关键词: 冲洪积扇, 高精度地形数据, 起伏度, 粗糙度, 自动划分, 地貌演化

Abstract: Range-front alluvial fan deposition in arid and semiarid environments records vast amounts of climatic and tectonic information. Differentiating and characterizing alluvial fan morphology is an important part in Quaternary alluvial fan research. Traditional method such as field observations is a most important part of deciphering and mapping the alluvial fan. Large-scale automatically mapping of alluvial fan stratigraphy before traditional field observations could provide guidance for mapping alluvial fan morphology, thus improving subsequent field work efficiency. In this research, high-resolution topographic data were used to quantify relief and roughness of alluvial fan within the Laohushan. These data suggest that mean surface roughness plotted against the size of the moving window is characterized by an initial increase in surface roughness with increased window size, but it shows no longer increase as a function of windows size. These data also suggest that alluvial fans in this study site smooth out with time until a threshold is crossed where roughness increases at greater wavelength with age as a result of surface runoff and headward tributary incision into the oldest surfaces which suggests the evolution process of alluvial fan.
Researchers usually differentiate alluvial morphology by mapping characteristics of fan surface in the field by describing surface clast size, rock varnish accumulation, and desert pavement development and analysis of aerial photographs or satellite imagery. Recently, the emergence of high-resolution topographic data has renewed interest in the quantitative characterization of alluvial and colluvium landforms. Surface morphology that fan surface initially tends to become smoother with increasing age due to the formation of desert pavement and the degradation of bar-and-swale topography and subsequently, landforms become more dissected due to tectonics and climatic change induced increased erosion and channelization of the surface with time is widely used to distinguish alluvial fan types. Those characteristics would reflect various kinds of morphology metrics extracted from high-resolution topographic data. In the arid and semiarid regions of northwestern China, plenty of alluvial fans are preserved completely for lack of artificial reforming, and there exists sparse surface vegetation. In the meantime, range-front alluvial fan displaced by a number of active faults formed a series of dislocated landforms with different offsets which is a major reference mark in fault activity research. In this research, six map units(Qf₆-Qf₁), youngest to oldest, were observed in the study area by mapping performed by identifying geomorphic features in the field that are spatially discernible using hill-shade and digital orthophoto map. Alluvial fan relief and roughness were computed across multiple observation scales(2m×2m to 100m×100m)based on the topographic parameters of altitude difference and standard deviation of slope, curvature and aspect.
In this research, mean relief keeps increasing with increased window size while mean surface roughness is characterized by a rapid increase over wavelengths of 6~15m, representing the typical length scale of bar-and-swale topography. At longer wavelengths, surface roughness values increase by only minor amounts, suggesting the topographic saturation length is 6~15m for those fan surfaces in which saturation length of standard deviation of curvature is less than 8m. Box and whisker plot of surface roughness averaged over 8m² for each alluvial fan unit in the study area suggests that the pattern of surfaces smoothing out with age and then starting to become rougher again as age increases further beyond Qf₄ or Qf₃ unit. The younger alluvial fan is characterized by prominent bar-and-swale while the older alluvial fan is characterized by tributaries headward incision. Cumulative frequency distributions of relief and surface roughness in Figure 8 are determined in an 8m by 8m moving window for the comparison of six alluvial fan units in the northeast piedmont of Laohushan. From these distributions we know that Qf₆ and Qf₁ reflect the prominent relief which is related to bar-and-swale and tributaries headward incision respectively, while Qf₄ and Qf₃ reflect the moderate relief which is related to subdued topography.
Surface roughness, in addition to facilitating the characterization of individual fan units, lends insight to alluvial landform development. We summarize an alluvial landform evolutionary scheme which evolves four stages depending on characteristics of alluvial fan morphology development and features of relief and roughness. The initial stage in this study site is defined as the active alluvial fan channels with bars of coarse cobbles and boulders and swales consisting of finer-grained pebbles and sand which could be reflected by high mean relief and mean roughness values. As time goes, bar-and-swale topography is still present, but an immature pavement, composed of finer grained clasts, has started to form. In the third stage, the bar-and-swale topography on the fan surface is subdued, yet still observable, with clasts ranging from pebbles to cobbles in size and there exists obvious headward tributary incision. Eventually, tributary channels form from erosion by surface runoff. Headward incision of these tributaries wears down the steep walls of channels that are incised through the stable, planar surface, transforming the oldest alluvial landforms into convex hillslopes, leaving only small remnants of the planar surface intact. Those evolutionary character suggests that alluvial fans in this area smooth out with time, however, relief or roughness would be translated to increase at greater wavelength with age until a threshold is crossed.
This research suggests that relief and roughness calculated from high-resolution topographic data of this study site could reflect alluvial fan morphology development and provide constraint data to differentiate alluvial fan unit.

Key words: alluvial-proluvial fan, high-resolution terrain data, relief, roughness, automatic division, landform evolution