Data Management and Climatology


High quality data is needed for climate analysis with sufficient spatial coverage and duration. For climate applications projects, demands of meteorological data in terms of coverage, frequency and immediacy are more stringent than for climate analysis. Thus both for climate analysis and project work, data must be collected. The data resources in Sri Lanka are rich and are often only in paper form, inaccessible and expensive. Over the years, we have sought to acquire and organize data for our research in climate and climate risk management. We have carried out quality assessment of key data.

 

Activity
Objectives
Partners Status Next Steps
Data Management
Obtain and manage data resources including meteorological, hydrological, sectoral and societal data. Include in suitable databases.
Data librarian, MASL.
Sufficient data for work in hydro-climatic analysis at all available locations are being archived. Records have been collected from global and local data sources.
Collection of daily records needs improvement. Collection of monthly data and database development needs to be completed. Updating of data needs to be organized. Malaria data management in progress.
         
Quality Control of Data
Quality assessment of meteorology and agriculture measurements. Methods extended to hydrological data.
IRI

 QC of rainfall and temperature (min, max and mean) of all meteorological stations and agro-meteorological stations completed and Technical Reports prepared.
Create a PC-based tool for quality control. Extend QC methods to entire database. Conduct QC on malaria and hydrology data. Develop methodology for daily data records
         
Climatological Mapping
Prepare an Atlas of climatology in Sri Lanka
University of Peradeniya, Iowa State University, MASL, NASA/GSFC Land Surface group.
Climatologies of rainfall,mean, maximum and minimum temperature, solar radiation mapped. Modeled wind, soil moisture and evaporation.
 Mapping to be refined. Mock of publication to be completed. Digital map library to be completed.

Data Resources

Systematic weather observations started in 1868. By 1900, there were 90 observation stations. Since then there have been up to 1000 stations at one time or another and regular rainfall records are maintained in half of these.

Compared to other tropical countries, Sri Lanka has excellent meteorological data. The map below shows the stations that provided data to the global archive named Global Historical Climate Network (GHCN) in the 19th Century. Sri Lanka is the region in the tropics with the densest network of stations.

 


 

Quality Control of Climate Data

Errors arise in climate data due to instrument error, observation errors and transcription errors. Such errors can unduly influence climate analysis particularly for climate change. Therefore we need to assess the quality of the data. Such procedures are useful to infill missing data.

Climate analyses is sensitive to outliers arising from a variety of sources, like shifting of the stations, instrumental errors and human error. Hence, prior to analysis it is important to remove outliers.

Temperature fields should show spatial uniformity. An anomalous rise in one station is likely to be associated with a similar rise in neighbouring stations. Therefore we may compute the standard error for each year for the anomalies for each station to assess the degree of scatter in a given year for stations across the country as a measure of quality.

Such quality analysis gives a way to build homogeneous and complete data set and reconstruct the discarded or missing data.

/files/fect//qc.png

 

The precision of climate change estimates from instrumental data is sensitive to the quality of data. The quality of the monthly temperature and rainfall data at the main stations of the Sri Lanka Department of Meteorology and Agriculture were evaluated based on station histories, comparison with observations at neighbouring stations and internal consistency of the data. The Global Historical Climatology Network (GHCN) has previously carried out such quality control for 11 Sri Lankan stations. We have undertaken quality assessment for 18 stations each of which had a longer duration than used in the GHCN analysis and had fewer missing records. The GHCN quality checks lead to all data prior to 1950 from Trincomalee, Batticaloa and Colombo and data from Hambantota prior to 1921 being discarded. The present quality control leads to a much smaller set of data being discarded.
Inter Annual variations – Temperature at any station should relate to that in neighbouring stations. We may use a statistical techniques to estimate the temperature at station using the observations at neighbouring stations alone. The comparison of such reconstruction and observation is illuminating.

 

Climatologies

/files/fect//climat.png

Climatology

Given its size, Sri Lanka shows a remarkable variation of climate among its regions. The variation in temperature is modest with the mean value being 27oC with a mean daily range of 6oC. The relative humidity varies from 60% to 90%. Westerly winds prevail over the island from May to September and North-Easterly winds prevail from December till February. Rainfall measurements are available for Colombo from 1853 and there are a dozen stations which have a 130 year rainfall record and over 500 rainfall recording stations with a 50 year record. Some of these stations are shown in the figure below along with the topography.

Regionalization


Sri Lanka has been regionalized considering -
  • annual rainfall (as wet, dry and intermediate),
  • elevation (as up, mid and low country) and based on
  • agro- ecological zones (considering rainfall, elevation and soils).
Puvaneswaram and Smithson of University of Sheffield, UK. presented a scheme that considers seasonal rainfall tables (see map) It is useful as different climatic mechanisms dominate specific regions.

 

/files/fect//clmamap.png
Map: The three climatic regions of Western, Eastern and Plains. The Western and Eastern regions may be subdivided into the coastal (< 200 m altitude) and mountain (>200 m altitude). The Plains region may be divided among the Northern and Southern Plains. The climatologies of these different regions show seasonal distinctions.

 

Monthly Regional Rainfall

Rainfall is high from October to December and from April to June through all regions of the island. During the intervening periods, rainfall is augmented on the western hill slopes from May to September and on the eastern hill slopes from December till February.

/files/fect//rainfallgrapgh.png

Seasons


There is no clear cut method of specifying seasons for Sri Lanka. Traditional agricultural seasons are from -
  • October to March (Maha)
  • April to September (Yala).
  • Meteorologists have demarcated the seasons as the
  • North-East monsoon (DJF)
  • First Inter-Monsoon (MA)
  • South-West monsoon (MJJAS) and
  • Second Inter-Monsoon (ON)
The significant climatic processes that bring rainfall are the passage of the Tropical Convection Zone (TCZ) around April and November, orographic rainfall in the Western hill slopes (MJJAS) and in the Eastern hill slopes (DJF) and cyclonic storms particularly in November and December. The monsoon over the region modulates the rainfall particularly in the boreal winter (DJF) and summer (JJA).
Meteorological demarcation is useful for studies of wind and temperature patterns but does not separate modes of rainfall well. For analysis of rainfall, we often use a breakup of the agricultural seasons into their wetter early half (AMJ and OND) and drier later halves (JAS and JFM).

Output

Conference Proceedings

Dec 1999: with Gunaratne, A.U.K, A mass consistent wind model for Sri Lanka, Proceedings of the 8th Asian Fluid Dynamics Congress, Shenzen, China 8:817-820.
Dec 2002: with Heli Bulathsinhala, Quality Evaluation of Mean Historical Temperature Data in Sri Lanka, Sri Lanka Association for the Advancement of Science, 58th Annual Sessions, University of Colombo, Colombo, Sri Lanka.

Abstracts submitted:

Nov 2004, Climate change assessment for Sri Lanka from quality evaluated data. International conference on sustainable water resources management in the changing environment.
July 2004, Topographically informed interpolation of temperature in Sri Lanka. Submitted to: First national symposium on Geo-Informatics, PGIA, Sri Lanka.

Reports

November 2005: Manjula Siriwardhana, Janaki Chandimala, Siraj Razick, Upamala Tennakoon and Lareef Zubair, FECT Technical Report -7, 2005, Quality evaluation of diurnal temperature range in Sri Lanka.


November 2005: Manjula Siriwardhana, Heli Bulathsinhala, Kusalika Ariyarathna, Irugal Bandara, Janaki Chandimala, Siraj Razick, Upamala Tennakoon and Lareef Zubair, FECT

Technical Report-3, 2004, Quality Evaluation of Historical Rainfall Data in Sri Lanka.

May 2005: Manjula Siriwardhana, Kusalika Ariyarathne, Irugal Bandara, Heli Bulathsinhala, Janaki Chandimala, Siraj Razick, Upamala Tennakoon and Lareef Zubair, FECT Technical Report-2, Quality evaluation of historical maximum and minimum temperature data in Sri Lanka.

February, 2005, Current Climate and Climate Change Assessments for Coconut and Tea Plantations in Sri Lanka. Submitted to: START, Washington, D.C., USA,

December 2004: Irugal Bandara, Kusalika Ariyarathne, Janaki Chandimala, Siraj Razick, Manjula Siriwardhana, Upamala Tennakoon and Lareef Zubair, FECT Technical Report-4, Quality evaluation of historical maximum and minimum temperature data in the agro-meteorological stations in Sri Lanka.

December 2004: Heli Bulathsinhala, Manjula Siriwardhana, Kusalika Ariyarathne, Irugal Bandara, Janaki Chandimala, Siraj Razick, Upamala Tennakoon and Lareef Zubair, FECT

Technical Report-1, 2004, Quality evaluation of historical mean temperature data in Sri Lanka.

Data Products

Archive of monthly quality controlled rainfall and temperature data for main agro-meteorological and meteorological stations. Spreadsheets for quality control work including identification of outliers, statistics, calculation of anomalies, reconstruction of data and programs to identify and correct for stations shifts and instrument miscalibration.