This page contains additional information regarding observation
locations, observation values, and conditioning/prediction schemes
used in
"Approximating likelihoods for large
spatial datasets"
Michael L. Stein, Zhiyi Chi and Leah J. Welty
Section 4:
Locations of points on 100 x 100 grid.
Section 6:
Chlorophyll Fluorescence Values: The first row consists of the number of observations
(13,815) along with the number one. The remaining rows consist of
log(fluorescence) values corresponding to the locations given below.
Chlorophyll Fluorescence
Locations: The first row consists of the number of observations
(13,815) along with the number two. The remaining rows consist of
distance from shore (km) and depth (m) of the corresponding
observation. We have smoothed the locations in order to account for
discretization in the distance from shore measurements.
The chlorophyll fluorescence measurements were provided by Henry
Vanderploeg, Great Lakes Environmental Research Laboratory. The
measurements were collected as part of the Episodic Events Great Lakes
Experiment (EEGLE) Project, supported by the National Oceanic and
Atmospheric Administration's Coastal Ocean Program and the National
Science Foundation.
Chlorophyll Fluorescence Collection Path (numbers indicate the first and last observations, numbered 0 and 13814):
In the conditioning/prediction set designs, we refer to the points
in order from the 0th point to the 13,814th point. Note that the
fluorescence profile was taken from off-shore to near-shore, so the
0th observation, which has log(fluorescence) value 3.97 log(volts),
was taken ~37.05 km from shore and at a depth of ~4.65 m. The
13,814th observation has log(fluorescence) value 3.61 log(volts) and
was taken ~6.73 km from shore at a depth of ~1.86 m.
The conditioning/prediction sets start with the number one. The
second number is the number of integers that follow after the second
number. The third number is the total number of
conditioning/prediction blocks. The remaining numbers consist of
conditioning and prediction sets specified block by block. Each block
specification consists of the total number of of points in the
conditioning set, the total number of points being predicted, the
vector of points in the conditioning set, and the vector of points in
the prediction set. For example, the specification for the
ith block consists of the numbers ci
pi x1 x2
... xci y1 y2
... ypi where ci is the number of
conditioning points, pi is the number of prediction points,
the xjs refer to the conditioning points, and the
yks refer to the prediction points.
Conditioning/Prediction Sets:
DN(10)
DNF(10)
DNMF(10)
DN(20)
DNF(20)
DNMF(20)
This page was last updated August 2003.