JOURNAL OF GEOPHYSICAL RESEARCH, VOL. 99, NO. B10, PAGES 20,267-20,280, OCTOBER 10, 1994 Kinematics of arc-continent collision in the eastern Central Range of Taiwan inferred from geomorphic analysis James H. Willemin 1 and Peter L.K. Knuepfer Department of Geological Sciences and Environmental Studies, State University of New York at Binghamton Abstract. Activearc-continent collision in Taiwan is anexample of a typical collisional orogen, with a deformed continental margin (Foothills Belt,Hsueshan Range, and Central Range) separated froman accreted island arc(Coastal Range) by a collapsed forearc basin (Longitudinal Valley). The geomorphology of theeastern Central Range reflects spatial and temporal variations in upliftrateand pattern along theorogen during lateQuaternary time. Digitized drainage basin perimeters, stream channels, andthemountain-piedmont junction of a 200-km- long segment of theeastern Central Range provide data for calculating quantitative geomorphic parameters. Statistical analyses of these parameters define fourdomains of distinct landscape character. The northern domain extends northward along thecoast from the northern endof the Longitudinal Valley (LV) and has landscape characteristics suggesting uplift near themountain front. The central two domains areadjacent to theLV and show evidence of uplift concentrated in theinterior of therange, awayfromthefront. The boundary between thetwo central domains occurs near themiddle of theLV. The southern domain extends southward along thecoast from the southern endof theLV and has characteristics suggesting a transient response to increased upliftrate. Analysis of fluvial terrace profiles in theCentral Range adjacent to the southern half of theLV indicates a change in upliftpattern through time,with migration of maximum uplift ratefrom themountain frontto theinterior of therange. This change probably occurred just prior to late Pleistocene time. Introduction Tectonic geomorphology, broadly defined,is the studyof the influence of active tectonism on the landscape [Keller, 1986; Mayer, 1986]. As with otherbranches of geology, tectonic geo- morphology has two distinct modes of inquiry: the forward problem, in this case the study of the effects of a known tectonic perturbation on the landscape [Merritts and Vincent, 1989; Beaumont et al., 1992; Koons, 1990], and the inverse problem, one of inferring the nature and relative activity of tectonicele- ments from observed variations in landscape character [Fonseca, 1988; Wells et al., 1988; Menges, 1990]. In regions where all primary controls of landscape characterexcept tectonismare relatively constant (i.e., geologic structure, lithology, (micro)climate,biota, and human land use), variations in rate and style of tectonism are dominantly responsible for variations in landscape character [Wells et al., 1988], and the inverse problem becomes tractable. Such a situation exists in the eastern Central Range of Taiwan, which has dominantly orogen-parallel structural and lithologic trendsin a relatively uniform climatic setting [Ho, 1982, 1988]. In this study, quantitative geomorphic parameterssuch as mountain front sinuosity [Bull and McFadden, 1977] and stream gradient index [Hack, 1973] are used to definegeomorphic domains (regions of distinct landscape character) along theeastern Central Range. The characteristics of 1Now atDepartment of Geology, St.Lawrence University, Canton, New York. Copyright1994 by the American Geophysical Union. Paper number 94JB00731. 0148-0227/94/94JB-00731 $05.00 these geomorphic domains are then used to infer variations in rate and style of tectonism alongthe front. This study of the tectonic geomorphology of the eastern Cen- tral Range of Taiwanis designed to test two hypotheses: (1) Qua- ternary uplift is concentrated along the mountain front(consistent with uplift alonga westdipping thrust fault beneath the eastern Central Range); and (2) uplift ratesvary alongthe range, possibly asa result of the southward decrease in age of collision between the Asian and Philippine Sea plates. Geomorphic analysis of fluvialsystems (particularly stream long profiles and fluvialterrace profiles) is a valuable tool for locating zones of uplift with respect to a mountain front [Rockwell et al., 1984; Bull and Knuepfer, 1987; Wells et al., 1988]. For example, activefaultsmay produce oversteepened or backwater reaches along a stream [Wells et al., 1988], and uplifttends to generate flights of terraces whoseheight abovethe present channel suggests thedegree of relative uplift [Rockwell et al., 1984; Bull, 1990; Merritts et al., thisissue]. Mountain front analysis [Bull andMcFadden, 1977;Mayer, 1986;Fonseca, 1988;Wells et al., 1988] similarly provides information about variation in relative uplift rates along a mountain range front. Thelandscape along a mountain front is the result of a balance between competing processes: uplift tendsto steepen and straighten a mountain front, while stream erosion tends to cut embayments into the front, and hillslope processes reduce the frontal slope. Geologic and Tectonic Setting The island of Taiwanis the result of active, oblique collision between the Asiancontinental margin and the overriding Luzon arcof the Philippine Seaplate(Figurel a). The collision is com- plex,involving reversal of subduction polarity fromeast dipping 20,267