Difference between revisions of "NGEE Teller Top Notes"
Jump to navigation
Jump to search
(37 intermediate revisions by 4 users not shown) | |||
Line 1: | Line 1: | ||
+ | ==Data QA== | ||
+ | * https://docs.google.com/spreadsheets/d/1ogCWyAmechT0oiUTxEl3KzThwVwx5LK7cjOV9SVK4oM/edit#gid=0 | ||
+ | * August 2021 Summary: | ||
+ | ** Replaced AT/RH in June | ||
+ | ** All other met data looks good | ||
+ | ** Thermal conductivity in pit | ||
+ | |||
+ | ==Extra Notes== | ||
+ | * For data grouping routine there is something off with the pit temperatures... more specifically one set is reported as 'Avg' so I must not have correct headers. | ||
+ | * Just a historical note that the Top snow stand on the lump had been bent over sometime prior to Ken H visit in fall 2020... I sorta think this happened in the winter of 2019-2020 but need to circle back to that to confirm. | ||
+ | ** Ken visit: 10/31/20 @ 19:40 is final data in Snow.dat. | ||
+ | |||
+ | ==Main== | ||
+ | [[NGEE]] | ||
+ | ===Latest=== | ||
+ | * Met Data QA'd through last visit (9/2018) | ||
+ | * Radiation Data QA'd through last visit (9/2018) | ||
* Partway through processing soil. | * Partway through processing soil. | ||
+ | ===Met 2017=== | ||
+ | * Initial teller top was an instrumented survey tripod (May 2016 to August 2016) | ||
+ | * A musk ox knocked it over and was discovered in July I think. | ||
+ | * permanent station in August 2016. | ||
+ | * The back up air temperature was not in place in 2016. The HMP155 is super susceptible to excessive moisture like found at Teller top and Utqiagvik. So, HMP155 fails almost every winter. | ||
+ | * Every time I review data I delete more subtle but error AT/RH/ DP data. | ||
+ | ** Re-reviewing the first year data in January 2022 I have caught a few more things. | ||
+ | ** original review included for the air tmemperature data & dew point from bottom station to weed out the in range but in error data as the HMP155 was failing in 2016. I am redoing the analysis now but made another dataset that includes longwave temperatures and skin temperature... that shows a bit more to cull from the record. | ||
+ | |||
===Met 2017=== | ===Met 2017=== | ||
* HMP155 August 2017 | * HMP155 August 2017 | ||
Line 15: | Line 41: | ||
* Wind | * Wind | ||
** all rime conditions corrected for through 3/2018 in QC spreadsheet. | ** all rime conditions corrected for through 3/2018 in QC spreadsheet. | ||
− | * Rain NAN through | + | * Rain NAN through snowmelt |
+ | * Radiometer leveled and desiccant replaced 6/12/2018 | ||
+ | * Radiometer leveled again 9/24/2018 | ||
+ | * Tipping bucket leveled 6/12/2018 & 9/24/2018 | ||
+ | * AT/RH corrected through 3/30/2018 20:00 after sensor destroyed over winter season. | ||
+ | ===Met 2022=== | ||
+ | * Figured out this year that the HMP155 cable was the source of the bad data issue. | ||
+ | * However, musk ox near station in May 2022 visit prevented fixing until October 2022. | ||
+ | * Back up air temperature began to fail in 2022 and will need replacing, too (sadly). | ||
===Radiation 2017=== | ===Radiation 2017=== | ||
* There is a gap 10/2/2017 0:00 to 10/2/2017 22:00 | * There is a gap 10/2/2017 0:00 to 10/2/2017 22:00 | ||
** I think this is when we removed the sensor from the field to change the desiccant. | ** I think this is when we removed the sensor from the field to change the desiccant. | ||
+ | *** as in, the screws were stripped and so we drilled and retapped back in town. | ||
===Radiation 2018=== | ===Radiation 2018=== | ||
* All good through March visit. | * All good through March visit. | ||
+ | ** There is a weird thing in 2018 where the lower SW has values higher than the upper SW. I assume this is due to snow/ frost up top. Perhaps related to how close the snow surface is to the sensor this year resulting in a lower wind speed later in the winter. | ||
+ | |||
+ | ===Radiation 2022=== | ||
+ | * Sensor swapped for recalibration but I didn't modify the program so, proper data requires 2 key file changes for this interval. | ||
+ | ** Changes: | ||
+ | ** Shortwave sensors were already set up with data_type '''swrad''' so only a value change is necessary. | ||
+ | *** Finally, '''Coef_2''' for SW sensor units are now non-zero: | ||
+ | Coef_2 = OriginalCNR4_calib_value / Replacement_CNR4 | ||
+ | ** Longwave sensors remain data_type ''num'' | ||
+ | *** Use additional sensor SBTempC_Avg.csv to: | ||
+ | **** Compute longwave of the sensor body | ||
+ | **** Set up a spreadsheet and create a new time series: | ||
+ | corrected_LW_step1 = [SBCTemp_C-> Watts (Stefan-Boltzmann Law)] - [LW From .dat] | ||
+ | corrected_LW_step2 = [corrected_LW_step1] * [original_LW_Calibration] / [Current_LW_Calibration] | ||
+ | corrected_LW_for_archive = [corrected_LW_step2] + [SBCTemp_C-> Watts (Stefan-Boltzmann Law)] | ||
+ | ** Net Radiation needs to be recomputed after corrections are applied | ||
+ | |||
+ | * 2022-04-01 02:00:00 -- First data point with replacement unit | ||
+ | ** Keyfile: | ||
+ | teller_top_micromet_2022-04-01_params_20-cnr4_090013.csv | ||
+ | * 2022-10-05 19:00:00 -- Last data point with replacement unit | ||
+ | ** Keyfile: | ||
+ | teller_top_micromet_2022-10-05_params_20-cnr4_131352.csv | ||
+ | |||
+ | * All data for this interval is contained in data file: | ||
+ | teller_top_MicroMet-2022_10_05.dat | ||
+ | * Calibration Info 2016 to 2022-04-01 01:00:00 -- SN 120966 | ||
+ | SW Upper -- 13.83 | ||
+ | SW Lower -- 13.08 | ||
+ | LW Upper -- 13.46 | ||
+ | LW Lower -- 14.06 | ||
+ | * Calibration info 2022-04-01 01:00:00 to 2022-10-05 19:00:00 -- SN 090013 | ||
+ | SW Upper -- 15.12; Coef_2 = 13.83 / 15.12 | ||
+ | SW Lower -- 10.51; Coef_2 = 13.08 / 10.51 | ||
+ | LW Upper -- 7.40; Coef_2 = 13.46 / 7.40 | ||
+ | LW Lower -- 6.72; Coef_2 = 14.06 / 6.72 | ||
+ | * Calibration info 2022-10-05 19:00:00 and onward: -- SN 131352 | ||
+ | SW Upper -- 12.70; Coef_2 = 13.83 / 12.70 | ||
+ | SW Lower -- 12.44; Coef_2 = 13.08 / 12.44 | ||
+ | LW Upper -- 10.35; Coef_2 = 13.46 / 10.35 | ||
+ | LW Lower -- 11.70; Coef_2 = 14.06 / 11.70 | ||
===Soil 2017=== | ===Soil 2017=== | ||
* 2017 heat flux plates prior to fall were being measured on wrong channel. | * 2017 heat flux plates prior to fall were being measured on wrong channel. | ||
− | * 2017-03-12 02:00:00 Upper heat flux plate has a value that isn't ever flipped by the qc program. | + | * 2017-03-12 02:00:00 Upper heat flux plate has a value that isn't ever flipped by the qc program. |
− | * 2017-03-12 02:00:00 Lower heat flux plate has a value that isn't ever flipped by the qc program. | + | * 2017-03-12 02:00:00 Lower heat flux plate has a value that isn't ever flipped by the qc program. |
* Easy Perl one liners to accomplish this: | * Easy Perl one liners to accomplish this: | ||
perl -pi -e 's/"2017-03-12 02:00:00",149.2/"2017-03-12 02:00:00",6999.00/g;' $TT_SOIL_DIR/outputs/SoilPit_HeatFlux_Lower.csv | perl -pi -e 's/"2017-03-12 02:00:00",149.2/"2017-03-12 02:00:00",6999.00/g;' $TT_SOIL_DIR/outputs/SoilPit_HeatFlux_Lower.csv | ||
Line 34: | Line 110: | ||
* 2017 Soil heat flux plates also all corrected | * 2017 Soil heat flux plates also all corrected | ||
* with Vlad determined only good VATP are 2/5/25/50/60/100 | * with Vlad determined only good VATP are 2/5/25/50/60/100 | ||
+ | * Soil Temperature: VATP | ||
+ | ** I think I have everything corrected though I don't entirely believe the '''100cm probe. | ||
+ | * 8/16/2017 from Lily's field book page 50: | ||
+ | ** '''VATP was heaved 9.5cm above standing water table. I pushed it back down''' | ||
===Soil 2018=== | ===Soil 2018=== | ||
− | * 2018 soil pit soil moisture -- all corrected through most recent visit | + | * 2018 soil pit soil moisture -- all corrected through most recent visit (3/2018) |
− | * | + | * |
+ | ===Soil 2021=== | ||
+ | * I've decided the top site experiences high heat fluxes in spring / summer and it seems likely the output is overranging the measurement thresholds set in the program. | ||
+ | |||
+ | ==Snow== | ||
+ | ===Snow Depth 2016-2017 Lump=== | ||
+ | Season Starts: '''11/3/2016 08:35''' | ||
+ | * Winter 2021 re-analysis and discussion: | ||
+ | ** I'm still hung up on the precise start. If you look at the end of season zero offset, it meshes up with a season start of 10/26/16 18:00ish or 10/29/16 11:05ish. But from met and radiation data I think this might be rain / slush that melts but it's enough to smoosh down the vegetation briefly but it springs back up again. So, I guess I'm still going with 11/3 start. | ||
+ | ** Albedo suggests no snow (centered on midnight) 10/26 to 10/28. Temperature on these days is around +2 C. So, precipitation in tipping bucket is probably rain. It looks like a litle snow fell and melted though looking at the std deviation of snow depth and the lack of noise in the snow depth data for a couple hours starting 10/26/16 1:40. | ||
+ | ** albedo on the UTC late day of 10/27 & 10/28 @ 0:00 is back to no snow. | ||
+ | ** 10/29 0:00 0.39 albedo but probably top dome covered in rain/slush/snow as the tipping bucket records a bunch. | ||
+ | ** Net Radiation is also mostly positive over this interval. that changes after 10/31 0:00 where there seems to be clearing skies at night. | ||
+ | ** Then no more real precipitation of note until 11/3 after 8:00 UTC. | ||
+ | ** with respect to the snow depth signal over this say 10/26 to 11/2 0:00 interval it's noisy like exposed ground / the signal reflecting off grass rather than a snow surface. So, roughly I think any snow that fell in this interval melted. | ||
+ | ** After sun is up on 11/2/2016 the albedo is 0.72 so looks like good snow. Value never drops below 0.6 in next several days. So pretty good confidence on that. | ||
+ | ** However, still hard to pin down a start. WIth all that said, 11/2 to 11/4 is still a mess with grass / organics not completely buried. The tipping bucket shows no precip on 11/1 or 11/2 and then 9:00 (UTC) to 14:00 on 11/3 it's tipping. So maybe 11/2 albedo is skewed high by heavy clouds / frost. air temperatures are +2 to +4 C on 11/1 and 11/2 until late in day. | ||
+ | ** So, I guess 11/3/2016 08:35 is official start but there is still ground exposed at the sensor as compared to below the radiometer. And it sort of looks like snow starts at 8:35 then transitions to rain around 15:00. If you go by the highest confidence quality numbers then it looks like snow at 8:35 which melts over the day and then a bit more falls by end of day (as illustrated by the declining standard deviation of snow depth. | ||
+ | ** Then things are just kind of a push as to what is going on. Just minimal snow but likely snow covered until next event of note on 11/16/2016. | ||
+ | *** So, going at the start of the season following whatever the furthest away high confidence Q value snow depths probably wouldn't be the worst way to go. | ||
+ | * November / December 2016 have data spikes in the high confidence Q values. I'm pretty sure this is due to the failure of the HMP155 air temperature which was the only AT active this first winter. | ||
+ | ** If you look at the Column B snow depth values, which are adjusted for temperature there is a mix of 2 decimal data and lots of decimal data ('''TT_Lump_2016-2017.ods'''). The bunch of decimal data is the infilled from Teller Bottom air temperatures (or data logger panel temperature I can't remember at the moment). Anyway, the spikes look to be in the bunches of data where the HMP155 seemed to be working but wasn't actually. | ||
+ | ** I've been seeing this in the spreadsheets periodically an interval in late winter, this time 3/12/17 2:00 to 2:55 where the data came in as text rather than date format. It only shows up scrolling by briefly but something to watch for in final outputs. | ||
+ | Season Ends: '''05/15/2017 10:35''' | ||
+ | ===Snow Depth 2017-2018 Lump=== | ||
+ | Official Start: '''10/20/2017 04:55''' | ||
+ | ==== Winter 2021/22 Reanalysis and commentary==== | ||
+ | * In fall, air temperatures drop consistently below freezing 10/17/17 or 10/18/17 depending on how you feel about popping up over 0C | ||
+ | * Summer precip record ends the 16th. | ||
+ | * albedo is 0.14 to 0.25 through 10/19/17 3:00 (UTC) and then albeod on 10/19 after sunrise (so 21:00-23:00) suggests snow and following days look like 100% cover. | ||
+ | * general snow depth signal is basically just bouncing off leaves and such through the official start. I think you could make a case maybe for 10/19/17 12;45ish as an earlier start but the fifteen house in between these dates looks mostly like the leaves are crushed down a hair without any accumulation. Accumulation starts at the official start I believe. | ||
+ | * re-review pretty anti-climatic. Nothing much to clean up beyond erroneous high confidence quality number values here and there. | ||
+ | |||
+ | |||
+ | Official End: '''06/04/2018 02:15''' | ||
+ | ===Snow Depth 2018-2019 Lump=== | ||
+ | Original Start: '''10/29/2018 03:35'' | ||
+ | Revised: '''11/08/18 10:55''' | ||
+ | Winter January 2022 Reanalysis | ||
+ | * Air temperatures don't begin dropping much below 0 C until 10/17/18 16:00ish | ||
+ | * Hover around freezing from then till 10/27/18 2:00 | ||
+ | * Then they come back up and do a bit of melting it looks like. | ||
+ | * Albedo is barren ground until roughly 10/23/18 1:00 then inches up but at least under radiometer not full snow coverate until 11/10/18 22:00. | ||
+ | * Looking at air temperature albedo simultaneously it's clear there was some melting that took place the first week of november. | ||
+ | * Looking at the snow depth data it appears there were a couple periods of snowing then melting. 10/24/18 23:10 to 10/25 23:25 perhaps | ||
+ | ** to my eye in the fall when the std deviation values for snow depth are below one I tend to thin that is snow even if it isn't amounting to anything. | ||
+ | ** This isn't an absolute but the highest confidence reported readings tend to be the ones bouncing off the ground rather than reflecting off leaves. So can kind of use that as a proxy for snow if nothing else. | ||
+ | ** Basically the first snow of note falls around 10/25 00:00 and melts, snows, melts cycles and then between 11/8/18 11:45ish and 11/9/18 21:35 there is some hard to pin down exactly accumulation but at 21:35, that is the beginning of 100% snow cover and you see that clearly in both albedo and snow depth data. | ||
+ | * Not much done beyond this start of season adjustment. | ||
+ | Official End: '''5/31/2019 23:50''' | ||
+ | Revised End: '''5/31/2019 23:00''' | ||
+ | |||
+ | ===Snow Depth 2018-2019 Tripod=== | ||
+ | * Winter 2021 re-analysis: | ||
+ | ** For tripod depth shifted the start of the season earlier compared to what was originally in archive. Took a look at tipping bucket data and thought snow arrived sooner. | ||
+ | |||
+ | ** I don't know what the shift is on 4/4/19 00:50 (UTC) values go from consistent 148.8cm to 144.5. | ||
+ | *** I think this was somehow related to a station visit. Magnaprobe data ends at top 4/3/19 16:52... tz most likely akstd | ||
+ | ** For end of data set in June 2019, standing water is present so the end of season value isn't zero-ish. Final val is 5.0cm | ||
+ | *** I'm sure I must have knocked some rime ice off or something but pretty odd. | ||
+ | *** to verify the time & place I popped open the 2019 end of winter kmz that I made once upon a time. | ||
+ | ====Start of Season==== | ||
+ | * 2018-10-24 16:15 | ||
+ | Air temperature is around +1C at this time step. Precipitation falls into the tipping bucket but I think it is slush / snow. The snow depth goes from 0 to 6.3cm over three hours and the readings are erratic with respect to quality number. Albedo is 0.23 during the day on 10/23 and increases to 0.96 on 10/24. Longwave values are within 3 Watts suggesting warm clouds / low cloud cover etc. | ||
+ | |||
===Snow 2018=== | ===Snow 2018=== | ||
* Snow temperature profile | * Snow temperature profile | ||
** Weird shift (looks like mux wonky 1/22 to 1/28/2018) | ** Weird shift (looks like mux wonky 1/22 to 1/28/2018) | ||
+ | |||
+ | ===Snow 2021-2022=== | ||
+ | * This winter season the air temperature thermistor begins to fail so we'll need to correct data using probably downward longwave tempearture. | ||
+ | |||
+ | ==Partial Data Downloads== | ||
+ | Site Downloads / visits: | ||
+ | * Lily / Bob 6/28/2017 | ||
+ | * Lily 8/16/2017 | ||
+ | * Lily 8/20/2017 | ||
+ | * Bob 6/12/2018 | ||
+ | |||
+ | ==Info for metadata and file level metadata== | ||
+ | ===Instrumentation and Heights=== | ||
+ | * AT/RH @ 2.2m | ||
+ | ** Vaisala HMP155 | ||
+ | ** https://www.campbellsci.com/hmp155a | ||
+ | ** Temperature Range: -80 to +60 Celsius | ||
+ | ** Temperature Accuracy: (0.226 - 0.0028 x temperature)°C (-80° to +20°C) | ||
+ | ** Humidity Range: 0 to 100 % Relative Humidity | ||
+ | ** Humidity Accuracy: ±(1.2 + 0.012 × reading) % RH (at -40° to -20°C); ±(1.0 + 0.008 × reading) % RH (at -20° to +40°C) | ||
+ | * WS/WD @ 4.12m | ||
+ | ** RM Young 01503 | ||
+ | ** https://www.campbellsci.com/05103-l | ||
+ | ** Range: 0 to 100 m/s | ||
+ | ** Accuracy: +-0.3m/s | ||
+ | ** Startup threshold: 1.0 m/s | ||
+ | ** Wind Direction measurement band: 0 to 355 degrees | ||
+ | ** Accuracy: +-3 degrees | ||
+ | * Snow Depth @ Tripod | ||
+ | ** Campbell Scientific SR50A | ||
+ | ** https://www.campbellsci.com/sr50a | ||
+ | ** Resolution: 0.25mm | ||
+ | ** Accuracy: +-1.0 cm | ||
+ | * Snow Depth @ Lump | ||
+ | ** Campbell Scientific SR50A | ||
+ | ** https://www.campbellsci.com/sr50a | ||
+ | ** Resolution: 0.25mm | ||
+ | ** Accuracy: +-1.0 cm | ||
+ | * Net Radiation @ 1.5m | ||
+ | ** Kipp & Zonen CNR4 | ||
+ | ** https://www.campbellsci.com/cnr4 | ||
+ | ** Shortwave Measurement Range- 305 to 2800 nm | ||
+ | ** Shortwave Uncertainty in total daily - < 5% (The uncertainty values are for a 95% confidence level.) | ||
+ | ** Longwave Measurement Range - 4500 to 42,000 nm | ||
+ | ** Longwave Uncertainty in total daily- < 10% (The uncertainty values are for a 95% confidence level.) | ||
+ | * Skin Temperature @ 1.9m: | ||
+ | ** Apogee Instruments SI-111 | ||
+ | ** https://www.campbellsci.com/si-111 | ||
+ | ** Operating Range: -55 to +80 C | ||
+ | ** Accuracy: ±0.2°C (-10° to +65°C) & ±0.5°C (-40° to +70°C) | ||
+ | ** Field of View: 22° half angle | ||
+ | * Summer Precipitation | ||
+ | ** Texas Instruments TE525MM Metric model | ||
+ | ** https://www.campbellsci.com/te525mm-l | ||
+ | ** Accuracy: 1.0% up to 50 mm/hr | ||
+ | * Soil Moisture | ||
+ | ** Stevens Hydrasense II | ||
+ | ** https://amarillo.nserl.purdue.edu/ceap/metadata/Stevens_Hydra_soil.pdf | ||
+ | ** Range Dielectric Constant: 1 = Air to 78 = Distilled Water | ||
+ | ** Accuracy Dielectric Constant: ±1.5% or 0.2 whichever is greater | ||
+ | ** Soil Moisture Range: Completely Dry to completely saturated | ||
+ | ** Soil Moisture Accuracy: ±0.03 water fraction by volume in typical soil | ||
+ | ** Conductivity Range: 0.01 to 1.5 Siemens per meter | ||
+ | ** Conductivity Accuracy: ±2.0% or 0.005 S/m whichever is greater | ||
+ | ** Temperature Range: -10 to +65 Degrees Celsius | ||
+ | ** Temperature Accuracy: ±0.1 Degrees Celsius | ||
+ | * Soil Heat Flux | ||
+ | ** https://www.campbellsci.com/hfp01 | ||
+ | ** Sensitivity: 50 μV W-1 m-2 (nominal) | ||
+ | ** Accuracy: Within -15% to +5% in most common soils (12 hour totals) | ||
+ | * Snow Temperature | ||
+ | ** Thermistor - US Sensor PS302J2 | ||
+ | ** https://www.littelfuse.com/~/media/electronics/datasheets/leaded_thermistors/littelfuse_leaded_thermistors_interchangeable_thermistors_standard_precision_ps_datasheet.pdf.pdf | ||
+ | ** Range: -80 to +75 Degrees Celsius | ||
+ | ** Accuracy: ±0.1 Degrees Celsius | ||
+ | * Backup Air Temperature | ||
+ | ** Thermistor - US Sensor PS302J2 | ||
+ | ** https://www.littelfuse.com/~/media/electronics/datasheets/leaded_thermistors/littelfuse_leaded_thermistors_interchangeable_thermistors_standard_precision_ps_datasheet.pdf.pdf | ||
+ | ** Range: -80 to +75 Degrees Celsius | ||
+ | ** Accuracy: ±0.1 Degrees Celsius | ||
+ | * Soil Temperature | ||
+ | ** Thermistor - US Sensor PS103J2 | ||
+ | ** https://www.littelfuse.com/~/media/electronics/datasheets/leaded_thermistors/littelfuse_leaded_thermistors_interchangeable_thermistors_standard_precision_ps_datasheet.pdf.pdf | ||
+ | ** Range: -80 to +75 Degrees Celsius | ||
+ | ** Accuracy: ±0.1 Degrees Celsius | ||
+ | * Soil thermal conductivity | ||
+ | ** East30 Sensors Thermal Conductivity | ||
+ | ** Thermal Conductivity Accuracy: ±5% (Except at soil temperature range about -1.5 to 0.0 Degrees Celsius) | ||
+ | ** https://www.east30sensors.com/thermalconductivity | ||
+ | ** Temperature Accuracy: ±0.1 Degrees Celsius |
Latest revision as of 10:04, 28 October 2022
Contents
Data QA
- https://docs.google.com/spreadsheets/d/1ogCWyAmechT0oiUTxEl3KzThwVwx5LK7cjOV9SVK4oM/edit#gid=0
- August 2021 Summary:
- Replaced AT/RH in June
- All other met data looks good
- Thermal conductivity in pit
Extra Notes
- For data grouping routine there is something off with the pit temperatures... more specifically one set is reported as 'Avg' so I must not have correct headers.
- Just a historical note that the Top snow stand on the lump had been bent over sometime prior to Ken H visit in fall 2020... I sorta think this happened in the winter of 2019-2020 but need to circle back to that to confirm.
- Ken visit: 10/31/20 @ 19:40 is final data in Snow.dat.
Main
Latest
- Met Data QA'd through last visit (9/2018)
- Radiation Data QA'd through last visit (9/2018)
- Partway through processing soil.
Met 2017
- Initial teller top was an instrumented survey tripod (May 2016 to August 2016)
- A musk ox knocked it over and was discovered in July I think.
- permanent station in August 2016.
- The back up air temperature was not in place in 2016. The HMP155 is super susceptible to excessive moisture like found at Teller top and Utqiagvik. So, HMP155 fails almost every winter.
- Every time I review data I delete more subtle but error AT/RH/ DP data.
- Re-reviewing the first year data in January 2022 I have caught a few more things.
- original review included for the air tmemperature data & dew point from bottom station to weed out the in range but in error data as the HMP155 was failing in 2016. I am redoing the analysis now but made another dataset that includes longwave temperatures and skin temperature... that shows a bit more to cull from the record.
Met 2017
- HMP155 August 2017
- failure looks legit... not a data gap.
- HMP155 Cuts out on 10/25 8:00
- replaced AT portion with new AT thermistor.
- DP & RH left NAN
- HMP155 replaced on March 2018 visit
- Rain corrected through most recent download.
- Wind Speed / Direction
- Propeller replaced 6/28/2017
- No Rime fall 2017
Met 2018
- Wind
- all rime conditions corrected for through 3/2018 in QC spreadsheet.
- Rain NAN through snowmelt
- Radiometer leveled and desiccant replaced 6/12/2018
- Radiometer leveled again 9/24/2018
- Tipping bucket leveled 6/12/2018 & 9/24/2018
- AT/RH corrected through 3/30/2018 20:00 after sensor destroyed over winter season.
Met 2022
- Figured out this year that the HMP155 cable was the source of the bad data issue.
- However, musk ox near station in May 2022 visit prevented fixing until October 2022.
- Back up air temperature began to fail in 2022 and will need replacing, too (sadly).
Radiation 2017
- There is a gap 10/2/2017 0:00 to 10/2/2017 22:00
- I think this is when we removed the sensor from the field to change the desiccant.
- as in, the screws were stripped and so we drilled and retapped back in town.
- I think this is when we removed the sensor from the field to change the desiccant.
Radiation 2018
- All good through March visit.
- There is a weird thing in 2018 where the lower SW has values higher than the upper SW. I assume this is due to snow/ frost up top. Perhaps related to how close the snow surface is to the sensor this year resulting in a lower wind speed later in the winter.
Radiation 2022
- Sensor swapped for recalibration but I didn't modify the program so, proper data requires 2 key file changes for this interval.
- Changes:
- Shortwave sensors were already set up with data_type swrad so only a value change is necessary.
- Finally, Coef_2 for SW sensor units are now non-zero:
Coef_2 = OriginalCNR4_calib_value / Replacement_CNR4
- Longwave sensors remain data_type num
- Use additional sensor SBTempC_Avg.csv to:
- Compute longwave of the sensor body
- Set up a spreadsheet and create a new time series:
- Use additional sensor SBTempC_Avg.csv to:
- Longwave sensors remain data_type num
corrected_LW_step1 = [SBCTemp_C-> Watts (Stefan-Boltzmann Law)] - [LW From .dat] corrected_LW_step2 = [corrected_LW_step1] * [original_LW_Calibration] / [Current_LW_Calibration] corrected_LW_for_archive = [corrected_LW_step2] + [SBCTemp_C-> Watts (Stefan-Boltzmann Law)]
- Net Radiation needs to be recomputed after corrections are applied
- 2022-04-01 02:00:00 -- First data point with replacement unit
- Keyfile:
teller_top_micromet_2022-04-01_params_20-cnr4_090013.csv
- 2022-10-05 19:00:00 -- Last data point with replacement unit
- Keyfile:
teller_top_micromet_2022-10-05_params_20-cnr4_131352.csv
- All data for this interval is contained in data file:
teller_top_MicroMet-2022_10_05.dat
- Calibration Info 2016 to 2022-04-01 01:00:00 -- SN 120966
SW Upper -- 13.83 SW Lower -- 13.08 LW Upper -- 13.46 LW Lower -- 14.06
- Calibration info 2022-04-01 01:00:00 to 2022-10-05 19:00:00 -- SN 090013
SW Upper -- 15.12; Coef_2 = 13.83 / 15.12 SW Lower -- 10.51; Coef_2 = 13.08 / 10.51 LW Upper -- 7.40; Coef_2 = 13.46 / 7.40 LW Lower -- 6.72; Coef_2 = 14.06 / 6.72
- Calibration info 2022-10-05 19:00:00 and onward: -- SN 131352
SW Upper -- 12.70; Coef_2 = 13.83 / 12.70 SW Lower -- 12.44; Coef_2 = 13.08 / 12.44 LW Upper -- 10.35; Coef_2 = 13.46 / 10.35 LW Lower -- 11.70; Coef_2 = 14.06 / 11.70
Soil 2017
- 2017 heat flux plates prior to fall were being measured on wrong channel.
- 2017-03-12 02:00:00 Upper heat flux plate has a value that isn't ever flipped by the qc program.
- 2017-03-12 02:00:00 Lower heat flux plate has a value that isn't ever flipped by the qc program.
- Easy Perl one liners to accomplish this:
perl -pi -e 's/"2017-03-12 02:00:00",149.2/"2017-03-12 02:00:00",6999.00/g;' $TT_SOIL_DIR/outputs/SoilPit_HeatFlux_Lower.csv perl -pi -e 's/"2017-03-12 02:00:00",140.8/"2017-03-12 02:00:00",6999.00/g;' $TT_SOIL_DIR/outputs/SoilPit_HeatFlux_Upper.csv
- 2017 Soil pit soil moisture -- all corrected
- 2017 Soil heat flux plates also all corrected
- with Vlad determined only good VATP are 2/5/25/50/60/100
- Soil Temperature: VATP
- I think I have everything corrected though I don't entirely believe the 100cm probe.
- 8/16/2017 from Lily's field book page 50:
- VATP was heaved 9.5cm above standing water table. I pushed it back down
Soil 2018
- 2018 soil pit soil moisture -- all corrected through most recent visit (3/2018)
Soil 2021
- I've decided the top site experiences high heat fluxes in spring / summer and it seems likely the output is overranging the measurement thresholds set in the program.
Snow
Snow Depth 2016-2017 Lump
Season Starts: 11/3/2016 08:35
- Winter 2021 re-analysis and discussion:
- I'm still hung up on the precise start. If you look at the end of season zero offset, it meshes up with a season start of 10/26/16 18:00ish or 10/29/16 11:05ish. But from met and radiation data I think this might be rain / slush that melts but it's enough to smoosh down the vegetation briefly but it springs back up again. So, I guess I'm still going with 11/3 start.
- Albedo suggests no snow (centered on midnight) 10/26 to 10/28. Temperature on these days is around +2 C. So, precipitation in tipping bucket is probably rain. It looks like a litle snow fell and melted though looking at the std deviation of snow depth and the lack of noise in the snow depth data for a couple hours starting 10/26/16 1:40.
- albedo on the UTC late day of 10/27 & 10/28 @ 0:00 is back to no snow.
- 10/29 0:00 0.39 albedo but probably top dome covered in rain/slush/snow as the tipping bucket records a bunch.
- Net Radiation is also mostly positive over this interval. that changes after 10/31 0:00 where there seems to be clearing skies at night.
- Then no more real precipitation of note until 11/3 after 8:00 UTC.
- with respect to the snow depth signal over this say 10/26 to 11/2 0:00 interval it's noisy like exposed ground / the signal reflecting off grass rather than a snow surface. So, roughly I think any snow that fell in this interval melted.
- After sun is up on 11/2/2016 the albedo is 0.72 so looks like good snow. Value never drops below 0.6 in next several days. So pretty good confidence on that.
- However, still hard to pin down a start. WIth all that said, 11/2 to 11/4 is still a mess with grass / organics not completely buried. The tipping bucket shows no precip on 11/1 or 11/2 and then 9:00 (UTC) to 14:00 on 11/3 it's tipping. So maybe 11/2 albedo is skewed high by heavy clouds / frost. air temperatures are +2 to +4 C on 11/1 and 11/2 until late in day.
- So, I guess 11/3/2016 08:35 is official start but there is still ground exposed at the sensor as compared to below the radiometer. And it sort of looks like snow starts at 8:35 then transitions to rain around 15:00. If you go by the highest confidence quality numbers then it looks like snow at 8:35 which melts over the day and then a bit more falls by end of day (as illustrated by the declining standard deviation of snow depth.
- Then things are just kind of a push as to what is going on. Just minimal snow but likely snow covered until next event of note on 11/16/2016.
- So, going at the start of the season following whatever the furthest away high confidence Q value snow depths probably wouldn't be the worst way to go.
- November / December 2016 have data spikes in the high confidence Q values. I'm pretty sure this is due to the failure of the HMP155 air temperature which was the only AT active this first winter.
- If you look at the Column B snow depth values, which are adjusted for temperature there is a mix of 2 decimal data and lots of decimal data (TT_Lump_2016-2017.ods). The bunch of decimal data is the infilled from Teller Bottom air temperatures (or data logger panel temperature I can't remember at the moment). Anyway, the spikes look to be in the bunches of data where the HMP155 seemed to be working but wasn't actually.
- I've been seeing this in the spreadsheets periodically an interval in late winter, this time 3/12/17 2:00 to 2:55 where the data came in as text rather than date format. It only shows up scrolling by briefly but something to watch for in final outputs.
Season Ends: 05/15/2017 10:35
Snow Depth 2017-2018 Lump
Official Start: 10/20/2017 04:55
Winter 2021/22 Reanalysis and commentary
- In fall, air temperatures drop consistently below freezing 10/17/17 or 10/18/17 depending on how you feel about popping up over 0C
- Summer precip record ends the 16th.
- albedo is 0.14 to 0.25 through 10/19/17 3:00 (UTC) and then albeod on 10/19 after sunrise (so 21:00-23:00) suggests snow and following days look like 100% cover.
- general snow depth signal is basically just bouncing off leaves and such through the official start. I think you could make a case maybe for 10/19/17 12;45ish as an earlier start but the fifteen house in between these dates looks mostly like the leaves are crushed down a hair without any accumulation. Accumulation starts at the official start I believe.
- re-review pretty anti-climatic. Nothing much to clean up beyond erroneous high confidence quality number values here and there.
Official End: 06/04/2018 02:15
Snow Depth 2018-2019 Lump
Original Start: '10/29/2018 03:35 Revised: 11/08/18 10:55 Winter January 2022 Reanalysis
- Air temperatures don't begin dropping much below 0 C until 10/17/18 16:00ish
- Hover around freezing from then till 10/27/18 2:00
- Then they come back up and do a bit of melting it looks like.
- Albedo is barren ground until roughly 10/23/18 1:00 then inches up but at least under radiometer not full snow coverate until 11/10/18 22:00.
- Looking at air temperature albedo simultaneously it's clear there was some melting that took place the first week of november.
- Looking at the snow depth data it appears there were a couple periods of snowing then melting. 10/24/18 23:10 to 10/25 23:25 perhaps
- to my eye in the fall when the std deviation values for snow depth are below one I tend to thin that is snow even if it isn't amounting to anything.
- This isn't an absolute but the highest confidence reported readings tend to be the ones bouncing off the ground rather than reflecting off leaves. So can kind of use that as a proxy for snow if nothing else.
- Basically the first snow of note falls around 10/25 00:00 and melts, snows, melts cycles and then between 11/8/18 11:45ish and 11/9/18 21:35 there is some hard to pin down exactly accumulation but at 21:35, that is the beginning of 100% snow cover and you see that clearly in both albedo and snow depth data.
- Not much done beyond this start of season adjustment.
Official End: 5/31/2019 23:50 Revised End: 5/31/2019 23:00
Snow Depth 2018-2019 Tripod
- Winter 2021 re-analysis:
- For tripod depth shifted the start of the season earlier compared to what was originally in archive. Took a look at tipping bucket data and thought snow arrived sooner.
- I don't know what the shift is on 4/4/19 00:50 (UTC) values go from consistent 148.8cm to 144.5.
- I think this was somehow related to a station visit. Magnaprobe data ends at top 4/3/19 16:52... tz most likely akstd
- For end of data set in June 2019, standing water is present so the end of season value isn't zero-ish. Final val is 5.0cm
- I'm sure I must have knocked some rime ice off or something but pretty odd.
- to verify the time & place I popped open the 2019 end of winter kmz that I made once upon a time.
- I don't know what the shift is on 4/4/19 00:50 (UTC) values go from consistent 148.8cm to 144.5.
Start of Season
- 2018-10-24 16:15
Air temperature is around +1C at this time step. Precipitation falls into the tipping bucket but I think it is slush / snow. The snow depth goes from 0 to 6.3cm over three hours and the readings are erratic with respect to quality number. Albedo is 0.23 during the day on 10/23 and increases to 0.96 on 10/24. Longwave values are within 3 Watts suggesting warm clouds / low cloud cover etc.
Snow 2018
- Snow temperature profile
- Weird shift (looks like mux wonky 1/22 to 1/28/2018)
Snow 2021-2022
- This winter season the air temperature thermistor begins to fail so we'll need to correct data using probably downward longwave tempearture.
Partial Data Downloads
Site Downloads / visits:
- Lily / Bob 6/28/2017
- Lily 8/16/2017
- Lily 8/20/2017
- Bob 6/12/2018
Info for metadata and file level metadata
Instrumentation and Heights
- AT/RH @ 2.2m
- Vaisala HMP155
- https://www.campbellsci.com/hmp155a
- Temperature Range: -80 to +60 Celsius
- Temperature Accuracy: (0.226 - 0.0028 x temperature)°C (-80° to +20°C)
- Humidity Range: 0 to 100 % Relative Humidity
- Humidity Accuracy: ±(1.2 + 0.012 × reading) % RH (at -40° to -20°C); ±(1.0 + 0.008 × reading) % RH (at -20° to +40°C)
- WS/WD @ 4.12m
- RM Young 01503
- https://www.campbellsci.com/05103-l
- Range: 0 to 100 m/s
- Accuracy: +-0.3m/s
- Startup threshold: 1.0 m/s
- Wind Direction measurement band: 0 to 355 degrees
- Accuracy: +-3 degrees
- Snow Depth @ Tripod
- Campbell Scientific SR50A
- https://www.campbellsci.com/sr50a
- Resolution: 0.25mm
- Accuracy: +-1.0 cm
- Snow Depth @ Lump
- Campbell Scientific SR50A
- https://www.campbellsci.com/sr50a
- Resolution: 0.25mm
- Accuracy: +-1.0 cm
- Net Radiation @ 1.5m
- Kipp & Zonen CNR4
- https://www.campbellsci.com/cnr4
- Shortwave Measurement Range- 305 to 2800 nm
- Shortwave Uncertainty in total daily - < 5% (The uncertainty values are for a 95% confidence level.)
- Longwave Measurement Range - 4500 to 42,000 nm
- Longwave Uncertainty in total daily- < 10% (The uncertainty values are for a 95% confidence level.)
- Skin Temperature @ 1.9m:
- Apogee Instruments SI-111
- https://www.campbellsci.com/si-111
- Operating Range: -55 to +80 C
- Accuracy: ±0.2°C (-10° to +65°C) & ±0.5°C (-40° to +70°C)
- Field of View: 22° half angle
- Summer Precipitation
- Texas Instruments TE525MM Metric model
- https://www.campbellsci.com/te525mm-l
- Accuracy: 1.0% up to 50 mm/hr
- Soil Moisture
- Stevens Hydrasense II
- https://amarillo.nserl.purdue.edu/ceap/metadata/Stevens_Hydra_soil.pdf
- Range Dielectric Constant: 1 = Air to 78 = Distilled Water
- Accuracy Dielectric Constant: ±1.5% or 0.2 whichever is greater
- Soil Moisture Range: Completely Dry to completely saturated
- Soil Moisture Accuracy: ±0.03 water fraction by volume in typical soil
- Conductivity Range: 0.01 to 1.5 Siemens per meter
- Conductivity Accuracy: ±2.0% or 0.005 S/m whichever is greater
- Temperature Range: -10 to +65 Degrees Celsius
- Temperature Accuracy: ±0.1 Degrees Celsius
- Soil Heat Flux
- https://www.campbellsci.com/hfp01
- Sensitivity: 50 μV W-1 m-2 (nominal)
- Accuracy: Within -15% to +5% in most common soils (12 hour totals)
- Snow Temperature
- Thermistor - US Sensor PS302J2
- https://www.littelfuse.com/~/media/electronics/datasheets/leaded_thermistors/littelfuse_leaded_thermistors_interchangeable_thermistors_standard_precision_ps_datasheet.pdf.pdf
- Range: -80 to +75 Degrees Celsius
- Accuracy: ±0.1 Degrees Celsius
- Backup Air Temperature
- Thermistor - US Sensor PS302J2
- https://www.littelfuse.com/~/media/electronics/datasheets/leaded_thermistors/littelfuse_leaded_thermistors_interchangeable_thermistors_standard_precision_ps_datasheet.pdf.pdf
- Range: -80 to +75 Degrees Celsius
- Accuracy: ±0.1 Degrees Celsius
- Soil Temperature
- Thermistor - US Sensor PS103J2
- https://www.littelfuse.com/~/media/electronics/datasheets/leaded_thermistors/littelfuse_leaded_thermistors_interchangeable_thermistors_standard_precision_ps_datasheet.pdf.pdf
- Range: -80 to +75 Degrees Celsius
- Accuracy: ±0.1 Degrees Celsius
- Soil thermal conductivity
- East30 Sensors Thermal Conductivity
- Thermal Conductivity Accuracy: ±5% (Except at soil temperature range about -1.5 to 0.0 Degrees Celsius)
- https://www.east30sensors.com/thermalconductivity
- Temperature Accuracy: ±0.1 Degrees Celsius