The article shows a logarithmic trend (i.e. doubling your income increases your happiness by a fixed amount), which is a straight line on the log plot. This insinuates (if the trend continues, which is assumed from the results) that there is no point at which increased income won't make someone happier.
The traditionally-held view is that there is an income (I've heard something like 70k in the USA) where the plot should become a horizontal line, i.e. increasing your income increases your happiness by zero.
These are very different conclusions. The plot isn't really falling "significantly towards the top of the scale". Log is the correct way of looking at income, because people consider changes to their income proportionately to their current income (for example, a 5% raise to someone making 1M dollars a year is about as significant to his life as a 5% raise for someone making 100k, even though the actual additive amounts are very different).