Sampling the Landscape

1
Sampling the Landscape

• Scenario You have been told from people who fish
  cypress creek regularly that the species density
  and composition of both fish and shellfish change
  dramatically below the golf course (the old
  Florence landfill)... You are a curious
  scientist.
• How do you find the answer?

2
(No Transcript)
3

• Scenario You are working on creating a land
  cover/land use image for central Asia in the Tien
  Shen mountains. You have spent many hours
  creating what you believe is the BEST
  landcover/landuse classification possible...
• How do you know your classification is correct?

4
(No Transcript)
5

• Scenario You are traveling in a place you have
  never been before. You want to make sure you
  have a good basis to evaluate this new place...
  You want to avoid being influenced by wow
  factors nor do you want to only remember the
  first and last place you visited.
• How can you insure an objective assessment?

6
(No Transcript)
7
Options to characterize apopulation

• Census observe every element in a population
• Sample observe selected elements in
• population and extrapolate properties to
• population characteristics

8
Statistical Sampling.

• http//yosemite.epa.gov/R10/OEA.NSF/34090D07B77D50
  BD88256B79006529E8/F3974A8B1ED416A188256CA800759EF
  A?OpenDocument

9

• BY A SMALL SAMPLE, WE MAY JUDGE THE WHOLE PIECE
• ---- Miguel de Cervantes

10
Typical objectives of a sampling design
forenvironmental data collection are

• To determine whether certain characteristics of
  two populations differ by some amount,
• To estimate the mean characteristics of a
  population or the proportion of a population that
  has certain characteristics of interest,
• To identify the location of hot spots (areas
  having high levels or concentration) or plume
  delineation,
• To monitor trends in environmental conditions or
  indicators of health.

11
Guidance on Choosing aSampling Design for
EnvironmentalData Collectionfor Use in
Developing a QualityAssurance Project PlanEPA
QA/G-5S
http//yosemite.epa.gov/R10/OEA.NSF/34090D07B77D50
BD88256B79006529E8/F3974A8B1ED416A188256CA800759EF
A?OpenDocument
12

• ...A well-planned sampling design is intended to
  ensure that resulting data are representative of
  the target population and defensible for their
  intended use....

13
SAMPLING DESIGN CONCEPTS AND TERMS

• Target population is the set of all units that
  comprise the items of interest in a scientific
  study
• Sampled population is that part of the target
  population that is accessible and available for
  sampling.
• sampling unit is a member of the population that
  may be selected for sampling, such as individual
  trees, or a specific volume of air or water.
• measurement protocol is a specific procedure for
  making observations or performing analyses to
  determine the characteristics of interest for
  each sampling unit.

14
Two main categories of sampling designs

• Probability-based sampling designs apply sampling
  theory and involve random selection of sampling
  units.
• (quantitative conclusions about the sampled
  population are produced)
• Judgmental sampling designs involve the selection
  of sampling units on the basis of expert
  knowledge or professional judgment.
• (statistical analysis tools cannot be used,
  conclusions can only be drawn on the basis of
  professional judgment)

15
Simple Random Sampling

• Selected using random numbers, and all possible
  selections of a given number of units are equally
  likely.
• Traditional statistical technique, familiar to
  all statistics courses
• most useful when the population of interest is
  relatively homogeneous
• can be more costly due to difficulties in
  obtaining samples (inaccessible areas,
  transportation difficulties)
• Can be invalid in a heterogeneous environment
• no bias in the sample selection (can occur if
  units are selected subjectively.

16
Stratified Random Sampling

• target population is separated into strata, or
  subpopulations that are thought to be more
  homogeneous
• There is less variation among sampling units in
  the same stratum than among sampling units in
  different strata.
• potential for achieving greater precision in
  estimates of the mean and variance
• Water/land

17
Systematic or Grid Sampling

• samples are taken at regularly spaced intervals
  over space or time.
• useful for estimating spatial patterns over a
  large area or trends over time.
• ensures uniform coverage of a site.

18
(No Transcript)
19
How many points need to be sampled?

• Somebody is always going to argue that you need
  more points.
• Traditionally, 20 is the number of observations
  needed to be considered significant
• 2-80 rule.... Unofficially, from Compton
  Tucker NASA-GSFC ... A 2 sample of the
  population will describe 80 of the variability
  within the sample

20
The transect or windshield survey

• A regular sampling scheme
• A linear transect allows the geographer to
  extrapolate a limited set of field observations
  into a description of a fairly large area.
• A well done transect avoids observational bias.
  (primacy, recency, oh wow!)

21

• Linear transects are usually laid out along paths
  of easy transport while this can limit their
  statistical viability and leads to a biased
  sample the short answer is that they are cheap,
  easy to lay out and often one of the best tools
  available for initial reconnaissance of a field
  area

22
across the grain

• it is desirable to construct the transect across
  as many features / factors of the landscape as
  possible. In many places the gradient on the
  landscape will be elevation and or climatic.
• In Oregon, a north/south transect has little
  utility as the topography and climatic
  variability are greatest in the E-W direction
• In North Alabama, the grain appears to be related
  to the Tennessee River

23
Stratified Linear Random Sampling

• A defensible method
• Assumes that true random sampling will be
  expensive
• Assumes that populations should be divided into
  sub-populations with lower variability (land use
  classes for example)
• Assumes transport corridors will provide a
  reasonable sub-set of the target population.

24
(No Transcript)
25
(No Transcript)
26
(No Transcript)
27
(No Transcript)
28
(No Transcript)
29
There is something fascinating about science.
One gets such wholesale returns of conjecture out
of such a trifling investment of fact. ---- Mark
Twain
30

• Observation
• In the space of one hundred and seventy six
  years, the Lower Mississippi has shortened itself
  by two hundred and forty two miles.
• That is an average of a trifle over one mile and
  a third every year.

31

• Therefore any calm person, who is not blind or
  idiotic, can see that in the Old Oolitic Silurian
  Period, just a million years ago next November,
  the Lower Mississippi was upwards of one million
  three hundred thousand miles long, and stuck out
  over the Gulf of Mexico like a fishing rod.