Relative Humidity Calibration Notes
These are my rough notes. The slope values are
pretty much meaningless. The scatter is the standard
deviation of each calibration period and level group,
no outliers removed. The figures indicated are working figures
(like Figure 7.), that I used for
my analyses. They do not refer to the figures on these www pages.
<![CDATA[
li = linearly interpolated from other periods
Bottom hazer vs level 2 (figures 19,20) 1-2
Var Pe Bias Slope Scat Comments
rh 1 -1.0 ~-0.1 0.0984 Big cluster at -1 bias drops lower at higher RH
rh 2 -0.75 li
rh 2.5 -0.5 li
rh 3 -0.3 li
rh 4 -0.2 ~0 0.3831 Large scatter like t but well centered on median bias
rh 5 -0.2 ? 0.7949 Large scatter
rh 6 -0.2 ? 0.2125 Nice cluster
Bottom hazer vs level 3 (figures 21,22)
rh 1 -1.3 ~0 0.5567 Nice cluster, one outlier causing large scat
rh 2 -0.3 * 0.3722 Big group around bias
rh 2.5 0.2 li
rh 3 0.5 * 0.3653 A few low outliers at low humidities near end of cal period
rh 4 1.1 0 0.1854 Bias very consistent over wide RH range
rh 5 1.45 ? 0.8657 Large scatter, with cluster around bias = 1.2
rh 6 0.7 ? 0.2651 Tight group at high RH
Bottom hazer vs level 4 (figures 23,24) bias = bottom(1) - 4
rh 1 -2.1 ~0 0.5908 Moderate scatter, but well centered around median bias
rh 2 -1.2 li
rh 2.5 -0.5 li
rh 3 0 * 0.2848 Tight cluster at rh =91 and bias 0.1
rh 4 0 * 0.3826 Pretty consistent over wide RH range
rh 5 0.3 ? 0.2036 Bottom hazer pegged at RH=100 due to previous editing so comparison not valid
rh 6 0.6 ~0 0.4929 Moderate scatter but no trends and well centered around median bias
Period:r=1,y=2,g=3,b=4,m=5,k=6
bottom hazer vs top hazer (1st figures 4)bias = bottom-top
rh 1 0.8 ? 0.4411 A few high outliers but mostly well clustered aroun mean bias
rh 2 0.8 li
rh 2.5 0.9 li
rh 3 0.9 * 0.1634 Tight group
rh 4 1.8 * 0.5579 Large Scatter at low wind speed, but rest well-clustered about median bias
rh 5 0? * (na) Both pegged at 100% - no good comparison possible
rh 6 0? * (na) Both pegged at 100% - no good comparison possible
Top hazer vs level 4 (2nd figures 4) bias = top - 4
rh 1 -2.7 * 0.1348 Tight group
rh 3 -0.8 ~0 0.1867 Moderate scatter around median bias
rh 4 -1.8 0.1 0.6942 Increase with higher rh, Large scatter, especially at lower temps
rh 5 0.3 * 0.1574 Level 5 pegged at 100% so not good comparison
RH conclusions
Similar to T all errors appear to be bias and not f(t) or f(rh)
Top Hazer comparisons with 1 and 4 are consistent with 1 vs 4 comparisons
Bottom hazer seems to get too high with time
Level 3 consistently low after start;
The suggested calibration is based on linear fits between calibration periods.
(No q fluxes available to test periods when q flux = 0. Maybe after Lyman-alpha
data processed. As with t, the "true" bias was assumed to be average of all humidities
during each calibrations period.
]]>
|
|
| Outline Page |
Relative Humidity Page |
|
Last update: 4/22/99
Please send all comments and suggestions to the author, Peter
Guest,