The graph you post is deceptive because it is so far out of date — it stops at 1990 — and it is not clear what units the measurements are in. Your point, I gather, is that the satellite extent reporting normally starts at 1979 (a local high) instead of 1974 which appears much lower on the graph. I don’t know why reporting normally starts in 1979 (uncertainty about earlier data?) but 1974 is only a little lower than 1984 on the graph. If you look at the differences between 1984 and 1979 on up-to-date datasets, you can see that they aren’t great compared to recent years e.g. 2016 as you can see using this tool: http://nsidc.org/arcticseaicenews/charctic-interactive-sea-ice-graph/ Just looking at the summer minimums we see:
1979: 6.9 km² * 10⁶
So including 1973–1979 would not change the overall conclusion.
Yes, piomas is a model but it is not just a number created by geeks in a basement computer lab. It takes input from satellite extent data and other observations. Obviously it would be better to have a direct measurement of sea ice volume but piomas is the best you can get — better than the altimetry-based approach which cannot distinguish ice from snow as demonstrated by the wide discrepancy of Zwally from direct (GRACE) mass measurements of the antarctic ice sheet. http://www.realclimate.org/index.php/archives/2015/11/so-what-is-really-happening-in-antarctica/
Your claim on increased evaporation from reduced ice coverage sounds plausible but no evidence that this cooling effect is of the same order of magnitude as the deceased albedo warming effect and it it is supposed to be a stabilizing factor on sea ice, it obviously isn’t working.