Introduction to Cratering Studies

PSI MGS Web Site: Table of Contents MGS Mission at PSI Introduction to Cratering Studies Published Results Current Research in Progress Study of Mars Landing Sites

What can we learn from crater counts in general? Every time an asteroid or comet hits a planet, it makes an impact crater. These craters accumulate with time. The older a surface, the more impacts it has experienced. If we are clever enough, we humans can use these circular features to interpret the history of the planetary surfaces we see. For example, we can rank surfaces, or geologic units, on any planet in order of age by noting that the more impact craters it has, the older it is. If we can estimate the rate of crater production (for example by using the known rates measured for the moon from Apollo exploration), then we can count up the total number of craters and calculate the actual age of the surface in millions of years.

Dr. Hartmann has published interpretation of Mars crater statistics since the 1960's, and pioneered many aspects of this work. He shared the Nininger Mereorite prize in 1966 for using this technique to count craters in Canada and on the lunar lava plains, and then estimated that the lava plains are, on average, about 3.4 billion years old. This estimate was proven exactly correct when Apollo astronauts brought back samples a few years later -- a development that gave confidence in the technique.

At a more subtle level, Hartmann (1966) coined the term "crater retention age" to emphasize that on a geologically active planet like Mars or Earth, the crater numbers give an estimate not necessarily of the original date of formation of the surface (e.g. a lava flow) but rather of how long features of a given size survive, given the erosion/deposition conditions on that surface. For example, a 1-kilometer crater might take 100,000 years to be obliterated, due to rains, glaciers, sedimentation, etc., wheras a 10-meter crater might last only a few years. The lifetimes at different scales give us an idea of the environmental activity and processes on that planet.

The technique was developed in the 1960s using lunar craters. We have to count not just the total number of craters, but the numbers of craters at different sizes. Figure 1 shows a typical size distribution for crater counts in the lava plains on the moon, which are about 3400 My (million years) old. The dashed line is a fit through those data points. It turns up at small sizes because there are many more small impacts, apparently due to "secondary" fragments blown out of craters on the moon and also blown out of craters on the asteroids themselves. German researcher Gerhardt Neukum and others have shown that this turned-up "secondary branch" is present in the incoming "primary" meteorite flux falling out of space. It has been seen among fragments in the asteroid belt. The solid line is a fit to the very heavily cratered uplands of the moon. This line appears to mark approximate saturation equilibrium of craters on the surface; that is, if you added a new crater at this crater density, the new one would obliterate older craters and on average the line would stay about the same.

Figure 1: Crater size distribution for craters on lunar lava plains. (Age approx. 3400 My).

Figure 2: Crater size distribution for older regions of Earth. Small craters have been lost by erosion, as measured relative to numbers seen on the moon (reference lines).

The lines on Figure 1 are used as reference lines on all our plots of Mars data. Figure 2 shows what happens if we apply these ideas to the Earth. As we all know, Earth has few craters. This is because Earth's surface is so geologically active that most areas are less than 500 million years old -- too young to have many impact craters! Therefore, as shown in Figure 2, the data points for Earth hover at the bottom of the diagram, far below the lunar data points. Also, note that the slope of the curve for the Earth is not as steep as for the moon. It is nearly flat, because (as common sense tells us) small craters are obliterated by geologic processes much faster than large craters. At a crater diameter of 500 km Earth has about as many craters as the lunar lava plains, because they last a long time. But notice that for 250 m diameter craters, the number visible is only around one ten-millionth of that number. Assuming the same influx of impactors on Earth and moon, the smaller craters on Earth have not survived.

An important lesson from this: flattening of the size distribution, relative to what is seen on the moon, can be an indication of some form of geologic processing that erodes or obliterate craters.

Figure 3 shows the background reference lines that we use on a typical Mars plot. The dotted line is the same reference line seen in figures 1 and 2 for lunar lava plains, and the solid line is for saturation equilibrium. In addition are two short heavy lines defined by the U.S. Geological Survey to divide Martian history into three broad periods. NOACHIAN PERIOD: this is the oldest period with the most impact craters, in which there may have been ancient seas, lakes, and/or rivers. HESPERIAN PERIOD: a middle period. Erosive and volcanic activity may have been declining. An open question is the degree to which river channels were still active. AMAZONIAN PERIOD: the most recent period, with the fewest impact craters. Many of the sparsely cratered lava plains of Mars formed during this period.

Figure 3: Basic background reference lines for mars crater count plots.

What do Martian impact craters tell us about Mars?

The goal of crater-count studies on Mars is to characterize relative age relations, make absolute age estimates, and to characterize the erosion regime as far as it can be judged from the loss of small craters.

Figure 4: Crater size distribution for young lava plains on Mars. Shape of distribution at larger sizes matches the shape found on the moon.

Figure 4 shows a typical set of crater counts from the Mariner 9 era (1972), on the young (Amazonian) lava plains of the broad Tharsis region and other lava plain regions of Mars. Large-diameter craters were counted over a large region; smaller diameter craters were counted in a sampling of smaller regions and averaged to characterize the area. The main point is that in this young area, the crater populations (at least at larger diameters above D = 2 km) are a very good match to the general shape, or slope of the size distribution of the lunar lava plains. This means that the same size distribution of impactors is hitting both areas (as expected from the inner solar system's population of asteroids and comets). In recent, Amazonian Martian history, there has not been enough erosion on Mars to obliterate craters larger than 2 km. This is why the two curves are parallel.

Can we see effects of erosion in older Martian regions?

Yes. Figure 5 shows our new results from our counts of craters on one of the first MGS images -- showing the old cratered plains around Nirgal Vallis. Hartmann (1971 and current work in preparation) has used a simple theoretical model of crater obliteration to predict the steady state distribution of craters on Mars. The model assumption is simply that craters act as holes, or potential wells, that preferentially fill with windblown dust (or any other assumed material) faster than the surroundings. The key assumption in the model used here is that the rate of net infill is a constant. (Time varying rates could also be assumed in order to make other predicted curves.) Under this model, craters have a lifetime proportional to their depth. The solid curve shown here is for a net infill rate of about 0.0004 cm/y. This rate was derived using this technique in its early stages, as far back as our early analysis of Mariner 4 data in 1966 (Icarus, 5:565-567). The first suggestion that small craters have been obliterated on Mars due to dust infill was made by Ernst Öpik in 1965 when he wrote about the Mariner 4 pictures. For this reason, we have begun to call this loss, and the flattening of the crater size-distribution, the Öpik Effect.

Figure 5: New crater count data for one of the first MGS images. Crater counts on old lava plains near Nirgal Vallis show losses of small and mid-size craters and fit a steady-state due to net dust deposition in craters of about 0.0004 cm/year.

The high resolution MGS pictures give good evidence in support of this model. They show drifts of windblown dust engulfing many older, smaller craters, and deposits of dunes or smooth dust on the floors of many larger craters. We can confirm that the depositional-obliteration process is actually occurring on Mars.

The interpretation of Figure 5 is that the craters are indeed giving evidence of progressive loss of smaller craters -- an environment different from the moon. Probably the agent is infill by windblown dust, which obliterates smaller craters. The fit to the predicted line suggests net deposition rates of the order 0.0004 cm/y, which may be enough to obliterate the oldest craters up to tens of kilometers across. We will be able to use these techniques to search for variations in obliteration rate, or net deposition rate, in different locations on Mars.