| Graphic: | 
I wrote a bit of code in: aster/isff/src/sfun/EBEX00/radcomp.q
to compare the Rsum and Rnet values. A plot over the last 4 days is
attached. Red lines are Rnet, Green Rsum using PSPs, and Blue Rsum using
K&Zs.
The nighttime Rsum values use just PIR data and are larger by ~20% than
Rnet values.
The daytime Rsum values usually are larger than Rnet over cotton (and smaller
over dirt!). Confounding this study is whether the domes are dirty. Stn7
at least, was cleaned 2 days ago, which appears to have increased the magnitude
of the K&Z somewhat. In general, the K&Zs read a bit higher than the PSPs.
123: radiation, Site all, Mon 21-Aug-2000 10:08:08 PDT, Dome and Krypton cleaning
Today is dome and krypton cleaning. All nets had a fine
layer of dust on them. The glass domes looked good.
Kryptons had minimal contamination.
time (local) site comments
8:45 3 dome and krypton done, dusty net
9:15 6 dome and krypton done, dust on net
10:40 1 dome and krypton done, dusty net
11:10 2 doem and krypton done, dust and spots on net
11:35 4 dome and krypton done, dusty net
12:05 5 dome and krypton done, dusty net
13:10 7 domes done, dust on uplooking domes and net
13:30 8 domes done, dust on uplooking and net
13:50 9 domes done, dust on uplooking and net
15:30 7 krypton cleaned
16:30 8 krypton cleande, dirt spots on probe arms
16:50 9 krypton cleaned
136: radiation, Site , Fri 25-Aug-2000 10:15:25 PDT, rad teardown
I started teardown of the rad/soil systems this morning ~7:00AM. The
order was: 3,1,2,4, and I will continue with 5&6 soon.
The Qsoil sensor at 2 was labeled "Bad", like the one we replaced at 1.
I'll have to reexamine the data from this site. I also accidentally cut
the insulation on the T or G probe while cutting the cable tie on this sensor.
I've taken photos of the rad stands as I've gone.
139: radiation, Site all, Fri 29-Sep-2000 10:52:55 MDT, Radiation summary
Here is a summary of comments affecting rad operations:
(not necessarily exhaustive, since I only looked at subject: radiation,
not status, log, etc.
7-25-00 17:30 GMT: cleaning of all radiometer domes at stand site 09.
check of level. All ok, except NCAR upward looking
pyrradiometer is only a little bit out of center.
nothing adjusted.
cr23x uni basel: time incorrect.
7-28-00 18:10 PDT: John rewired stn1 logger so these signals now look okay.
07/31/00 13:33 PDT:
The Kipp&Zonnen has been providing bad data through the night and
into the day. During the day after warming up, the data has been
OK. The wiring inside the ventillation plenum had a bad connection.
It was recabled and now appears OK.
The wiring arrangement for the K&Z radiometers needs to be improved.
7-30 16:00 - 16:15 GMT Cleaning of all domes at radiation stand site 09.
Dome of Q*7 of NCAR is dirty, especially at lower parts.
probably due to spraying. Completely cleaning was not possible.
Time Uni Bsel OK. Leveling of all instruments OK.
8-5 18:19PDT: The Rsw.in and Rsw.out may have been temporarily fixed.
These have never been right at this site since setup.
The Rsw.out looked like an 'in.'
The Rsw.in was basically garbage/missing.
Checked:
Wiring: consistent w/ Comments in logger pgm list..ie correct.
Msg Out: both aster prep and logger prm listing consistent..ie correct.
CR10 pgm: appears correct as best I can tell.
Possible Mux problem in logger box, or else pgm is actually incorrect.
It appears that the cr10 is sampling these on AM416 mux channels
3-L/H, and 4-L/H instead of 4 and 5. This implies we either need
another 'pulse port 2' to move the mux forward to the correct channels,
or else the thing is sticking on 1 channel.
Temp fix = plug Rsw.out cable into the 'in' port on the box which goes
to 4-L/H. Jumper plug 'Rsw.out' port on the box to 3-L/H
and plug Rsw.in cable to the 'out' port.
I presume the coefficients are being applied correctly in the program.
Keep an eye on it to see if we're OK, and maybe try that program step
also.
Radiometer Cleaning cycle:
st1 8/8 ~23Z
st2 8/9 ~0Z
st3 8/7 ~22Z
st4 8/9 ~0Z Q7 has those darn spots...
st5 8/7 ~21Z were in good shape.
st6
st7 8/6 ~0Z The Rsw.in/out definitely needed it but not extreme.
The Q7 had spots on it probably from spray?
Also cleaned the German's sensors.
st8 8/6 ~23Z Rsw.in/Rlw.in did need it but not extreme. Q7 spots
seemed to come off pretty well.
st9 8/6 ~20Z The Q7 had the worst spots seen so far. Cleaned
them some but not completely off.
8-6 19:00 PDT: s8
The Psp.out has been giving bad readings.
It appears to be a connection or the Eppley photopile
is bad.
8-7 15:19 PDT The TSFC probes at station 8 are working now.
The wrong calibration factors were used.
8-7 16:47 PDT: The Rsw.out (s8) sensor is working now.
It was another case of bad wiring inside the ventillation plenum.
8-8 11:20 PDT: Rsw.in and Rsw.out at site 7 looking good
Another program problem. The multiplexer was
not skipping forward 2 steps. The heat may have had
something to do with it. I added delay commands between
the increment commands. That seemed to fix it.
8-11 10:01 PDT: KZ at station 6 removed
The problem with the KZ seems to be in the
sensor so I brought it back to the base. I
was out there at the site from 9:20 - 9:40
local time.
8-11 10:24 PDT: KZ fixed at station 6
The KZ at station 6 is back up, ~10:40 local. The
internal connector was bad. removing the shrink tubing
near the connector fixed the problem so I left the
connector in place. My guess is the contacts between
male and female have opened up and little bit of twist
on the cable causes and open circuit.
8-14 17:29 PDT: Clean domes & krypton at 3
Due to the move the domes and krypton needed cleaning.
This was done at ~17:15 local.
8-16 11:32 PDT: rads cleaned s7
While Thomas was cleaning his radiometers at s7, he offered to clean ours
as well. He noted that the Q7 was rather dirty, but the others seemed okay.
8-21 10:08 PDT Dome and Krypton cleaning
Today is dome and krypton cleaning. All nets had a fine
layer of dust on them. The glass domes looked good.
Kryptons had minimal contamination.
time (local) site comments
8:45 3 dome and krypton done, dusty net
9:15 6 dome and krypton done, dust on net
10:40 1 dome and krypton done, dusty net
11:10 2 doem and krypton done, dust and spots on net
11:35 4 dome and krypton done, dusty net
12:05 5 dome and krypton done, dusty net
13:10 7 domes done, dust on uplooking domes and net
13:30 8 domes done, dust on uplooking and net
13:50 9 domes done, dust on uplooking and net
15:30 7 krypton cleaned
16:30 8 krypton cleande, dirt spots on probe arms
16:50 9 krypton cleaned
140: radiation, Site , Fri 29-Sep-2000 16:55:13 MDT, Epply radiation errors
- Previous -
Next -
Index
| Graphic: | 
| Graphic: | 
We've seen the old Epply problem of negative PSP values at night.
Just for fun, we've compared these at one site (8) to the dome and case
temperatures of an adjacent PIR. The correlation is shown in the first
plot. A regression of Rsw = 12 (Tdome-Tcase) + 3 fits rather well!
To see what this error would be during the day, the second plot shows this
regression function in yellow, along with the nighttime Rsw data that it is
fitting (black). The daytime error is about 2%, which is rather large!
Nevertheless, this error is small compared to the energy balance residual.
Note that this error is not exactly correctable, since we don't have a
Tdome and Tcase for the PSPs. (But we might use nearby PIRs anyway,
if we are desperate.)
10/10: P.S. I've tried the fit with (Td-Tc)^4 and with Tc^4 + B(Td-Tc)^4,
and the simple linear fit still appears to be the best. Tony suspects that
this indicates a heat transfer by convection, conduction and/or diffusion,
rather than by radiation. He says that Marshal Haeffelin (NASA/LARC) was
modeling (on 8 Oct 98) the Epply PSP and may have some insight.
I've also done a similar fit for Rsw.out at s8 and get a fit of 13(Td-Tc)-2,
which is a similar slope but different offset.
- 142: radiation, Site all, Tue 31-Oct-2000 09:52:06 MST, Campbell coefficients
Post-EBEX00 the integrity of the Campbell coeffs was checked.
Consistency in the sequence:
CalLab data | Coeffs in LabView loading files | Coeffs inspected post-EBEX
was examined and found to be correct. See delany/deployments/ebex00/coefftab.fm
There was an apparent discrepancy which was resolved. The coeffs in Campbell
St7NW were not loaded in the memory, but rather were embedded in the code.
This circumstance probably occurred because originally there were only
eight Campbell 10X's available and one Campbell 10. The Campbell coefficients
need to be loaded in the code as the memory for the Campbell 10 is volatile.
Then an additional Campbell 10X was acquired and allocated to St7NW.
The original code was retained. This would cause no problem, although it
did cause puzzlement.
Neither JM nor SS (nor anyone else) changed coefficients in the field.
However, JM did find and correct (see LB# ????) cable misconnections.
JM did find and correct (see LB# ????) wiring shorts
SS, did adjust multiplexer delays (see LB# ????)
- 143: radiation, Site all, Tue 07-Nov-2000 15:21:38 MST, Radiometer coefficient checks,flagged bad data
Inspection of the coefficients indicated that the correct coefficients were
used throughout the deployment.
The procedure of checking coefficients is described in the file:
delany/deployments/ebex00/coefftab.fm
and involved checking the coefficients throughout the sequence:
cal lab data
coefficients listed in LabView coefficient Campbell loading files
manually cycling through the memory of the Cambpbell after return from EBEX00
The problem data resulted from wiring breaks and the bad data is to be
flagged but not corrected.
This flag file is formed by defining when the value of the output of an
individual radiometer deviates significantly (say 20%) or more than
20 Wm-2 from the value of a credible reference radiometer.
These reference short-wave down-welling and short-waveup-welling radiometers
were chosen by inspecting the traces of all the radiometersfor reasonable
and continuous data.
For the Rsw.in radiometers the four traces were examined to define which
was the most continuous.
Rsw.in 9 is taken as that credible reference radiometer.
For the Rsw.out radiometers all 15 traces were examined to define which
was the most continuous.
Although 3, 4, and 5 had fairly good continuous traces the best was chosen.
Rsw.out 4 is taken as the most credible reference radiometer.
- 144: radiation, Site all, Tue 07-Nov-2000 15:24:49 MST, Dome cleaning and calc of Rsw.in/out values
The dome cleaning procedure during EBEX was not as systematic as
could have been wished. A summary of dome cleaning undertaken during
EBEX00 was made and listed as Logbook# 139. Plots of the radiometers'
responses were made and inspected.
There were four up-looking short-wave radiometers:
Station 7 unventilated Eppley PSP -> Rsw.in 7
Station 7 ventilated Kipp & Zonen -> Rsw.in.kz 7
Station 8 ventilated Eppley PSP -> Rsw.in 8
Station 9 unventilated Eppley PSP -> Rsw.in 9
As the four radiometers viewed the identical sky the are redundant and
a single value of Rsw.in is all that is required. Inspection of the
traces for the four were made to decide how best to generate this single
trace of Rsw.in
The effect of cleaning had noticeable effects on
Rsw.in.kz 7 on 16 Aug of ~ 50 Wm-2
Rsw.in 8 on 21 Aug of ~ 80 Wm-2
other cleaning events did not make much difference in radiometer response.
The other two up-looking short-wave radiometer yielded more consistent traces.
with Rsw.in 7 ~ 1.04 X Rsw.in 9 .
I propose to use for the the single value of Rsw.in
Rsw.in.mean <- (Rsw.in 7 + Rsw.in 9)/2 for the period 6 Aug - 23 Aug
when both thes radiometers had good output, and
Rsw.in.mean <- (Rsw.in 9 X (Rsw.in.mean/Rsw.in 9) for the period before
6 Aug when only Rsw.in 9 had good output
In general the dust from the traffic along
the dirt road together with the effect of agricultural aerial spraying
was more apparent on the up-looking radiometers than on the down-looking
radiometers.
There were fifteen down-looking short-wave radiometers:
Station 1-6 unventilated Eppley PSP -> Rsw.out 1-6
Station 1-6 ventilated Kipp & Zonen -> Rsw.out.kz 1-6
Station 7 unventilated Eppley PSP -> Rsw.out 7
Station 8 ventilated Eppley PSP -> Rsw.out 8
Station 9 unventilated Eppley PSP -> Rsw.out 9
Each of these radiometers viewed different surfaces although the paired
Eppley and Kipp and Zonen radiometers at the six station, stn1 - stn6 could
be assumed to have had a very nearly identical view.
The ratio of Rsw.out/Rsw.out.kz for stations 1-6 were:
(1) 1.076, (2) 1.08, (3) 0.975, (4) 1.07, (5) 1.065, (6) 1.04, or (1-6) 1.05
The value taken for each of the up-welling short-wave radiation should be the
mean of the Eppley and the Kipp 1and Zonen Radiometers.
The value for the other three stations should be the value of the single
radiometer at that site.
| |
|