This was one of the points I raised in my "ephemeral atmospheres" work
Back in about 2001, I showed that a variety of landforms on early "pluvial" Mars did show good signs of fluvial erosion, but they spanned a period of Mars history of several hundred million years, yet each one was only weathered to the same extent of a few million years on Earth - this immediately implied that Early Mars was "warm and wet" only about 1% of the time. From this came the concept of impact-driven ephemeral atmosphere which I think is gradually gaining strength as a useful concept for early Mars...