This is not so much a review of a recent paper as a review of a significant paper. “An intense terminal epoch of widespread fluvial activity on early Mars:1. Valley network incision and associated deposits” by Alan Howard, Jeff Moore, and Ross Irwin is the first of a pair of papers published in 2005 that make the case that instead of a gradual transition from warm and wet to cold and dry on Mars, there was a frenzy of flowing water at around 3.7 billion years ago (the boundary between two geologic epochs called the Noachian and Hesperian).

Howard et al. point out that in the Noachian period (~4.1 to 3.7 billion years ago) there was enough erosion to fill craters 10s of kilometeres across with sediment, and they estimate that on average 100s of meters of erosion happened in the highlands (with variations according to local climate, slope, etc.). Some drainage networks in ancient Noachian terrain are 100s of kilometers long, suggesting that there was enough flowing water in some places to make relatively “mature” networks. You can think of the “maturity” of a drainage network as how long the water has been flowing. If you go outside and dump a bucket of water on a pile of dirt, you’ll end up with some erosion and a little channel carved into the dirt, but it isn’t connected with any other channels – it is “immature”. But if you pour water on a landscape for thousands or millions of years, the channels eventually connect up with each other to form a dense branching network or tributarys – a “mature” system.

But at the same time, it doesn’t look like early Mars was particularly wet by Earth standards. Howard et al. point out that not many Noachian-aged craters have breached rims, which is what you would get if the crater filled with water and overflowed. The authors estimate that the long-term erosion rate in the Noachian period was comparable to the erosion rate in the driest deserts on earth, like the Atacama in Chile or the dry valleys of Antarctica.

The central puzzle that is considered in this paper is that there are some valley networks on Mars that are much less degraded than others. Howard et al. consider these fresher-looking valleys on the border of the Isidis basin, south of the crater Schiaparelli, and in Margaritifer Sinus, and try to come to some conclusion about where the fresher valleys came from.

In the Isidis basin margin, the fresh-looking valleys stretch up to 800 km, and can be multiple kilometers wide at their thickest points. These valleys tend to cut several hundreds of meters into plains of material that appears to be sediment from erosion that occurred in the Noachian era. By carefully studying the topography of the area, the authors found that the valleys are deepest where the local slope is greatest, exactly where you would expect the most erosion to occur. But in a river system that is allowed to flow for a long time, this effect tends to smooth out the river’s profile (if you get the most erosion where it is steepest, pretty soon those areas become less steep!).

Over in Evros Valles south of Schiaparelli, the authors point out a drainage system that consists of a well-defined main valley with lots of faint tributaries. Again, the deepest parts of the valley are also the steepest parts. The same is true for the Parana  Valles in Margaritifer Sinus. In this area there is also an interesting example of erosion that occurred in multiple stages: early erosion created a gently sloping surface composed of fans or “pediments” (ramps of eroded bedrock), but then a later round of erosion carved channels into that, which then widened to create a new fan surface which was incised again! The details of this are not as important as the fact that clearly there was more than one round of erosion.

Howard et al. also take a look at alluvial fans and possible deltas on Mars, and find that large alluvial fans are found in craters >50 km in diameter. Crater counts on the fans give them an age that falls near the boundary between the Noachian and Hesperian eras, and the fans don’t have drainage networks on them (something that is common on Earth when it rains after a fan has been formed). In terms of deltas, of course Eberswalde (a former candidate landing site for Curiosity) is the best example but the authors point out a couple of other possible deltas too. They note that the deltas often show inversion of relief which occurs when river beds are more resistant to erosion than their surroundings, and so they end up standing above the surrounding plains when erosion occurs. They also note that the volume of the Eberswalde delta is not far off from the volume of the valleys that feed into it, suggesting that they formed at the same time.

At this point, the authors take a step back and start looking for common characteristics of the “fresh” valleys. They note that the main valleys tend to carve sharply into plains in the highlands, and tend to have steep side walls. Most of the valley networks mapped from Viking-era data behave like this, and the authors interpret them to be valleys carving the sediment that was generated in the Noachian era when the highlands eroded.

The fact that there are some deltas of an age comparable with the Noachian-Hesperian boundary suggests that there was enough water flowing for long enough to create lakes, and the large alluvial fans are also about this age. The authors say that the fans “imply appreciable,  probably repeated flows extending  over a time interval of many hundreds or thousands of years”. Also, since the deltas and fans are not incised by later channels, the authors believe that the flows that formed them shut off pretty abruptly.

Based on the width of some channels, the authors estimate the amount of water that was flowing, and it is pretty large, suggesting that rain or snow was involved. It’s harder to get sudden large amounts of water from groundwater sources – those tend to be more of a steady trickle. And with groundwater there is the problem of refilling the aquifer, which would require precipitation anyway. Plus many of the valleys occur at high elevations where you wouldn’t expect groundwater to be discharged. Of course, the authors point out that it’s impossible to know how often the discharge occurred or how long it lasted, but making some assumptions based on rates on Earth, they estimate that the delta in Eberswalde would take thousands to 100s of thousands of years to form. If the water was flowing non-stop, it could form in tens of years.

So what explains the fact that these valleys carve into sediment deposited from erosion in the Noachian, why didn’t the sediment just keep piling up? Basically you can switch from accumulating sediment to eroding it by either increasing the flow of water or decreasing the supply of new sediment. The authors list three scenarios:

1. To increase the flow of water, you could have the climate warm up (possibly related to volcanic eruptions, changes in the planet’s tilt, or big impacts), driving an increase in rain and snow.
2. You can also get big discharges by switching from mostly rain to mostly snow: snowmelt tends to lead to big pulses of water flowing through the system, but relatively low amounds of new sediment, so that can cause incision of new channels.
3. You can also do it by cutting into soft rocks below a resistant cap, such as a duricrust, which is a hard layer that forms at the surface of a soil.

Of course, you can have combinations of these scenarios, and there are lots of unresolved questions, but the bottom line is, there is good evidence that something changed toward the end of the Noachian period causing new valleys to be carved into the sediment accumulated earlier in the Noachian, and leading to the formation of alluvial fans and deltas.

Howard, A., Moore, J., & Irwin, R. (2005). An intense terminal epoch of widespread fluvial activity on early Mars: 1. Valley network incision and associated deposits Journal of Geophysical Research, 110 (E12) DOI: 10.1029/2005JE002459