On May 31, an inland tsunami swept across the south and east shores of Lake Manitoba. It was rapid and devastating. Driven by hurricane-force winds, the lake rose two feet in less than an hour and more than four feet in an afternoon.
My neighbour crawled from a bedroom balcony onto a sandbag dike and into swirling waters to barely escape the rapidly advancing tide. Others along the lakeshore near St. Laurent were rescued by zodiac, backhoe and even helicopter.
That no one lost their life was a near miracle and much credit goes to fast-acting municipal officials who co-ordinated the evacuation.
This unfolding drama, though, was largely ignored when it happened. It occurred at exactly the same time the NHL returned to Winnipeg.
The effect on lakeside communities is still not fully known but massive and far from over. The lake will continue to swell for another month to its highest-ever recorded level and it will remain high for many months, if not years.
Unfortunately, the worst is far from over. More damage will likely occur when the usual fall storms hit the beach, and even next spring when shoreline ice pileup along a lake that sits underneath flooded structures will destroy many of the homes and cottages that survive the fall.
Those who have lost homes, cottages, farms and ranches are disconsolate and angry. I and many of my neighbours on Lake Manitoba were in a state of shock after seeing the destruction along the lakeshore.
And we were all asking one question: Why did this happen?
The answer is complex. Some, including provincial government officials, have argued that the phenomenon is natural.
This is partly true. Given the extraordinarily high river levels and heavy rainfall, Lake Manitoba would naturally (that is, without the Portage Diversion or Fairford Water Control Structure) have suffered extensive flooding, perhaps even comparable to the flood of 1955, when the lake reached 816.3 feet at Delta Beach.
But the landscape of southern Manitoba has been so radically altered by flood-control measures such as dikes, reservoirs, diversions and floodways, it is hard to know what is natural any more.
Moreover, to play the what-is-natural game is just a hair's breadth from blaming the victim. If we do play that game, then remember that Portage la Prairie and everything downstream to Winnipeg is flooded, too. The Portage Diversion was used to avert that "natural" outcome.
So a simpler and more direct question is what role did the Portage Diversion play in the flood on Lake Manitoba?
The 1955 flood provided the impetus to build the Fairford Water Control Structure (FWCS), which became operational in 1961. It increased outflow from the lake, allowing greater control of water levels and, most importantly, averted floods such as that which occurred in 1955. Until this year, it had worked well.
Over the last 50 years, the lake has remained in a narrow and actively regulated range between 810 and 813.5 feet, rarely even falling outside its ideal range of 810.5 to 812.5 feet. It never reached its flood level of 814 feet. That compares to levels above 816 feet reached prior to the completion of the FWCS.
The key to managing lake level was that the maximum outflow from the control structure normally exceeds the maximum inflow from natural inputs, in particular the Waterhen (about 13,000 cubic feet per second) and Whitemud rivers (about 3,500 cfs). The maximum output of the FWCS ranges from about 12,000 cfs to 19,000 cfs, depending on lake level.
If natural inputs do not exceed outflow capacity, keeping the lake level stable is simple.
Precipitation and evaporation are wild cards. Both are major contributors to the water budget but harder to measure.
Evaporation accounts for about half the water loss from the lake over the year, but acts too slowly in the short term to help much when water levels rise rapidly, as they have now.
In most years, there is surplus outflow capacity from the FWCS that can accommodate an added and artificial input from the Portage Diversion.
That averages 250,000 acre feet per year, representing less than 1/16th of the natural inflow to the lake.
But by the time it is all over this year, the Portage Diversion will have added an extra four million acre feet to Lake Manitoba, doubling the normal annual input to the lake in just four months. The consequences are easily predictable.
The Portage Diversion was designed to protect Winnipeg -- a city built on the flood plains of two rivers -- and points upstream on the Assiniboine from flooding.
It diverts water from the Assiniboine River into Lake Manitoba and in non-flood years there is no direct connection between the two unless the diversion is open.
But the Assiniboine River is higher than Lake Manitoba. A little more than 2,000 years ago, it drained directly into Lake Manitoba and not the Red River, as it does today.
Now when the Assiniboine floods and overflows its banks, as would have occurred this year without the operation of the Portage Diversion, some of that water ends up in channels on the Assiniboine's old flood plain and heads toward Lake Manitoba.
How much is uncertain. A best guess is that when the Assiniboine floods, about 20 per cent of the flow would end up in Lake Manitoba even without the diversion.
One can now estimate the impact of the Portage Diversion on Lake Manitoba water levels.
The surface area of Lake Manitoba is normally 1.14 million acres: given this year's flooding it is a bit bigger now. As of June 13, the Portage Diversion had sent three million acre feet of water into Lake Manitoba, more than double the previous record of 1.42 million acre feet in the flood of 1976.
Spread that water over Lake Manitoba and it raises the lake level 2.6 feet.
We now need to reduce that 2.6 feet by the amount the lake that would have risen from Assiniboine River overflow even without the diversion: that is perhaps half to two-thirds of a foot.
What is left over, about a two-foot rise in the level of Lake Manitoba, is what is directly attributable to the operation of the Portage Diversion this year.
High flows through the diversion began on April 15 when Lake Manitoba sat at 813 feet. That level is now three feet higher.
Roughly speaking then, 2/3 of the rise can be attributed directly to the operation of the Portage Diversion.
More importantly, without the diversion, the lake just reaches its flood level of 814 feet, the point at which serious damage begins.
Last year, three windstorms on Lake Manitoba, in May, July and October, reached wind speeds comparable to those of this year's storm. Because the lake then was more than two feet lower than it was on May 31 this year, those storms caused comparatively minor damage.
Why did this year's disaster happen? Quite simply the inflow capacity of the Portage Diversion far exceeds the maximum outflow of the Fairford Water Control Structure, particularly when the FWCS has to handle natural inputs as well.
Since 1970, when the Portage Diversion first went into operation, we have been playing Russian roulette with Lake Manitoba water levels.
This year we lost. The residents around Lake Manitoba never stood a chance.
Scott Forbes is a behavioural ecologist at the University of Winnipeg. He has owned lakeside property at Twin Beaches for the last dozen years.