Laserfiche WebLink
<br />structures and (2) Ihe effectiveness of the check dams in mitigating the effects of arroyo erosion <br />and river flow stage. Thermal infrared is very sensitive to subtle differences (or changes) in <br />density and grain size (Hussein. 1982; Johnson et aI., 1998), Thus. minor disruptions of the <br />surface will change Ihe surface's densilY and/or grain size, which may be detected in TIR imagery <br />(Johnson et aI., 1998). Thermal-infrared images can also distinguish a degraded, buried ruin from <br />its surrounding alluvium. as long as the ruin and the alluvium have differenl compositions or <br />densities (Berlin et al.. 1977; Nash, 1985; Berlin et aI., 1990). Thus. we invesligated the use of <br />TIR data, as well as visible and near-infrared (NIR) imagery of different spatial resolutions, for <br />detection and mapping archaeological struclures at Unkar Delta (Davis, 2002b). Our results <br />showed that spatial resolution was Ihe key factor in mapping these structures for dala acquired <br />during Ihe middle ofa day. We could nol unambiguously identify the slruclures al Unkar Delta <br />using the daytime TIR dala al I-m resolution or using any of the ten reflectance wavelength band <br />images that were acquired at I-m resolution with the '1'1 R data (Figures I and 2). The ambient air <br />temperature in the canyon during our mid,day July data acquisition was close to 38 oc. T1R <br />sensors require liquid nilrogen cooling systems to maintain the sensor near absolute zero degrees, <br />but the cooling system could not compensate for Ihis high air temperalure. Thus, the sensitivity <br />of the TIR data was only 0.3 degrees, whereas a sensilivity of 0,1 degrees is required to detect <br />thennal anomalies associated with geologic materials (such as the archaeological slructure <br />relative to its alluvial surroundings; Nash, 1985). Panchromatic (black,and-white) imagery at 18, <br />em resolution was also found to be much less useful in uniquely identifying the structures than <br />CIR (green, red, and NIR composite) imagery at I I ,em resolution, which is attributed to the lack <br />of conlrasl between the alluvial and archaeological materials in black-and-white imagery. We <br />found that the most unambiguous detection oflhese structures requires CIR or natural-color <br />imagery with resolutions near II cm. The incipient stage of arroyo development may manifest <br />ilself more as subtle changes in surface materials due to recent exposure or transport Ihan as <br />changes in surface topography. However, the CI R imagery do not provide good discrimination of <br />alluvial geologic materials that may indicate mass Illovemenl and arroyo development. In order <br />to delecl subtle changes in alluvial surface materials, it would be better to acquire multispectral <br />dala that include at least one short-wave infrared wavelength band. where more distinctive energy <br />absorptions occur for geologic minerals. Such systems cannot obtain spatial resolutions better <br />than 0.5 m. Although this resolution will not detect morphometric changes in structures, the data <br />could better map changes in surface materials that may indicate mass movement. <br /> <br />For the detection of morphometric (dimensional) changes in archaeological structures and <br />in the arroyos that affect these slructures, even higher resolulion imagery (2-3 em or I: I ,600-scale <br />photography) is required (MacFarlane et aI., 2002; Petersen el aI., 2002), especially for <br />monitoring the very fine-scale (centimeter) changes in arroyos Ihat could ir.di-cate potential <br />impacts on such struclures. Photogrammetric analyses of such extremely high-resolution image <br />data, which were acquired within the remole-sensing initiative, produced an elevation accuracy of <br />6-10 em; the average vertical error at the 95% confidence level (for normally distributed errors) <br />of 18 cm. These data could resolve sub-meter-scale changes. but the vertical error exceeds Ihe <br />change-detection threshold for Ihe smaller, cm-scale gullies (MacFarlane el aI., 2002). Field <br />studies at four selected sites by Petersen et al. (2002) indicated that vegetation cover. soil <br />permeability, and soil shear slrength inversely related to gully erosion activity, These observed <br />trends, plus the factlhat gully head local ions can be predicted by slope and drainage area <br />relations derived from the phologrammelric data, suggeslthat up-calchment control of gully <br />initiation and knick-point refreat are more critical factors in structural degradation than is base- <br />level changes related to Colorado River stage or Glen Canyon Dam operations (Petersen et al.. <br />2002). The stereo-image data that were collected did however provide a catchment OEM that <br />allows geopmorphic modeling of potential impacts of rainfall on arroyo development and, in turn, <br />of potential impacls of arroyo development on down-slope archaeological Slructures. One of Ihe <br /> <br />8 <br />