78
February 2017
PHOTOGRAMMETRIC ENGINEERING & REMOTE SENSING
SECTOR
INSIGHT:
.
com
sufficed to indicate that despite the tendency of citizens living
in the most impoverished wards to use fewer testing kits, the
voluntary sampling was non-preferential in nature, leading
to an amazingly good reproduction of housing and neighbor-
hood characteristics. On the other end, the state-controlled
sampling program turned out to be surprisingly biased. The
impact of LSL on water lead levels was mainly investigat-
ed for houses built between 1935 and 1955 and located in
affluent neighborhoods. Particularly troubling was the total
absence of sentinel sites with LSL in two of the most impov-
erished wards, including the ward where the largest percent-
age of children with high blood levels was recorded in 2015.
Once preferential sampling has been detected, the next step is
to adjust the computed statistics accordingly; i.e. by weighing
the frequencies of data above 15 μg/L computed within each
class of housing characteristics, poverty levels, and wards to
reflect the proportions observed in the population (i.e. Flint
housing stock)
3
. However, such correction cannot compensate
for the absence of data in specific classes of housing character-
istics. In particular the lack of pre-1935 and post-1955 houses
with LSL in the sentinel network is a major weakness since
larger lead levels were recorded for this category of housing
in the voluntary program. Therefore this adjustment did not
reduce the gap between the results of two sampling programs.
Another potential culprit is the influence of the sampling pro-
3
Flint Safe Drinking Water Task Force Recommendations on MD-
EQ’s Draft Sentinel Site Selection. (February 2016). Retrieved from
on October 31, 2016.
ASPRS Certification validates your
professional practice and experience. It
differentiates you
from others in the profession. For more
information on the ASPRS Certification
program:
contact
,
visit
THE
IMAGING & GEOSPATIAL
INFORMATION SOCIETY
STAND OUT FROM THE REST
earn asprs
certification
ASPRS congratulates these recently Certified and Re-certified individuals:
tocol, in addition to the sampling design. In other words, how
the samples were collected by homeowners matters as much,
and potentially more, than their geographical location.
What are the take-home messages of this study? First, it is
critical for the person or organization conducting the sampling
to document the principles underlying the selection of sam-
pling sites, in particular when the results are expected to be
under heavy scrutiny by the public, the scientific community
and the media. In the present case, the sentinel monitoring
system which includes a total of 759 sites is described in a sin-
gle paragraph
3
. Second, trying to pursue too many objectives
using the same sampling scheme is often a recipe for failure.
The sentinel system had multiple goals, including contribut-
ing to the inventory of LSL and determining the effectiveness
of corrosion control additives, which might explain the prefer-
ential sampling. In particular, targeting LSLs in houses built
between 1935 and 1955 hampered our ability to disentangle
the effects of premise plumbing and LSLs on water lead levels,
which is critical to build accurate predictive models. Third, as
a spatial analyst leave no stone unturned. In most cases, you
won’t be consulted prior to data collection and will have to mine
datasets created through preferential or targeted sampling.
Then a good understanding of underlying physical processes
in action is as important as statistical skills to make sense of
results. If you are in charge of designing a sampling scheme,
a clearly illustrated and practical document is the EPA Guid-
ance on Choosing a Sampling Design for Environmental Data
Collection available at
.
CERTIFIED PHOTOGRAMMETRISTS
Michael Martin, Certification #1608
Effective December 3, 2016, expires December 3, 2021
Todd Schmidt, Certification #1609
Effective December 3, 2016, expires December 3, 2021
CERTIFIED MAPPING SCIENTIST
REMOTE SENSING
Rama Mulukutla, Certification #RS234
Effective December 3, 2016, expires December 3, 2021
RECERTIFIED MAPPING SCIENTIST
REMOTE SENSING
Chad Lopez, Certification #R179RS
Effective June 28, 2015, expires June 28, 2020
CERTIFIED MAPPING SCIENTIST GIS/LIS
Mark T. Reineke, Certification #GS291
Effective December 14, 2016, expires December 14, 2021
CERTIFIED MAPPING SCIENTIST LIDAR
Ryan Griffin, Certification #L010
Effective December 14, 2016, expires December 14, 2021
