Special Features of Environmental Surveys to Establishments

1
Jorge Saralegui-Gil Environmental Statistics
and Accounts Unit INE. Madrid (Spain)
jsaralegui_at_ine.es Quality Implications of
specific sampling techniques in environmental
surveys
Q2008 Rome July 2008
2
Environmental Statistics (INE, Spain).
3
Water use in Agriculture (I)

• Mixed Frame
• River Basin Authorities, (wholesale water,
  abstraction)
• Irrigators associations (distribution ,
  irrigation techniques, turnover)
• Agricultural Holdings (crops, self supply,
  costs).
• Area frame
• Difficulties to avoid duplicates of water flows
  supplied to agr. holdings.

4
Water use in Agriculture (II)
Estimates combine results of the Survey on Water
Use in Agriculture (INE) , the water use module
within the Survey on Waste Generation in
Agriculture (INE ) and the Area Frame Survey
(Min. of Agric.).

• - The basic step in the estimation procedure for
  the SWUA uses irrigated regional total area S,
  classified according to the irrigation technique
  k (leaking, sprinkling or gravity) as exogenous
  variable for calibration purposes.
• A correction c is needed to cope with
  self-supplied water a by holdings which consider
  the externally supplied water a insufficient.

5
Non irrigation water statistics (I)

• Three flow types (water abstraction, water supply
  and waste water collection and treatment)
  constitute estimation targets for surveys on
  the sector.

• Considered as public services, its final
  responsibility is entrusted to local
  authorities.
• Very complex management structure .
• Ad hoc observation units Water Managers
  (agregations of establishment type units) in a
  domain.
• Mixed frame for the Survey on Water Supply and
  Treatment The database links territorial
  districts with microdata identification of the
  corresponding CBR units .

6
Non irrigation water statistics (II)

• Take-all strata H Frame units serving medium
  and large districts.
• For the remaining water managers (i,j below) a
  ratio estimator is used it is based on the
  number of residents ( p) of the
  municipalities k being served by them (with a
  correction coeffic. c whenever the service is
  shared with other managers) .

7
Sewage Treatment.
Population figures cannot be used as totals for
calibration (coverage unknown )

• Exogenous (Min. of Env.) Equivalent population
  is used instead to estimate the total
  population receiving the waste water treatment
  service in the domain.
• The concept of equivalent inhabitant is defined
  as the biodegradable organic load carrying a
  biochemical 5-day oxygen (DBO5) demand of 60 gr.
  of oxygen a day.
• Equiv. Pop. can be calculated at the micro level
  for responding units (water managers).

8
Waste Generation Surveys (I)
Information needs on waste are approached by the
statistical system in the context of

• Estimates presented in satellite accounts
  (material flows accounts , satellite accounts on
  waste) .
• Structural statistics (to produce both
  environmental indicators and reports).
• Specific sections of sustainable development
  indicators systems.
• Other fields of waste statistics refer to the
  type of treatment applied to waste, and the
  treatment facilities characteristics.  

9
Waste Generation Surveys (III)
Quality challenges Reporting units
(establishments, holdings ) must cope with a
double coding problem, related to the use of
different classifications.

• Administrative regulations (particularly
  demanding for hazardous waste) use LoW (List of
  Waste, product oriented) . Statistical system (in
  EU) works with EWC-Stat (activity oriented
  categories
• Technical difficulties to treat blank cells
  either as a partial non-response or as a zero
  value.
• Strategic mandates to reduce reporting burden.
  Regional estimates are highly demanded but the
  estimation cannot be efficiently carried out
  using direct estimators based on small samples

10
Waste Generation Surveys (IV)
Estimators of waste type X at the regional level
C are of the post stratified ratio type. An
ancillary variable v is introduced (sales,
production, employment ) whose value is known
for every sample unit i of stratum h in post
stratum A , as well as its exogenous totals from
the active population survey and from the
structural annual business surveys respectively,
as follows

11
Waste Generation Surveys (V)
It becomes necessary for some wastes to make use
of a synthetic estimators based on nationwide
means in order to impute the latter to a post
stratum A within the domain C.

12
Urban Waste (I)
The Frame for the Collection and Treatment of
Urban Waste Survey (UWS hereafter ) consists of
territorial units linked to the
establishment-type units supplying the service

• Take-all strata coverage amounts to more than a
  40 of the territory as a whole , thus
  estimation-related biases are fairly bounded.
• As with the sewage treatment, the use of complex
  estimation procedures within sampling strata
  becomes necessary as an exact knowledge of
  territorial coverage for collection of particular
  categories of selective waste is lacking.

13
Urban Waste (II)

• As far as the distinction urban waste versus
  household waste is concerned, the procedure
  combines estimations from the UWS and from waste
  generation surveys
• In the waste generation survey questionnaires
  addressed to the service sector establishments
  and institutions , one section concerning the
  destiny of waste is included in order to know
  whether the waste is handed over to either a
  waste municipal manager or to a non-municipal one
  (licensed private managers)
• In doing so it becomes possible to estimate the
  fraction of waste collected by municipal
  managers, and consequently the amount of a
  particular category of waste generated by
  households, derived from the UWS estimated
  totals.

14
Waste Treatment (I )

• Perhaps the most important area of waste
  statistics.
• Two approaches to frame building
  facility-oriented approach the licensed waste
  managers approach.

• - In a first step, a sample of licensed waste
  managers is drawn- from administrative records
  as provided by the regional environmental
  authorities .
• - Both the number and total capacity of several
  types of waste treatment facilities are
  externally available for each domain. It is used
  as an ancillary variable when estimating total
  treated waste

15
Waste Treatment (II)
The estimator is of the type

• x denotes the amount of waste that have been
  treated by manager j of the frame, in stratum h
  by treatment type k. (either recycling, land
  filling, incineration with or without energy
  recovery, etc.) .
• Design weight w of unit j is adjusted within
  the stratum h by the coefficient of non response
  /out of scope c. t refers to maximum capacity
  for treatment k, as stated for those facilities
  managed by a sampled unit.
• T refers to total treatment capacity for type
  k, as given by external sources, within the
  estimation domain. 

16
Other environmental statistical operations

• To estimate the use of hazardous chemical
  products in agriculture Holdings fill out an open
  answer question set where they are inquired on
  the commercial brand, amount, crop and license
  code for each plan protection product used.
• To obtain estimates at active substance level a
  subsequent computer procedure is carried out by
  merging administrative records of the identified
  ppt .
• Supply-side environmental protection statistics
  production estimates from samples of
  establishments are to be combined with estimates
  for the demand-side (expenditure) variables,
  along with administrative sources or other
  external information such as data from
  sector-related entrepreneurial associations .

17
Some conclusions (I)

• Heterogeneous nature of environmental statistical
  operations which lie typically halfway between
  common practices for national accounting and the
  traditional survey-based statistics.
• In order to avoid double counting,
  underestimates or omissions, when estimating
  physical flows related to generation and
  treatment of either waste or water, work teams
  for environmental surveys have to be able to act
  with high methodological flexibility.
• They do need to be provided with sufficient
  resources in order to build survey frames and set
  complex designs following mixed procedures,
  which combine statistical registers and
  administrative sources with one or even more
  surveys, in one or several steps.

18
Some conclusions (II)

• Placing the personnel that carry out
  environmental accounts, indicators and surveys in
  the same technical units turns out to be very
  convenient.
• Staff in charge of data collection have to be
  aware that physical accounting will still take
  some time until it is familiar to business
  practices, then adapting instruments to this
  (hopefully) transitory scenario is of crucial
  importance for the quality of environmental
  statistics.
• These facts condition altogether the special
  profile within the statistical system of the
  environmental statistics area that needs to be
  taken into account by the statistical offices any
  time they design their medium or long term
  operating plans.

19

THANK YOU
Quality Implications of specific sampling
techniques in environmental surveys
jsaralegui_at_ine.es. INE . Madrid
Q2008 Rome July 2008