Regional aggregation of statistics on water abstraction and use Seminar on water statistics for countries in Eastern Europe, Caucasus and Central Asia September 2012, Almaty, Kazakhstan Stefan Tsonev National Statistical Institute Bulgaria
A good tool for better understanding, modeling and decision making is the spatial presentation of the information. Statistical offices have the practice to do this mainly by presenting data within administrative borders. For this purpose a list (nomenclature) of administrative/territorial units is used. Examples are the two regional nomenclatures developed by Eurostat: - NUTS ( Nomenclature of Territorial Units for Statistics ); - LAU ( Local Administrative Units)
The nomenclatures as files are available at Eurostat’s web page. This is an example with a screenshot of the LAU file for Bulgaria:
Using “LAU2_NAT_CODE” we can aggregate (summarize) the statistical data for the all administrative levels from lowest to highest (settlements, municipalities..., etc.). Armed with proper administrative maps as GIS layers we can easily present spatially many data from the environmental statistics.
Many, but not all. Household wastes can be collected and presented within administrative borders (aggregations), as municipalities are responsible for them. Other studied processes such as those related to waters depend on natural factors such as relief are generated in geographical borders, so different than administrative aggregations are used. Such aggregations in river basin districts are required by the Water Framework Directive.
WATER FRAMEWORK DIRECTIVE 2000/60/EC Article 3 Coordination of administrative arrangements within river basin districts (RBD) 1.Member States shall identify the individual river basins lying within their national territory and, for the purposes of this Directive, shall assign them to individual river basin districts Member States shall ensure the appropriate administrative arrangements, including the identification of the appropriate competent authority, for the application of the rules of this Directive within each river basin district lying within their territory.
So for policy making, physical and monetary data aggregated by river basins (RBD) and/or administrative regions are required by national authorities and EU institutions. Bulgaria is divided in 4 river basins districts (RBD), containing 27 sub-basins (Sub-RBD). The borders of the river basin districts are build by the shapes of the sub-basins. At LAU 1 and 2 levels, Bulgaria is divided in 264 municipalities and settlements
Settlements and municipalities (LAU 1 and LAU 2)
Maps of RBDs and LAU with crossing borders 64% of municipalities are situated in 1 Sub-RBD 30% - situated in 2 Sub-RBDs 6% - situated in 3 Sub-RBDs 1% - situated in 4 Sub-RBDs
SurveyPopulationType of survey 1. Water supply, sewerage and treatment 1.1. Public water supply; sewage and UWWTP Total coverage (100) 1.2. Irrigation systems 2. Water use 3. Water users Partial statistical survey carried out through studying the main population. It covers the big water consumers – criterion for coverage: enterprises using over than 36 m 3 water annually (100m 3 /day). Some enterprises below this criterion submit data on a voluntary basis. 4. Hydroelectric plantsTotal coverage Data collection - water statistics surveys
PROBLEM: Data collected from particular water supply companies in Bulgaria, refer to a specific region, neither basin, neither municipality or other administrative structure.
Survey and data processing technology (To make data suitable for regional aggregations) All data collected with questionnaires include: -ID Enterprise -NACE code -Activity’s location – code LAU2 (settlement) Information received from water supply companies refer to a specific region, but with the exception of water abstraction, the data collected with questionnaires are reported at a LAU 2 level within the region. Data received from water use survey are also collected on LAU 2 level.
To LAU nomenclature three additional fields are added: - RBD code (code of the river basin district where the settlement is placed); - Longitude and Latitude (geographical coordinates of the settlement as a point object for GIS applications).
Problems caused by crossing borders How statistical data referred to point units (settlements) which are situated in more than 1 Sub-RBD can be aggregated? How statistical data referred to irregular areas can be decomposed into the smaller units? How the statistical data produced by one type of region to be transposed to another type?
Practical approach for spatial aggregation of statistical data at RBD level 1. Data aggregation: from LAU2 to Sub- RBDs; 2. Data disaggregation using additional data “regionalizing factor”; 3. Additional collection of spatial data - if 1 & 2 are not applicable.
Data aggregation: from LAU 2 to Sub-RBDs The data collected by the statistical surveys allows us to refer them to a given point – LAU 2 settlement with it’s geographical coordinates. Combining this points we can build different shapes by aggregating data for an administrative code and/or for geographical region.
On the map above, the settlements (presented as points in different colors for each company) are showing the areas served by the companies. ( As it was already mentioned these areas have nothing common with the boundaries of sub-river basins /Sub-RBD/ ). For each settlement (point) from the map we have a set of data, collected by statistical surveys.
Some indicators calculated at Sub-RBD level by using point data - settlements Self-supply water abstraction, assumption: water source is near to the location of enterprise, identified by LAU2 Water use by supply category, by economic activity Wastewater generation and discharge by type of collecting system Sewage, UWWTP: Wastewater collected, treated and discharged; BOD5 – Incoming load and Effluent, treatment capacity; Sewage sludge production and disposal Population connected to: public water supply, sewage and UWWTP Other: drinking water purification plants (capacity, population connected), population with water-supply restriction, etc.
Data needed to be disaggregated by using “regionalizing factor” Data on waters, reported only on a company level: –Water abstracted for distribution by source; –Water transfer; –Water entered the water supply systems; –Losses of water during the transport –Price of water services Data from other sources – such as regional GDP (only available at NUTS 3 level).
Data disaggregation Example: Water price reported by companies partially included in one RBD or unit of NUTS 3 level. Task: The price of water for the RBD or NUTS 3 unit is needed. Solution: We have the quantities and price of the water delivered for each settlement within the boundaries of the RBD or NUTS 3 unit. Then we calculate weighted average by multiplying the totals of water delivered by each company (in the RBD or NUTS 3 unit) by the corresponding price. Then the sum of water price for the two companies is divided by the total quantity of water delivered in the RBD or NUTS 3 unit.
Calculation at sub-basin /Sub-RBD/ level: Price of “Water used” delivered by Public water supply Calculation method: data disaggregation Assumption : annual average price of water is the same for the whole region served by given company (known at “company level”) Regionalizing factor: Volume of water supplied to the users (known at NUTS 5 level) Calculation of the price as a weighted average by sub-basins (Sub-RBDs)
Calculation at RBDs level: GDP (GVA) Available data: GDP/GVA at NUTS 3 level Methods for data disaggregation: Assumption: Breakdown of GDP at RBDs is based on the assumption of homogeneity of GDP/GVA per person at district level (NUTS 3) Regionalizing factor: Annual average population
Difference between aggregations - Drinking water used by household per capita –at NUTS 3 and Sub-RBDs level 1. Average (l/day/per capita): NUTS 3: Sub-RBDs: Variation coeff.% NUTS 3: 17-21% Sub-RBDs: 39-52% Conclusion: The average and variance are changed mainly due to shift from administrative to geographical borders. Administrative aggregations are relatively homogeneous; at Sub- RBDs level - large dispersion
Thank you for attention!