Actually, I think you're misinterpreting the evidence in the photo.

This is a bit of a classic for flow dynamics on Mars. Flow came from the bottom left in a huge "flood" spilling across the valley floor. if it was an aqueous flow there would have been a wide lake. At a narrow point in the valley, a waterfall developed, that would put Niagara or Victoria falls to shame. The waterfall eroded back up the valley, forming an entrenched channel, but the scale of the flow was so great that material spilled over the meander bends (the "wet" looking patches like leaves on a stem). A gas-supported flow would also overspill like this. A massive fluid lava stream could also possibly produce the same patterns

At the upstream end the "lake" drained into the entrenched channel forming a dendritic pattern (tree-like or in this case, stubby fingers). At the downstream end another lake formed.

There are fascinating patterns of collapse and/or boulders close to the shorelines of the two "lakes". Perhaps ice-rich debris rafting down the lakes, or giant rocky lumps from the frothy gassy avalanche that could have poured down the valley in a non-aqueous scenario. Or solidified crusts of basalt from a lava river and lake, now drained away?