-
Wiegert’s Method?
- Pick the quadrat size or shape that minimizes the product of relative cost and
- relative variance
-
What size quadrat?
- want norm dist
- min edge effect
- different species different values per size quad
- easy to read
- cost-person min/quadrat
- stat precision high
-
Hendrick's assumptions?
- 1)
- Assumes that the slope of the line is between 0 and -1.
- •This
- is a very restrictive assumption, and if the slope is not within this range,
- the method can not be used
- 2)
- Assumes that the amount of time to read a quadrat is directly proportional to
- size (i.e. 2m2 is twice as costly as 1m2).
-
What's different about Hedrick's Method?
- Pointed
- out that variance usually decreased with larger quadrat size.
- The
- log of variance will fall linearly with the log of the quadrat size.
-
Two approaches for picking size and shape quadrate?
Go to lit, do same as everyone else.
Better, do pilot study
-
Sampling design?
Method of placement
- Number of samples
- ..also think about gear (2 practical components)
- decided before hand
-
Best sampling design..
maximizes...
best...
smallest possible...
at lowest...
- efficiency
- statistical estimates
- confidence intervals
- cost
-
To determine a sampling design...
specify...
decide on...
select a...
- statistical pop
- sampling/experimental units
- sample and adopt a variety of sample plans
-
Simple random sampling?
2 things
- N sampling unit pop
- every one has equal chance
note: sampling with replacement better
-
Stratified random sampling?
- divided in subpops that don't overlap
- randomly sample within each sub pop
-
Strata?
- a subpop in a stratified random sample
- together make up entire pop
-
random versus simple random sampling?
- random=every indiv has equal chance
- simple random=every possible group has an equal chance
-
Benefits of Stratification?
get estimates of means and confidence intervals for each subpop
sampling problems can be addressed at each level
RESULTS IN GAIN OF PRECISE ESTIMATES FOR WHOLE
maybe more than one lab studying same..admin convenience
-
Systematic sampling?
regular or systematic placement of samples
simple and even, hopefully objective
-
problems with systematic sampling?
- periodic variation, not common though in ecology
- still good if looking for gradients, or if you need cheap and easy
-
centric systematic sample?
equal strata, sample in middle
-
Adaptive sampling? Why?
Take advantage of spatial pattern to get a better estimate of abundance.
-
Multi-stage sampling?
two or more levels of sample selection; 7 lakes primary, benthic areas of 10m2 is element
-
Cover? how is it expressed?
area of ground covered by plants, rock, litter, bare ground
percent
-
Four types of cover?
basal, canopy, foliar, ground
-
cover, indicator of 2 things...
- ecological processes
- (dominance)
management--erosion etc
-
cover is a good substitute for....
biomass; too hard and too much disruption.
-
Three general approaches to measuring cover?
Points, lines, plots/quadrats
-
3 Spatial patterns in nature?
- Aggregated (clumped)
- Uniform (regular or spaced)
- Random
-
poisson distributions resemble what kind of habitat distribution?
uniform
-
Negative k binomial distributions are like what in nature?
synonymous with aggregation
-
3 measures from Line Transect Sampling?
- sighting distance
- sighting angle
- perpendicular distance
-
Problems with Line transect sampling?
- detectability falls with distance
- undercount bias
-
Detection function?
estimate the zone of detectability, falls with distance
-
distance methods, 2 general approaches?
- random organism to nearest neighbor
- random point to nearest organism
- (object to object, or point to object)
-
problems with object to object and point to object approach in distance methods?
hard to pick a random organism
point to organism is biased towards isolated individuals
-
distance methods for plants?
n,n,p,r
- nearest neighbor
- point-centered quarter
- nearest individual
- random pairs
there's more
|
|
