LIMITATIONS IN WATER QUALITY DATA

from: Gleick, P.H. 1993. About the Data. In: Water in Crisis: A Guide to the World's
Fresh Water Resource (Ed., P.H. Gleick), Oxford University Press, New York, pp 117-119.

“Good data are hard to come by. Data are often not collected regularly and systematically.
They may not be reported in useful or consistent forms If collected, they may never be
compiled or distributed. Standards and techniques of measurements differ from region to
region, or worse from year to year Units of measure vary around the world -different
quantitation measures may have the same name, or a single measure may be called different
things in different places. Data are collected by individuals with differing skills, goals, and
intents. Some data are collected objectively; other data are collected to support different
ideological or political biases. The more we know about a data set and how it was collected,
the better equipped we will be to evaluate it and use it.”

Problems with data format

• primary data
  • original (ie., collected)
    • should list qualifiers
      • where, when, how, problems, limitations
• secondary data
  • recycled data
    • often do not list qualifiers
    • “For example, all estimates for the sediment discharge of the Irrawaddy
      River are based on one set of measurements made by the British in the
      1870s and quoted over and over since then.”
• need to use standard (SI) units
  • standard international metric units
  • use exponents rather than prefixes
    • billion may mean 10^9 or 10^12
• problems with US conventions
  • rain is measured in inches
  • water supplies are measured in acre feet
    • cover an acre with one foot of water
    • ~ annual requirements for a family of four

Errors in the data

• inadvertent propagation from errors in original sources
• introduced in transcription and conversion

Difference between “measured” data and “derived” data

• direct measurements of geophysical hydrologic data
  • time consuming, expensive, not comprehensive
• =>derived data
  • estimates, based on assumptions
  • computer models
  • gigo (garbage in, garbage out)

Time-series data and time of estimate

• temporal variables
  • single estimate
    • made at a specific time
  • estimates of a fixed estimate
    • changing “t”, changing investigators, changing methods
  • estimates of changes over time
    • time series data
    • hourly, daily, seasonal & yearly variability

Uneven regional data coverage and oneven data quality

• International Drinking Water Supply and Sanitation Decade (1981-1990)
  • “... ,we are still missing large blocks of information about access to clean water
    and sanitation. In some regions, data are available for less than half the
    population, typically excluding poorer rural areas almost entirely.”
• most analyses have been made in temperate environments
  • poor understanding of water quality in the tropics and polar regions
• emphasis on newer problems in the industrial world
  • e.g., organics in ground water
  • these problems also exist in third world countries

Variability, uncertainty, and illusory precision and accuracy

• variability
  • stochastic nature of hydrologic processes
    • time and space
  • human factors (techniques, needs, economics,..)
• uncertainties
  • incomplete knowledge
    • time and space
    • analytical capability
• illusory precision
  • data reported with too much precision
    • e.g. 6.37" of rain in 24 hours
  • rounding off data may hide differences
    • e.g., 5'7" ,. 6' ,. 6'5" tall
• false accuracy
  • single error, multiple errors

Aggregation of data

• problems in processing raw data
  • presentation in simplified figures or tables
    • engineers and lawyers want “a number”
    • annual rainfall or river runoff
  • processed to prove a point
    • a water body “is” or “is not” polluted
• averaging masks differences
  • average ground water withdrawal (changing wells)
  • average pollutant discharge rate

Inconsistent definitions, standards, and boundaries

• standards for common measurements
  • e.g., fecal coliform in Monterey Bay
• no standards for other parameters
  • e.g, bioavailable copper in SF Bay
    • “safe drinking water”
• changes in boundaries
  • e.g., international rivers now in former USSR

International politics and hydrologic data

• restrict data on water and environmental quality
  • “national security”
    • pollution in former USSR
    • water resources in Middle East

“What is in a number? Each number has a a value, such as 2.5 rather than 6.7. Each number
has a unit that defines its character, such as a volume, a flow, or a level of contamination.
Each has a level of precision, 2.5 rather than 2.531. But each also represents more than just a
physical or scientific fact. It represents something real, something tangible about the world in
which we live, something that may be far more complicated than it appears on the surface.
For example, data on per capita water availability in a country is a measure of the what is
theoretically available, not what is actually supplied to each human living there, which is
often far less. Data on potential hydroelectric resources in a country may suggest enormous
untapped energy reserves, but tapping them may be possible only at the cost of strangling our
fisheries, or ruining our Grand Canyons into flat-water reservoirs. Data on the fraction of the
population without access to basic sanitation and clean water gives a sense of the staggering
raw numbers of people without the most basic human services. But there are other meanings
hidden in this single number as well: it means 1,300 million people carmot turn on a tap in
their household and get clean water, as the readers of this book can. It means that tens or
hundreds of millions of (primarily) women and children must spend hours each day searching
for water for their basic needs And it means that millions die every year of preventable
water-related diseases.”