First, it's important to set the stage. The Western Aleutians, which include Kiska, have some truly awful weather. One source I found claimed that there is an average of only eight clear days a year. Whether that's true or not, the predominate summer weather pattern is heavy fog, clouds, and cold rain. Temperatures are extremely low- most of the Aleutians are above the 10C isotherm, meaning the average temperature of the warmest month is less than 10 degrees C, which is more typical of areas in the Arctic Circle. It's not a hospitable area, even in August.

The Allied landing on Kiska was following on the heels of the invasion of Attu, an island right nearby, whose garrison had not been withdrawn. The defenders of Attu, numbering less than 2,500, had left their shoreline positions in favor of fortified areas in the highlands and mountains, so that the landing forces initially proceeded unopposed. However, the Allies then ran into fierce opposition from these elevated strongpoints, with the Japanese using the terrain and weather to shield themselves from air and naval attacks and exacting a withering toll. After weeks of battle, the remaining Japanese defenders executed a massive banzai charge, with many of the survivors committing suicide with grenades afterwards. Nearly the entire Japanese garrison died over the course of the battle, and the Allies suffered almost 4,000 casualties.

Now we come to Kiska. Larger, better fortified, and with twice the number of men defending it. After the horrific experiences on Attu, Allied forces were expecting more of the same. When aerial reconnaissance data (hampered by the weather, of course) began to indicate a Japanese withdrawal, the assumption was that they had moved into the mountains to execute the same sort of strategy as on Attu. So when the landings proceeded unopposed, it only served to confirm this prevailing theory. Psychologically, it would have felt like a real invasion to the Allied troops on the ground- constant naval and air bombardment, the chaos of a landing, and of course, the fog to conceal the supposed Japanese defenders. In such an environment, a panicky soldier shooting at half-glimpsed silhouettes and sparking a firefight with friendlies is unfortunately all too easy.

In total, the Allied forces suffered 92 killed and 221 other casualties- of these, 70 KIA and 47 WIA were from a destroyer that struck a mine, meaning the landing forces had 22 deaths and 174 wounded. Many of these casualties can be attributed to non-combat factors- the landing force totaled 34,000 soldiers, and a certain number of casualties can be expected due to disease, frostbite, hypothermia, accidents and such. The Japanese also had planted explosives and other booby traps as part of their defenses. The 10th Mountain Division (of which one regiment, the 87th, was part of the invasion force) history states that 11 men of the regiment were killed due to friendly fire and 7 due to booby traps and accidents.

Most or all of the friendly fire casualties appear to have occurred in firefights between U.S. and Canadian forces, which would be the most likely place- fratricide tends to occur along boundaries between different units, especially those in different services. Friendly fire is far more common than people think- sources vary, but the total proportion of U.S. WWII casualties due to friendlies could be reasonably estimated at 10-12%. The only difference in this case is that, unbeknownst to the men on the ground, there was no actual enemy there to justify these losses.