Central America  Data Solutions: Central America Data Solutions


Data solutions for risk management: Data solutions for risk management The complete package Historical observations 10 years + No missing or erroneous observations (cleaned) Discontinuities detected and corrected Near real-time data updates (feeds) Continuous stream of observations – no feed failures No missing or erroneous observations (cleaned) Settlement data Reliability and accuracy is critical Forecasts Reliability and accuracy Without all 3 components in place the deal(s) may never get off the ground.


Data solutions for risk management: Data solutions for risk management As with any region of the world we begin by looking at the two most widely used types of data and then we will consider other options Climate Data Synoptic Data Other Solutions


Data solutions for risk management: Data solutions for risk management As with any region of the world we begin by looking at the two most widely used types of data and then we will consider other options Climate Data Data is subject to the highest levels of quality control Is the basis of the official records of historical weather and climate for a country (legal quality data) Climate data typically is observed over a 24-hour period (true max / min) Current standard for weather risk transactions Often expensive and difficult to obtain Synoptic Data Other Solutions


Climate Data: Belize: Climate Data: Belize Instituto Meteorológico Nacional (http://www.imn.ac.cr) 34 Stations Currently Reporting (May 2007) Limited real-time observations available on internet Few Stations with more than 20 years history


Climate Data: Costa Rica: Climate Data: Costa Rica Instituto Meteorológico Nacional (http://www.imn.ac.cr) 68 Stations Currently Reporting (Feb 2007) 13 automatic stations (real-time data available on the internet) 28 precipitation only 40 precipitation and temperature 20+ years History


Climate Data: El Salvador: Climate Data: El Salvador Servicio Nacional de Estudios Territoriales (www.snet.gob.sv) Center of Pronóstico Meteorológico (CPM) Center of Información and Agrometeorología Center of Predicción Climática (CPC) 5 Synoptic stations 27 Climate stations 72 Precipitation Stations 42 Stations available online Displaying near-real time hourly observations Temperature, wind, relative humidity, precipitation,… Historical data available up to 80 years


Climate Data: Guatemala: Climate Data: Guatemala Instituto Nacional de Sismologia, Vulcanología, Meteorología e Hidrologia (http://www.insivumeh.gob.gt/index.html) 9 Synoptic stations 65 Meteorological stations (13 Automatic stations) 71 Hydrological stations (daily and hourly observations) Historical data available for 48+ years (limited selection)


Climate Data: Honduras: Climate Data: Honduras Servicio Meteorológico Nacional – SNM (http://www.smn.gob.hn) 14 Synoptic stations 20 precipitation stations Basic data feeds (14 Synoptic stations) available on internet Historical data: many stations with 60+ years of data


Climate Data: Nicaragua: Climate Data: Nicaragua Nicaraguan institute of Territorial Studies (INETER) (http://www.ineter.gob.ni) 20 primary weather stations Substantial historical records (40+ years) Climate summaries available online


Climate Data: Panama: Climate Data: Panama Hydrometeorology Management of Empresa de Transmisión Eléctrica S.A. (ETESA) (http://www.hidromet.com.pa) 12 Climate stations (updates available online – broken link since Feb 2005) 91 active weather stations


Climate Data: Summary: Climate Data: Summary Facts Each Meteorological Office has some digital archives of historical observations Many countries are currently providing limited real-time observations on their web sites (no historical data archives available on the internet) Reliable and accurate observations in a near-real time environment is available for a limited number of core stations Feasibility / comments Preliminary discussions with these Met Offices indicate that data can be made available for market use We are not able to comment on the quality of the data, we are assuming moderate quality at best Sourcing data will be somewhat time consuming (redistribution contracts) and costly Cleaning techniques will play a major role in assuring the accuracy and completeness of time series Settlement data (per the U.S. and Europe) will probably not be available The complete package? History Feeds Settlement = may be a problem


Data solutions for risk management: Data solutions for risk management As with any region of the world we begin by looking at the two most widely used types of data and then we will consider other options Climate Data Synoptic Data Synoptic data is often observed over a 12-hour period (12-hour max / 12-hour min) Undergoes minimal amount of quality control (erroneous observations problematic) Exchanged among National Meteorological Services to produce forecasts (ECMWF, GFS,…) Easier Access to Data / less expensive Lack of settlement data in the traditional sense Other Solutions


Synoptic Data: : Synoptic Data: Synoptic data in Central America is observed using the following convention (all times in UTC) Minimum temperatured = 0000d to 1200d Maximum temperatured = 1200d to 0000d+1 Precipitationd = 1200d to 1200d+1 Missing and erroneous observations are extremely problematic Precipitation, sometimes missing is denoted by “0.0” “good” stations the are ones missing less than 20% of observations (good climate stations are ones missing less than 5% of data) Settlement data is non-existent in its traditional form (Met Office edited data)


Synoptic Data: Feeds : Synoptic Data: Feeds GTS (Global Telecommunications System) transmitted Synoptic reports An analysis of the past 3-months of archived reports finds that 18 stations have reliable data feeds (max, min, precipitation) Summary: Feeds are reasonable Feeds are useless without historical data for pricing


Synoptic Data: Historical data availability : Synoptic Data: Historical data availability Archived Synoptic reports from the ISH (Integrated Surface Hourly dataset) / Global Summary of the Day database Inventory (mid 1970s to present) Total of 49 stations with historical data Belize = 3 Costa Rica = 6 El Salvador = 5 Guatemala = 8 Honduras = 15 Nicaragua = 7 Panama = 5 Maximum Temperature: 13/49 are missing more than 50% of all observations Minimum Temperature: 13/49 are missing more than 50% of all observations Precipitation: 18/49 are missing more than 50% of all observations


Synoptic Data: History & Feeds : Synoptic Data: History & Feeds History without feeds = useless Feeds without history = useless The COMPLETE PACKAGE is available for 18 locations Country Summary Belize = 1 Costa Rica = 0 El Salvador = 0 Guatemala = 5 Honduras = 9 Nicaragua = 0 Panama = 3


Synoptic Data: Summary: Synoptic Data: Summary Facts The GTS feeds appear to be reliable with reasonable amounts of missing data The historical data is reasonable with stations missing on average 23% of observations This solution is feasible for some countries and not others (Costa Rica, El Salvador, and Nicaragua have no data) Feasibility / comments The data cannot be cleaned – not enough surrounding stations for traditional methods Without cleaning the risk exists for erroneous observations There is no reliable source for settlement data We would recommend using this data as-is only if desperate The complete package? History (potential problems) Feeds (potential problems) Settlement


Data solutions for risk management: Data solutions for risk management As with any region of the world we begin by looking at the two most widely used types of data and then we will consider other options Climate Data Synoptic Data Other Solutions… Private networks / system upgrades Re-analysis data


Other Solutions…: Other Solutions… Private networks / upgrade existing systems Install instrumentation – working with the Met Office or private companies to install observation systems for weather risk management purposes Specific geographical placement Known data quality / reliability Certified data for settlement purposes Reduced tampering risk Generate synthetic history / clean existing archives Use existing data from 2nd and 3rd tier climate networks Proven ability to create representative observations The complete package? History Feeds Settlement Re-analysis data Historical data Feeds / updates


Review / comments: Review / comments Climate data Best historical data solution (after cleaning), feeds questionable, settlement data problematic Difficulty sourcing data / cost Synoptic data Easiest solution for ‘reliable’ feeds, historical data ok at best, settlement data non existent To be used as a last resort Private networks / upgrade existing systems Best solution for settlement data and feeds Costly and time consuming Synthetic historical data not preferable


Recommendations: Recommendations Scenario #1 (one-off trade, small transaction, limited data budget) Consider initial pricing using historical synoptic data or climate data Focus on stations that already report via Synoptic or web site Carefully word contracts with regards to settlement Scenario #2 (one-off trade, large transaction, moderate budget) Focus on stations with archives of climate observations (source data and CLEAN data) Focus on stations that already report via Synoptic or web site Employ an independent 3rd party for certification of data for settlement Scenario #3 (strategic business direction, multiple transactions, large budget) Select regions of interest, review climate data availability (sourced data, CLEAN data, synthesize data) Work with Met Office to upgrade existing hardware and communication systems (install stations) Employ an independent 3rd party for certification of data for settlement