the user community would be of great benefit. This should include training the user community to understand, demand and use climate information, as well as training climate scientists to understand the needs of the users.

This three-track approach of simultaneously improving data availability, access and use is being implemented in Ethiopia in collaboration between NMA and the International Research Institute for Climate and Society (IRI) at Columbia University in the United States. This project, funded partly by, has three major components:

  • Improving data availability by quality controlling station data from the national observation network as well as integrating station data with the best satellite products.
  • Improving access via an online facility installed at NMA allowing query, visualization and downloading of data and information products.
  • Improving the use of climate information through strengthening the capacity of the user community to understand and use existing and new information products.

Improving data availability

Data availability is improved by filling spatial and temporal gaps in climate observations. The spatial gaps are a result of sparse station network, while temporal gaps are due to interrupted observations or lost data due, for example, to communication problems. Cleaning national climate observations and combining them with satellite proxies could help to fill these gaps.

This approach has been implemented in Ethiopia to generate a 30-year time series of rainfall and temperature data at 10 daily timescales for every 10-km grid over the country. The combined rainfall dataset draws on more than 600 raingauge stations merged with 30 years of satellite-derived rainfall estimates. For temperature, data from over 300 stations are combined with Moderate Resolution Imaging Spectroradiometer (MODIS) satellite Land Surface Temperature (LST) estimates.

The project has collaborated with the Tropical Application of Meteorology using the Satellite (TAMSAT) group at the University of Reading, United Kingdom, which is generating a 30-year time series of rainfall estimates for Africa based entirely on data from the Meteosat satellite calibrated against local gauges. These estimates are combined with the observations from over 600 raingauges. The result is a unique, high-quality data set, which has been shown to be more accurate than any other long-term, satellite-based time series (Dinku et al., 2011). The main strengths of this data set include the following:

  • The satellite rainfall estimates are calibrated specifically over Ethiopia. The relatively high density of gauges facilitated a highly detailed calibration, which takes account of topographic and seasonal variations.
  • The satellite estimates are produced using a single satellite algorithm and the same satellite information going back 30 years. This makes it a very consistent time series. This consistency is critical for detecting climate trends, amongst other things.
  • The satellite estimates are combined with contemporaneous data from the same set of 600 stations. This improves the quality of the rainfall estimates significantly. This emphasizes the importance of the local approach to the rainfall estimation. Other satellite products that use raingauge observations from Ethiopia have access to 20 stations at most.
  • The combined data covers the whole country. This improves data availability particularly for rural communities where there are no weather stations.
  • The combined data are in a format that is easily imported into GIS (geographic information system) browser software for combination with other data of interest.
  • NMA staff members who had been trained on data quality control, satellite rainfall and temperature retrievals, as well as combining station and satellite data did most of the work at NMA. This ensured the continuity of the project.

Improving access

One way to improve access to climate information is to make data, tools and targeted products available through the Internet. This may not be enough to reach all potential users, but would be a good starting point. The main purpose is to reach those who can then reach others by other means.

The current project has redesigned NMA’s Website for better presentation of its existing products and services and for the delivery of product derived from the new data set. The new website (Figure 5) is designed in such a way that users can easily locate the different products and services offered by NMA. It includes a Climate Analysis and Applications Map Room.

This Map Room has five parts: Climate Analysis, Climate Monitoring, Climate and Agriculture, Climate and Water, and Climate and Health (Figure 5). These Map Rooms are created by IRI using IRI Data Library (IDL) tools, and then transferred to NMA. The Climate Analysis and Climate Monitoring Map Rooms have been completed, while the others are still under construction. These will be completed in consultation with the user community in each sector.

screenshot of website
Figure 5 – NMA’s improved Website, designed by IRI and developed by a local company, is designed in such a way that the user can find information easily. It presents existing and new products from simple station history to more sophisticated maps. It also makes locating and ordering data sets easier. The sector-specific Map Rooms on the right facilitate the use of climate information.

The Climate Analysis Map Room provides information on the mean climate (in terms of rainfall and temperatures) at any point or for any administrative boundary (Figure 6). It also shows the performance of the rainfall seasons over the years as compared to the mean. This kind of information service is unprecedented in the whole Africa. The authors are not aware of any national meteorological services that provide this kind of information. But Ethiopia’s NMA offers more than a look at the past climate. Its Climate Monitoring Map Room enables monitoring of the current season. Different maps and graphs compare the current season with the mean or recent years (Figure 6).

Figure 6 — A sample of the Climate Analysis Map Room. The top-left map shows monthly climatology for June, the bar graph is rainfall climatology for the administrative region indicated in red, while the line graphs are maximum and minimum temperatures. The bar graph at the lower left shows the performance of rainfall seasons over the years as compared to the mean.

This information could be extracted at any point or for any administrative boundary. Data are updated every 10 days, thus enabling close monitoring of the season. Extracting and presenting information at any administrative level enables focusing on a specific area of interest. The services offered by NMA will be improved further when the other three Map Rooms are completed. These Map Rooms will enable the targeting of individual products by users.

The combination of quality controlled higher resolution climate products derived from station observations and satellite proxies will help to overcome a critical impasse in current data dissemination policies. NMA is usually very protective of their databases.

The approach presented above maintains the control of the national observations by NMA while enabling the dissemination of a wide array of value-added climate information products targeted to specific user needs.

NMA is also in the process of formulating a data policy to make the combined time series freely available for research and other non-commercial purposes. Once available, it is envisaged that an expanded array of new products could be created to answer the specific needs of a wider range of stakeholders.

The products described above will provide basic information to aid climate risk management in a number of climate-sensitive sectors. These include vital sectors such as agriculture, food security, water resources and health. The project is expected to culminate with an online, dynamic and digital Climate Atlas of Ethiopia, which will offer wider access to the generated climate information.

However, all these may not be enough to ensure the appropriate and effective use of these products. Targeted training should be provided to the user community at all levels so as to understand and use the information products. The project has already trained a number of experts from the health sector in the use of climate information. This training included an intensive two-week summer course at IRI for seven health and climate professionals, as well as a couple of training courses in the country. The IRI scientists and Ethiopians who attended the IRI summer courses conducted the local trainings. NMA is also planning to train users at different levels and backgrounds on the use of its new products and services.

Figure 7 — A sample of the Climate Monitoring Map Room. The four plots compare the performance of the current season with respect to the mean and recent years. The information is extracted for the locality shown on the map.

Lessons learned

This is the first project of its kind and was only a concept when presented to NMA management. The project is a success only because NMA bought into the project with an open mind.

NMA staff did most of the work. They were first trained on different aspects of the project work. All the training courses were conducted at NMA, which enabled many more people to participate. The other benefit of the training was that it was targeted towards specific tasks.

Obtaining and processing 30 years of raw satellite data, which covers the whole of Africa, was a time consuming but very important task.

Calibration of the satellite rainfall retrieval algorithm using all locally available raingauge data significantly improved the quality of the estimates. The algorithm used in this project was a very simple one and it only used thermal infrared data, yet it outperformed more sophisticated algorithms that incorporate passive microwave data.

Combining station observations with global products such as satellite proxies can significantly alleviate the problem of spatial and temporal gaps in station observations.

Improving data availability by itself will not guarantee the use of climate information in development practices; extra efforts are needed to make the information accessible and useful.

The Map Rooms were built using IDL tools, which were then transferred to NMA. The IDL enables users to access and analyse hundreds of data sets and download results in different formats. The main accomplishment was the transfer of the IDL tools to NMA without the data so that it could be used with NMA’s own data. This will now enable NMA to create more products from its own and other data sets.

training sessions

A promising template

The Ethiopian experience is a promising template for improving climate services throughout Africa. It should be improved and scaled up to all countries in Africa, adapting it to national needs and circumstances. Carrying out similar projects in other countries should now be cheaper and faster because of three main factors:

The raw satellite data obtained and processed for Ethiopia covers the whole of Africa.

Methodologies and computer codes developed for Ethiopia can easily be adapted for another location.

The portable IDL tools on which the Ethiopian Map Rooms were built can easily be adapted for any another country.

The task that may take time and resources is the organization and quality control of station data. This varies from country-to-country depending on the level of data organization. Some countries may need to digitize their data, or even rescue some of their data.


This work has been funded by and a grant/cooperative agreement from the National Oceanic and Atmospheric Administration (NOAA), NA050AR4311004. The views expressed herein are those of the authors and do not necessarily reflect the views of NOAA or any of its sub-agencies.


Carter, M.R. et al., 2004: Shocks, Sensitivity and Resilience: Tracking the economic impacts of environmental disasters on assets in Ethiopia and Honduras. BASICS Collaborative Research Support Program, University of Wisconsin, 38 pp.

Dinku, T. et al., 2011: “Combined Use of Satellite Rainfall Estimates and Raingauge Observations to Produce Rainfall Time Series over Data-Sparse Regions of Africa”. In review, Meteorological Applications.

Grey, D. and C.W. Sadoff, 2006: “Water for Growth and Development”. In, Thematic Documents of the World Water Forum. Comision National del Agua, Mexico City.

Hellmuth, M. E. et al. (eds.), 2007: Climate Risk Management in Africa: Learning from Practice. International Research Institute for Climate and Society, New York.

Stern, N., 2007: The Economics of Climate Change: The Stern Review. Cambridge University Press, Cambridge, UK.

World Bank, 2006: Ethiopia: Managing Water Resources to Maximize Sustainable Growth. (Report No. 36000-ET). The World Bank, Washington, DC.


1International Research Institute for Climate and Society (IRI) The Earth Institute at Columbia University, USA
National Meteorology Agency, Addis Ababa, Ethiopia
Department of Meteorology, University of Reading, UK
Honorary Fellow, School of Environmental Sciences. University of Liverpool, UK

Share this page