probabilities very well. Figure 5 illustrates this for
6 FINDINGS
Milford, where the Pg/Pmax ratio is 1.76 and the log-
normal value at a 2% annual probability of being ex-
Our answers for the 140 towns are presented in Table
ceeded (50-year mean recurrence interval) greatly ex-
1. Some of the towns listed in Table 1 are only par-
ceeds the data trend there. With this evidence, we gave
tially in the CS zone. At this time for those towns, we
little weight in our analysis to stations with high
recommend that the ground snow load be determined
Pg/Pmax ratios.
using the information in Table 1 rather than from the
Once we had all 140 case study answers, we com-
map in ASCE 7-95. The case study process is a more
pared them to the answers on the upper and lower plots
detailed and thus, in all likelihood, a more accurate
on the last page of the case study form. The upper
assessment of the ground snow loads in these towns.
"nearest 6" plot answers did not agree with our an-
This is consistent with the guidance in the commen-
swers well at all. Only 59 of the upper plot answers
tary attached to ASCE 7-95, which states that "detailed
were within 5 lb/ft2 (0.2 kN/m2) of our 140 case study
study of a specific site may generate a design value
answers. For 50 stations the upper plot answers were
lower than that indicated by the generalized national
from 10 to 38 lb/ft2 (0.5 to 1.8 kN/m2) away from our
map. It is appropriate in such a situation to use the
answers. The lower "all values" plot answers were
lower value established by the detailed study. Occa-
within 5 lb/ft2 (0.2 kN/m2) of our answers for 116 of
sionally, a detailed study may indicate that a higher
the 140 case studies (i.e., 83% of the time). However,
design value should be used than the national map in-
for eight stations, the "all values" answers were from
dicates. Again, results of the detailed study should be
10 to 20 lb/ft2 (0.5 to 1.0 kN/m2) away from our an-
followed"(ASCE 1996).
swers. Thus, while the "all values" answers provide
After discussing the pros and cons of having a por-
good indications of the "correct" answers most of the
tion of New Hampshire defined by the ASCE 7-95 map
time, further study will occasionally result in signifi-
and the remainder defined by our case studies, we con-
cantly different, better answers.
cluded that it would be best to expand our case studies
The elevation correction factor can also be exam-
to cover the entire state. We have agreed in principle
ined on the upper and lower plots. On the upper
to do that and will revise the CRDA between CRREL
plot that factor varied widely between 13.5 lb/ft2
and SENH to increase the scope of work accordingly.
per 100 ft (2.12 kN/m2 per 100 m) and minus 9.0
We expect that once we have done the entire state and
lb/ft2 per 100 ft (minus 1.41 kN/m2 per 100 m).
examined all of our answers, some of the values in
The average value of this widely divergent and physi-
Table 1 may change a little. Thus, we advise readers
cally unrealistic set of numbers was 1.8 lb/ft2 per
to consider those values as interim in nature.
100 ft (0.28 kN/m2 per 100m). We place little value
To determine the ground snow load at elevations
on this average, as it is significantly influenced by
other than those listed in Table 1 (i.e., at elevations
other than those where the case studies were con-
some values that are physically unrealistic. Stations
like Mt. Washington create these inappropriate values.
ducted), the values in Table 1 should be increased or
On the "all values" plot, the slopes make somewhat
decreased by an elevation correction factor of 2.1 lb/
ft2 per 100 ft (0.33 kN/m2 per 100 m). For example, in
better physical sense, but Mt. Washington and a few
Hanover where the Table 1 value is 75 lb/ft2 at 1300 ft
other stations still create problems. Slopes vary from
5.3 lb/ft2 per 100 ft (0.83 kN/m2 per 100 m) to
(3.6 kN/m2 at 396 m), at an elevation of 900 ft (274
minus 3.0 lb/ft2 per 100 ft (minus 0.47 kN/m2 per 100
m) the answer would be 75 + (2.1/100)(9001300) =
m) and average 2.4 lb/ft2 per 100 ft (0.38 kN/m2 per
75 8 = 67 lb/ft2 (in SI units: 3.6 + (0.33/100)(274
396) = 3.6 0.4 = 3.2 kN/m2).
100 m).
We further examined the elevation correction fac-
We have not fully investigated the upper limit above
tor by studying each station in our database. We elimi-
which our elevation correction factor does not apply.
nated stations with less than 15 years of record, others
At this time it seems safe to use it up to an elevation of
with an elevation above 2500 ft (762 m), and others
2500 ft (762 m) in New Hampshire. At higher eleva-
with Pg/Pmax ratios less than 0.9 or greater than 1.7.
tions a larger elevation correction factor may be
For the remaining, high quality stations, the line of
needed.
best fit of their elevation to their 50-year ground snow
load, Pg, produced a slope of 2.1 lb/ft2 per 100 ft (0.33
kN/m2 per 100 m), as shown in Figure 6. While we
7 CONCLUSIONS AND RECOMMENDATIONS
expect that the elevation correction factor varies from
place to place in New Hampshire, we do not have
The current case study plots contain some data of lim-
enough data to support such differences. Thus, we have
ited value that mask rather than define trends. Perhaps
used this elevation correction factor for all New Hamp-
stations with fewer than about 14 years of record should
shire towns.
be eliminated from the plots on the case study forms
and perhaps stations with Pg/Pmax ratios exceeding
about 1.7 should also be eliminated from those plots.
319
Previous Page
