Environmental Engineering Reference
In-Depth Information
data collection. This, in turn, makes it easier for the agencies to deal with financial
and technical constraints. For example, an initial investment on model develop-
ment and data collection could be subsequently improved and enhanced as addi-
tional data collection and model development phases are completed. In this
context, the subsequent stages of research and development would progressively
address weaknesses and limitations of the initial work.
The base alternative is simply not to collect data (Case 0 for each category). In
this case, analyses would be limited to the use of generation rates from the lit-
erature, and delivery tour data would heavily rely on OD synthesis, thus limiting
the scope and applications of the resulting models. This approach could be
improved by collecting additional data. With a limited budget, selecting Cases 1
for freight generation, delivery tour and agent spatial distribution data categories;
although allowing the estimation of freight demand models for certain industry
sectors, would not provide county specific models. In terms of tour data, the use of
a sample from GPS aggregators could lead to potential bias in the data and the
inability to understand the factors that determine the trip or the trip purpose. In
addition, with the ZIP Code Business Pattern data it would not be possible to
precisely geo-locate the agents, thus limiting the ability to complement the data
with other commercially available economic/land use datasets.
The data collection effort could be progressively improved by selecting larger
sample sizes (Cases 2-6). In summary, increasing the number of surveys for
freight generation data would allow the estimation of disaggregate freight trip
generation models for different industry segments and the ability to capture dif-
ferences among areas. A larger delivery tour data collection effort would meet
modeling needs and include all industry segments. Collecting cordon surveys
provides internal-external, external-internal and external-external trip data that
would complement the delivery tour collection effort granting access to a set of
data for backup and validation purposes. Acquiring a larger sample of records
from data aggregators allows accurate geo-location. Considering the freight/freight
trip generation characteristics of large traffic generators (large establishments or
building that generate a disproportionate number of freight trips or tonnage) would
provide a more complete picture of freight demand in large urban areas, given that
they—although small in quantity—can account for a large proportion of urban
freight movements. For special purpose/choice analyses, any of the three cases
proposed could be selected, bearing in mind that fewer observations would come
at the expense of not having enough data for validation purposes. It is also
important to stress that the alternatives have vastly different implications in terms
of quality of the modeling effort that they could support. While not collecting any
freight data would definitely minimize collection costs, it would lead to large (and
of unknown magnitude) errors during the modeling process. On the other hand,
conducting an extremely comprehensive data collection effort may not necessarily
improve the quality of the modeling process because of the models inherent
limitation to replicate the system under study.
Search WWH ::
Custom Search