Deep diving into buoy data

[uniquorn]

052460 dtc is not working reliably but 551610 and 569620 are. Here is the difference between top of buoy temp and air temp.

Location on worldview   https://go.nasa.gov/3Q8rZfg  (light contrast)
From top down: 052460, 569620 then 551610

569620 peak buoy temps have been a lot warmer than air recently.


nullschool weather

[uniquorn]

A reminder of 569620 since deployment on sep3. Here we are looking at the temperature difference over 2cm along the length of the vertical buoy. In particular the effect of the snow layer. After a long winter protecting the ice from the cold it is now trying to protect it from the warmth.

The second image shows more detail of the many layers within the snow layer and how they change over time with air and ice temperature.

The third image gives more numerical detail of just the snow layer with the estimated ice surface as the thicker light blue line. The coldest day at the top of the snow layer was jan30 at -39.5C. The bottom of the snow layer was 16.5C warmer at -23C.
The trend lines are buoy temp above the snow layer(red) and near ice surface(blue).

[uniquorn]

WHOI TOP2 is co-located with 569620 with itp127 nearby. The profile parameters change during early march and the frequency of profiles changed from 4hourly to daily to 2hourly so the date labels may not be accurate. Anyway we missed cold eddy on apr24.

water temps at 2m depth were very slightly warmer may12-jun6 with what looks like some slight warming from below during may30-jun4.

The SAMI temp data on ITP127 doesn't correlate well.

[uniquorn]

<>
Location on worldview   https://go.nasa.gov/3Q8rZfg  (light contrast)
From top down: 052460, 569620 then 551610
<>

Had to see what they looked like drifting on worldview
jun2-9  (13.7MB)
probably should have put the other 2dp in

[uniquorn]

Journal of Geophysical Research: Oceans
Research Article   Free Access
Multiyear sea ice thermal regimes and oceanic heat flux derived from an ice mass balance buoy in the Arctic Ocean
Ruibo Lei, Na Li, Petra Heil, Bin Cheng, Zhanhai Zhang, Bo Sun
First published: 31 December 2013
https://doi.org/10.1002/2012JC008731

Abstract

[1] The conductive and oceanic heat fluxes and the mass balance of sea ice were investigated utilizing an ice mass balance buoy (IMB) deployed in the Arctic Ocean. After IMB deployment, the ice thinned from 1.95 m in late August to 1.46 m by mid-October 2008. From then on, ice growth until mid-June 2009 increased the ice thickness to 3.12 m. The ice temperature and consequently the conductive heat flux at the ice surface exhibited persistent high-frequency variations due to diurnal and synoptic-scale atmospheric forcing. These signals propagated downward with damped magnitude and temporal lag. The competition of oceanic and conductive heat flux dominated the low-frequency variations of ice growth. However, high-frequency variations in ice growth were controlled largely by the oceanic heat flux. From mid-November 2008 to mid-June 2009, the average oceanic heat flux along a track from 86.2°N, 115.2°W to 84.6°N, 33.9°W was 7.1 W/m2. This was in agreement with that derived from an IMB deployed in 2005, about 1.5° to the north of our buoy. We attributed the relatively high oceanic heat flux (10–15 W/m2) observed during autumn and early winter to summer warming of the surface ocean. Upward mixing of warm deep water, as observed when our buoy drifted over the shallow region of the Lomonosov Ridge (85.4°–85.9°N, 52.2°–66.4°W), demonstrated the impact of bathymetry on the oceanic heat flux under ice cover, and consequently on the basal ice mass balance.

extract

3.2 Sea Ice Mass Balance

[11] Snow accumulation at the IMB site was 0.08 m from 22 August to 10 October 2008, due to either new or blowing snow. Melt at the ice base continued until 10 October 2008, yielding a cumulative loss of 0.49 m resulting in an ice thickness of 1.46 m with a mean melt rate of 0.01 m/d over 49 days. The latter corresponded to latent heat flux of 29 W/m2. Our data demonstrated that the ice base determined by the under-ice sounder and by the thermistor string matched well each other by mid-October 2008. This was evidenced by (1) the freezing front extending to the ice base, (2) the temperature at ice base remaining at the freezing point, and (3) the ice temperature gradient having been established across the basal ice layer (Figure 3f). Although this temperature gradient was very low, it distinctly differentiated this layer from the oceanic mixed layer underneath the ice. The ice growth rate was initially sluggish during October 2008. Post mid-November 2008, the ice growth rates repeatedly exceeded 0.01 m/d (Figure 3g). Thus, the transition from basal decay-freeze balance to distinct growth took about 1 month while heat was released by refreezing of brine pockets and increasing solid fraction within the ice. From 10 October to 12 November 2008, the heat released from ice cover was estimated to be 6.5 MJ/m2 by using the observed ice temperature by the IMB and the ice salinity and density derived from the ice core measurements. Perovich et al. [1997] found a similarly extensive transition for a floe in the Beaufort Sea during both, the deployment year of their IMB and the subsequent year. Thus, we inferred that this process was not related to the disturbance of the borehole by buoy deployment. Our data showed that the ice grew from 1.46 m on 10 October 2008 to an annual maximum of 3.12 m on 11 June 2009, yielding a mean ice growth rate of 0.007m/d. Surface ice melt started earlier than basal ice melt, and was detected after 13 June 2009, when the temperatures exceeded 0°C near the upper surface.

[12] Lowest sea ice temperatures were generally encountered in the upper layers in response to atmospheric forcing. Extended atmospheric warming events lasted for about 8–15 days in the first half of November 2008 and late January 2009 (Figure 3b). They resulted in marked increases of the ice temperature over the entire ice depth (Figure 3f) and a pronounced decrease of the ice growth rate (Figure 3g). In the deeper layers, the high-frequency variations of ice temperature were damped, and with substantial lag to those at the surface. During the initial ice growth season, as the freezing front approached the ice base, the brine channels refroze from the top of the ice downward, which delayed the ice cooling and resulted in a low ice temperature gradient, especially for the low layer. The linear fit gradient within the basal 0.40 m ice layer increased gradually from mid-November 2008 onward, due to further refreezing of brine pockets (Figures 3h and 3i). The refreezing process of brine pockets within multiyear ice during the initial ice growth season clearly differentiates it from first-year ice [Lei et al., 2010]. For comparison, we also calculated the temperature gradient of the basal 0.40 m ice layer at the HOTRAX site (Figure 3i). This temperature gradient showed a similar seasonality as that at the CHINARE site. Both had relatively small values in November. They distinctly increased in December, and peaked at about 7.1–7.4 K/m in early January. Both series reverted to relatively small values by the end of June due to the seasonal warming of the ice temperature and the decrease in ice growth rate. Both processes might result in a lower solid fraction in the basal ice layer [Wettlaufer et al., 1997; Hunke et al., 2011]. From May 2009 onward, the complete ice column warmed at the CHINARE site, with a distinct lag in lower layers because of thermal mass of the ice and the brine within the ice. With summer impending, the ice crystals surrounding brine pockets melted and enlarged the latter. The heat, afforded to melt the ice, in part counteracted the gain of thermal heat from atmosphere, hence delaying ice warming.

Interesting that they also get -4C at the bottom of the ice.

[uniquorn]

Summer is coming. 569620 ice warming from top and bottom and beginning to form the now familiar temperature contour.

Close up on June temps

Freeze/thaw conditions. Not quite melt pond weather today. Maybe some fresh snow.
The cold snow layer on jun4 possibly some phase change? Doesn't make sense otherwise.

[Jim Hunt]

The cold snow layer on jun4 possibly some phase change?

It looks as though the snow melted adjacent to the thermistor string, but not under the surface sensor?

A adjacent webcam like in the good old days would come in handy!

[uniquorn]

The cold snow layer on jun4 possibly some phase change?

It looks as though the snow melted adjacent to the thermistor string, but not under the surface sensor?
A adjacent webcam like in the good old days would come in handy!

Looking more carefully, air temp on 06/04:16h was -5.75C but top of buoy was +4.75C. The sunny day just didn't warm the snow up. The arrival of consistently warmer air on jun7-9 is the onset of snow melt imho. Still looks like refreeze at surface during low sun though. Probably some cloud over the last day.
Location is near terra modis swath but that could be the floe.

[uniquorn]

Located 569620 on S1A image from yesterday.
S1A_EW_GRDM_1SDH_20220613T164847_B84D_N_1.final.jpg

06/13:16  81.01528   -143.866288
white dot near centre
wv  https://go.nasa.gov/3HjPe1W

[uniquorn]

052460 still occasionally sending dtc data.

[uniquorn]

551610 is the first buoy to show signs of bottom melt, losing 2cm on jun11 despite an estimated 12cm of snow on top. Will probably waver at this level before significant melt begins.

[uniquorn]

1. Focusing on jun8-15 we can adjust the temp contour scale to match air temp from -4C to +2C which gives us a more detailed view of the snow layer and top of the ice. As Jim suggests above, heat from the top of the buoy is probably affecting the snow layer.

So we probably have a small melt pond forming around the buoy on jun9 which refreezes 'overnight' but melts again by afternoon jun10.

Refreeze 'overnight'  on jun11-12 again during a cloudier day then some warming in the 'snow' layer....

Melt pond starting to affect the ice by jun14, then more refreeze.

2. Note that sfc distance increases with melt.

[uniquorn]

2. 569620 has (my) estimated 25cm of snow, so much thicker. Freeze/thaw near the buoy down to maybe 12cm. Perhaps some brine drainage at the ice bottom late yesterday.

3. Surface distance was doing my head in so I've reversed the scale (red on right

[uniquorn]

052460, snow loss due to melt or drift ~6cm of ~20cm.
No ice thickening since jun7

[uniquorn]

The ocean beneath TOP4 (and 551610) warming up a little from lows of -1.61C in april to a high of -1.49C on jun16. Could be due to the proximity of the shelf break or the surprising amount of open water just visible under cloud (or both).

Last position on 2022/6/20 23046 UTC : 78.2782° N, 131.4957° W
rough loc  https://go.nasa.gov/3tM4Bug

Location from clear day on jun15. Hopefully I'm wrong and it's on a larger floe.
https://go.nasa.gov/3OacvGf

[uniquorn]

The ocean beneath TOP4 (and 551610) warming up a little from lows of -1.61C in april to a high of -1.49C on jun16.<>

Looking at 551610 we see 2cm of bottom melt or ablation on jun18 and some snow melt or drift.

Interpreting temps in the snow layer is tricky but I think we can assume some warming from buoy to ice (or melt pond) during sunny days due to horizontal advection and vice versa under cloud.

Taller format to show small thickness changes.
15 new buoys awaiting deployment

[uniquorn]

So with 4cm bottom thinning 551610 is the first of the 3 SIMB3's to show onset of bottom melt this year, starting before the snow layer has fully melted at an overall ice thickness of around 2m.

The diff between top of buoy temp and air temp suggesting some sunny days recently, but this buoy tending to confirm that warm air appears to affect the snow layer temps more than sunlight so far this season.

From 569620 up thread

Looking more carefully, air temp on 06/04:16h was -5.75C but top of buoy was +4.75C. The sunny day just didn't warm the snow up. The arrival of consistently warmer air on jun7-9 is the onset of snow melt imho.<>

I'd appreciate a second opinion on that.

edit: error on the 551610 surface distance chart so corrected and updated it to jun23. Will cover in detail down thread.

[Brigantine]

From 569620 up thread

Looking more carefully, air temp on 06/04:16h was -5.75C but top of buoy was +4.75C. The sunny day just didn't warm the snow up. The arrival of consistently warmer air on jun7-9 is the onset of snow melt imho.<>

I'd appreciate a second opinion on that.

What's the sampling rate on the thermistors? Is it 8-hourly? 4-hourly?

What I see is a suspiciously high (~ly negative) dT/dt gradient after that 06/04 event.
The time of day that normally has the highest daily temps (20h?) is cold that day.
(not that an 18⁰ difference in solar elevation is exactly night & day)

Is that +4.75 just one sample >>0⁰C with <0⁰C immediately before and after? If it was an ITP profiler doing that, I'd be pulling up the engineering graphs for a quality check. Any particular values in that average stand out?

Was anything really +4.75 that day? Like you say, the snow didn't seem to notice it

[uniquorn]

Thank you for your interest Brigantine.

What's the sampling rate on the thermistors? Is it 8-hourly? 4-hourly?

Sampling rate is 4hr

Is that +4.75 just one sample

chart attached. I think the dtc is in good condition and the buoy above surface warms up more when it is sunnier, cools more than air temp during low sun or heavy cloud.

Deployment area is probably flat, it was for 551610.

I think south facing ridged areas will be melting on sunny days, flat areas relatively unaffected till the consistently warmer air arrives.

[uniquorn]

551610 temps and deployment image for ref

[Brigantine]

Ok, looks legit. So it just does that sometimes. And up to that point 16h was the more common time of day for high temps.

Out of curiosity, in say early April at night, at times when there was clear(ish) sky and no sun how far the buoy temps drop below the air temps?

[uniquorn]

569620 winter temps

edit: though I see the drop below air temp as due wind chill rather than clear sky.

[Brigantine]

[Edited]

Thanks. That's a pretty consistent ⁻2-⁻3K max temperature difference from radiative cooling. That makes sense compared with the big spikes in direct sunlight. Up until the equinox, you even see a ⁻1K difference when the air temps are high.

In the first half of April there's also a fairly consistent offset. The thermistors (if they spike at all) spike and then drop again many hours before the air temps reach their daily high. I'm guessing the thermistors peak at (local) noon, and the air temps peak late-afternoon.

[551610] DTC0 and DTC1 seem to stick close to air_temp - 1K compared to the other DTCs.

Wind chill only drops the temperature of things that are internally heated e.g. mammals. When the air temp is warmer than the object, the effect of strong wind would be more like fan-bake

[uniquorn]

Thanks. Will check out radiative cooling.

[uniquorn]

So 551610 surface distance charts from spreadsheet may have been incorrect for some time.  The temperature contour charts were correct and correlate with cryoinno.
Only 2cm of snow left a couple of hours ago.
Possibly 3 fan bakes in a row.

nullschool temp jun21-23