April 18, 2018: More Science!

We had another warm day. One photo:

Looking west with Shoal Lake on the horizon. Town Island is at the left, Shragge’s Island is closest to the center of the picture, Devil’s Gap is towards the lower right. No real expansion of the open water, but some further loss of snow cover.

Now an update from Sean C. He and I agree that as of yesterday, our Mean Daily Temperature is above freezing.

First, his updated graph that finally gives 2018 a blue X, marking yesterday as this spring’s “Inflection Date” when we stopped making ice and started melting it.

The downward spikes indicate the length and severity of the winter, and now that we’re saying this one’s over, we can compare it to other recent winters. Not as severe as 2014, but worse than most. You can even make out our early April cold snap on the profile: it’s that tiny downward jab at the tip of the spike.

Next, Sean took a look at all the winters since 2003, and graphed them against each other.

For this graph, he lined up all the Inflection Dates together, at Day Zero, and then looked at how long it took to reach the same thawing index. The horizontal red line is his estimate of how much heat we should need to thaw the lake after a winter like this one. In the best case, it took 20 days to reach the equivalent of a full melt. The worst year took almost three times as long: 54 days.

By applying the average of 32 days, Sean found himself looking at an estimate of getting ice free on May 18th, the Friday of the Victoria Day weekend. In other words, right down to the wire for places like Clearwater Bay and Bigstone Bay.

Warning: this mathematical average does not directly factor in such things as abnormal sunshine, current, or snow cover.

I had a question: “Won’t it go faster in May than if we started in March?” You can now stop calling me Ice Captain and start calling me Captain Obvious.  Sean obliged me with this graph.

He knocked it off in a hurry, so it doesn’t have a pretty name, but it plots Thawing Index on the left edge versus Inflection Date across the bottom. The individual blue dots represent sample years and the diagonal line is a best fit for the overall trend. It shows, with a lot of variation from year to year, that the later you start the thaw, the faster the ice goes.

This fits with my old—but overly simple—notion that the thaw tends to even out. Keep in mind that until this year, I worked almost entirely from my aerial view of how much open water there was, and only regarded temperature as an influence on an inevitable process, rather than a fundamental factor. Sean has changed the way I look at it. Remind me to update the FAQ.

If we take into account both the thick ice from the cold winter, and the faster thaw due to starting late in the year, we get a rather more encouraging timeline. Instead of the 32 day average to melt a cold winter’s ice, we get a number that could be as low as 20 days to melt this thickness of ice when we start after mid-April.

That would put us at May 8, which is close to my estimate from before the cold snap. Hmm.

Neither Sean nor I have faith that that’s how it will play out. He remarked that, “…past performance doesn’t guarantee future profits.” He probably worries about how close that comes to the fastest thaw in his data. I don’t like how it fails to factor in a forecast for two more weeks of mean temperatures at or below normal.

Let’s allow more than three weeks, but less than four and a half. Four weeks would give us a lake entirely free of ice just days before the long weekend. Fine print: no guarantee is expressed or implied.  This offer does not apply to Shoal Lake. (Shoal Lake usually runs a few days later.)

Cheer up; even if it’s close, most lake dwellers don’t need the whole lake to be open.



April 17, 2018: Warm & Windy

We had sunny weather today, with a high of 7°C this afternoon. Better yet, it was breezy, with winds of up to 20km/hr from the north east.

It wouldn’t be realistic to expect one nicer day to make a big difference, and most of the photographs I took today look just like the ones I took yesterday, so I’ll only put this one up.

This is looking south over Devil’s Gap with Treaty Island stretching from near the center to the right edge. Beyond that, lots of ice out by Rogers Island, Town Island and so on. Zoom in, and you can see the ice roads still look pretty solid. Rat Portage Marina is visible at the lower left, (partly obscured by the digitally distorted propeller blade) and there’s still ice all around the docks.

However, to my eye, the ice looks distinctly more gray today. It’s not glaringly obvious in the photographs, but with the picture above you can zoom in on the lower right corner to check out the ice surface around Gun Club Island, and you might agree that it looks patchier.

Terra Satellite got a clear image today, so I have updated the satellite link. Usually the first patch of open water on Lake of the Woods big enough to “see from space” is down by Baudette, Minnesota, where the Rainy River spills into the lake and creates a dark patch on the south shore. That hasn’t happened yet.

The temperatures tomorrow are forecast to dip slightly, which will likely put Wednesday’s mean daily temperature right at the freezing point, but from Thursday on, we should be consistently trending warmer through to the end of April. Not fabulously warm, or even normal, but almost always above freezing, even at night.

That should mean we’re switching from making ice to melting ice. Finally.


April 16, 2018: Same Old

Today’s pictures are almost identical to Friday’s. We came in to Kenora on the same flight path, and nothing has changed.

When there’s a bit more open water developing, I’ll make a greater effort to range further out over the lake. For now, there’s nothing new to see.

Winnipeg River.

The Dalles, Myrtle Rapids. No significant change. In the distance, at the top of the picture, Shoal Lake, (on the right) and the US side of LotW (top center) are still pure white.

The powerline crossing.

No change.

Kenora to Keewatin.

No change.

Safety Bay.

No change. Note that you still cannot get a boat from Safety Bay to Channel Island, and that even the new ice near the Keewatin bridge has not melted.

However, I have high hopes for things to take a turn for the better tomorrow. Right now the forecast is calling for a high on Tuesday of about 7°C, and although temperatures sink a little on Wednesday, by the end of the week, we might be staying at or above freezing overnight. If that holds true, we should start to turn the corner.

Another guest contribution! Here’s a chart from John Harbottle, giving the ice-out dates he’s recorded for Kendall Inlet and Clearwater Bay.

John points out that the dates correspond well with the Bigstone Bay data from Madeleine Dreger. That makes sense to me, because both Bigstone and Clearwater are rather late melters.

Neither satellite has captured a cloud-free picture of Lake of the Woods since April 13, so I have not updated the link. Reminder: the sidebar and links do not appear in the emails, they are a feature of the full Ice Patrol website.

I am flying tomorrow, but I don’t really expect one warmer afternoon to make a visible difference. And remember, daytime highs in the single digits are still below normal for this time of year. But maybe we’ll lose some of that snow cover by the weekend- that would be nice.

April 14, 2018: Four Years Old

Yesterday was the fourth anniversary of Lake of the Woods Ice Patrol as a website. You can see the very first post here. Before that, I had an album on Photobucket that was technically public, but only a dozen or so people knew about it.

When I turned Ice Patrol into a WordPress blog, the site became searchable by anyone, and in four years the site has accumulated over 300,000 hits from all over the world and has acquired more than 2000 followers. Yesterday, I got my first interview with a US newspaper, the Duluth News Tribune.

Who’d have thought that many people would want to watch ice melt?

I’m celebrating the anniversary by rearranging the sidebar to make some of the popular features easier to find and use:

I’ve grouped the Archive links together. The Archive of Recent Months & Years offers a drop-down list for the spring months of the last four years. If you want to go further back,  Years Before 2014 is the old Previous Years list, renamed and moved up. Those are links to the old Photobucket Albums, which don’t have much commentary.

The link to Satellite Pictures has moved up to have its own heading. That lets me name the latest link with the date (and the name of the satellite). That’s handy because I don’t update the link on cloudy days, and now you don’t have to keep clicking on the undated (and dubiously named) “Recent Satellite Picture” link to see if there’s a new picture.

In other news, there was quite a lot of interest in Sean’s graph correlating mean daily temperatures to the thawing timeline. Rick Lord emailed me to say he found a scientific paper in the Hydrological Sciences Journal that talks about “the relationships between the mean winter air temperature and the duration of ice cover and its maximum thickness” on lakes in Poland that concludes that this measurement makes “a good base for forecasting and modelling.”

Sean tells me that he’ll be updating his data once we reach a mean daily temperature above freezing. According to the Weather Network, that should happen next Tuesday or Wednesday.


April 13, 2018: No Progress

We returned to Kenora from the north today, so I’ll start with a picture of the Winnipeg River. It’s about the only place I thought we might see improvement.

This is looking south west, with the Dalles near the middle of the picture. Click on the image to zoom in, and you can make out a little open water in the Myrtle Rapids area. Way off in the upper right corner, you can see Shoal Lake. It’s a sheet of white.

Next, a picture centered on Keewatin.

You’re looking south, with the Keewatin bridge close to dead center of the frame. Downtown Kenora is at the left, above the shiny propeller spinner. In the distance, you can see the Barrier Islands and beyond.

You’ll notice that the water on Safety Bay still doesn’t connect to the water in Keewatin Channel.

Here’s a closer look at that frozen patch.

Rheault Bay is right of center, just above the aircraft’s nose. There’s a little open water at the lower left, in the area of Mackie Island, but it still doesn’t reach beyond the tip of Yacht Club Island, and if you zoom in for a closer look, you can clearly see quite a lot of the water has refrozen. In the middle distance, the patch of water around Channel Island hasn’t expanded.

Last, a better look at Darlington Bay and Safety Bay.

Looking east, with the Keewatin bridge and the Kenora Forest Products stud mill in the foreground. There’s a lot of refrozen surface visible between the Keewatin bridge and nearby Lowe’s Island. Zoom in to see more fresh ice closer to Kenora, near the top center of the picture.

It was -10°C early this morning, and our afternoon high was -2°C, so it’s no surprise that nothing much melted. Next week should be better. Starting Tuesday, we might see days with more hours of above freezing temperatures than below.

April 12, 2018: Predictions

The recent correspondence with Sean, who used mean temperatures to extrapolate an ice-out date, got me thinking about all the different ways people try to predict the thaw.

My friend Linda reckons on about four weeks from the time the ice roads close until she can get to her camp on Treaty Island. That’s not a bad method; when the ice roads close, it’s because the ice is weakening.

My co-worker Cam is an ice fisher, so he pays a lot of attention to the thickness and quality of the lake ice. Way back in early March, he was talking about how it could take a long time to melt this year, and he even said the May long weekend could be jeopardized.

By mid-March, I was seeing nearly normal amounts of open water, so I felt more optimistic than Cam. I made allowances for the snow cover and thicker ice, but still felt we’d be clear in slightly more than six weeks, or around May 6-10. But then our little cold snap at the end of March turned into a prolonged stretch of very low temperatures, and my six-week timeline started to slip away.

Long-term forecasts showed temperatures persistently below normal, which made it even harder for me to figure out what might happen.

Then Sean wrote to say how he’d used temperature records to compare this winter to other recent ones. In his analysis, this winter resembles the winter of 2014/2015. In the spring of 2015, it took about four weeks (27 days) for the lake to melt after the mean daily temperature rose above freezing.

The Weather Network has a fourteen day forecast, and it looks as if that shift may happen on April 17 or thereabouts. If that holds up, Sean’s model would suggest an ice-out date of May 14 or 15.

I’m going to hedge a little, because of the snow cover, which is heavier this year than it was in the spring of 2015, and also because temperatures are still forecast to remain cooler than normal.

As of today, I feel Lake of the Woods could be ice-free around May 15-18, just days before the May long weekend.

On the other hand, you could argue that because this year is thawing later than 2015, things could happen faster. May is typically warmer than April, after all. Certainly the right combination of rain, sunshine and warm breezes could speed things up.

Remember: most lake dwellers don’t need the lake to be entirely ice-free to get to camp.

In the end, we’ll have to wait and see. Which brings us back to aerial photographs, my preferred way to show what’s actually happening. Stay tuned.

April 11, 2018: Science!

I didn’t fly today, so no new aerials. I’ll be talking to Ken O’Neil on Q-104 tomorrow morning, and then I’ll be flying on Friday, so I hope to have fresh photographs up for the weekend.

In the meantime, some historic data, and what it tells us about this year.

First, I heard from Madeleine Dreger. She has a place on Bigstone Bay, and she’s been recording the date of the ice-out there for thirty years. That’s great, because Bigstone is one of the last areas to let go each spring, so her dates are pretty close to the dates I have for Lake of the Woods overall. So I plotted all her numbers as a reflection below my own brick graph. It looks like this:

Brick Graph with Bigstone data corrected

As you can see, with more data, the bell curve is more pronounced. There’s a strong tendency for the ice to go out in late April or early May, but some years are better or worse. Insert ominous music here.

Some people have wondered why I use five-day periods, rather than specific dates. It’s because I’m never flying over the lake at the exact moment the last bit of ice melts. I came close one year, but usually I see ice one day, and none three days later, and I have to guess.

Speaking of guessing, each spring I look at Safety Bay, and when it starts to open up, I begin making regular patrols. That backfired badly this year, because we got to that point in March, and then it turned cold again.

Today I got an email from Sean C., who has a better way. He sent me a really cool graph that shows how the ice-out depends on the severity of the winter and the warmth of the spring. I’ll provide his full explanation, and then I’ll summarize some of the highlights.

Here’s his explanation.

I decided to take a look at some data to see if I could create a forecast of the date we are expecting Ice out on LOTW to see if we run the risk of a spoiled May Long weekend on Corkscrew. I shared it with some friends and they suggested sharing it with you.

I did a Freezing or Heating index over the last 5 years (including the winter of 2013/2014 since it was as bad as bad gets). The analysis is fairly simple. Its just a summation of all the daily mean temperatures through the winter for the freezing index and the same through the summer for the heating index. This data is used to determine how deep frost will penetrate into the ground in the winter as well as the thaw in the summer. I don’t know the correlation with ice depth and its subsequent thaw but the data speaks for itself.

In the attached graph, I plotted the cumulative heating and freezing indices for each season. The red dot is the date which your website has ‘called it’ for each season. Looking at the data, this winter was about as cold as the winter of 2014/2015. The Cyan X represents the date when the daily mean temperature was consistently above 0 which we can generalize as more heat was in the air than cold. So in the spring of 2015, we needed 196 heating index days to get to the ice out date. Using that same assumption we will need roughly 200 days this year. Right now, our daily mean temperatures are still consistently negative. So once we start seeing daily means above 0, it will be easier to forecast but right now I would say it is going to be close. It took from April 7th, 2015 when the mean temps were above 0 until May 3, 2015 to achieve those 196 heating index days which was 27 days. So a safe bet would be that it will take 27 days this year based on the temperatures we have seen. So right now its looking like a close call for may long weekend but as long as we start seeing daily mean temps in the positives in the next week or two, I think we will be ok.

Here are the highlights:

The upward spikes show how long and hot the summer was. 2015 and 2016 were hot summers. 2013, not so much.

The downward spikes show how long and cold the winter was.  2013/2014 was nasty, 2016/2017 was mild. This winter has been cold, more like the former.

The blue X’s mark the date when the temperatures switched over to mostly above freezing, and the red dots show when the ice went out.

Sean used those past dates to work out how long it takes the ice to melt after a winter like this one. See his last paragraph, above. Sean’s calculations do not include such short-term variables as snow-cover, sunshine, wind or rain.

Here’s the thing: we haven’t got our blue X yet this year, because we’re still spending more time below freezing than above. When we do reach that point, we’re still going to need roughly four weeks to open the lake. Let’s hope we can get started soon.