Why are these diodes able to handle less current at lower PWM duty?

http://www.nxp.com/documents/data_sheet/PMEG3010ER.pdf

Page 7 figure 9 shows the forward current handling at different PWM duty cycles. But the labels seem backwards to me. Wouldn't running the diode at lower duty cycles give it more time to cool down?

Or is there some kind of initial resistance that drops after a moment, kind of like when you have a MOSFET turned on partway, and the lower PWM duty keeps it in this region?

I see that as : the maximum allowable average forward current.
Suppose the delta is 0.1, and the max average is about 0.45 A (a graphical guess).
That means that pulses of 4.5A are allowed at 10% of the time at 20kHz.
Because pulses of 4.5A at 10% of the time make an average of 0.45 A.

That makes perfect sense, because the maximum continueus forward current is about 1.4 A

Take Fig10 for example:

DC - Iave = 1.4
50% - Iave = 1
20% - Iave = 0.63
10% - Iave = 0.45

So based on that:
DC - Ipeak = Iave / 1 = 1.4
50% - Ipeak = Iave / 0.5 = 2
20% - Ipeak = Iave / 0.2 = 3.15
10% - Ipeak = Iave / 0.1 = 4.5

So the peak current allowed gets larger at lower duty cycles.

Wouldn't running the diode at lower duty cycles give it more time to cool down?

Yes but it would also give it longer to heat up and that heat will build up too quickely and get too hot at the junction and damage it.

Caltoa/Tom:
Thanks guys, makes perfect sense!

Grumpy:

Wouldn't running the diode at lower duty cycles give it more time to cool down?

Yes but it would also give it longer to heat up and that heat will build up too quickely and get too hot at the junction and damage it.

I don't see what you're getting at. Unless there's something else going on that I'm unaware of, a 100% duty cycle should give the diode 10x as long to heat up as a 10% duty cycle.