Pro gradu -tutkielma Geologia Kallioperägeologia, taloudellinen geologia ja geodynamiikka

Global variations in erosion of young orogens: Swath profile comparison of climatic, erosional and topographic metrics to long-term exhumation rates

Niclas Blomqvist

2016

Ohjaaja(t): David Whipp

HELSINGIN YLIOPISTO MATEMAATTIS-LUONNONTIETEELLINEN TIEDEKUNTA GEOTIETEIDEN JA MAANTIETEEN LAITOS GEOLOGIA

PL 64 (Gustaf Hällströmin katu 2) 00014 Helsingin yliopisto Tiedekunta/Osasto Fakultet/Sektion – Faculty Laitos/Institution– Department Faculty of Science Department of Geosciences and Geography Tekijä/Författare – Author Niclas Blomqvist Työn nimi / Arbetets titel – Title Global variations in erosion of young orogens: Swath profile comparison of climatic, erosional and topographic metrics to long-term exhumation rates Oppiaine /Läroämne – Subject Geology Työn laji/Arbetets art – Level Aika/Datum – Month and year Sivumäärä/ Sidoantal – Number of pages

Masters thesis 02/2016 185 Tiivistelmä/Referat – Abstract The topography of the Earth’s surface is the result of the interaction of tectonics, erosion and climate. Thus, topography should contain a record of these processes that can be extracted by topographic analysis. The question considered in this study is whether the spatial variations in erosion that have sculpted the modern topography are representative of the long-term erosion rates in mountainous regions. We compare long-term erosion rates derived from low-temperature thermochronometry to erosional proxies calculated from topographic and climatic data analysis. The study has been performed on a global scale including six orogens: The Himalaya, Andes, Taiwan, Olympic Mountains, Southern Alps in New Zealand and European Alps. The data was analyzed using a new swath profile analysis tool for ArcGIS called ArcSwath to determine the correlations between the long-term erosion rates and modern elevations, slope angles, relief in 2.5-km- and 5-km-diameter circles, erosion potential, normalized channel steepness index ksn, and annual rainfall. ArcSwath uses a Python script that has been incorporated into an ArcMap 10.2 add-in tool, extracting swath profiles in about ten seconds compared to earlier workflows that could take more than an hour. Swath profile analysis is a relatively common method for geomorphological research. A swath profile is a rectangular extraction of a digital model to a cross-section, where statistical parameters (minimum, mean and maximum) are presented along the profile length. In previous studies swath profiles have been used to identify relationships between topography and major structures, to compare various climatic, erosional and topographic data across a given orogen and along strike, and to recognize fluvially and glacially eroded forms.

An unambiguous correlation between the topographic or climatic metrics and long-term erosion rates was not found. Fitting of linear regression lines to the topographic/climatic metric data and the long-term erosion rates shows that 86 of 288 plots (30%) have “good” R2 values (> 0.35) and 135 of 288 (47%) have an “acceptable” R2 value (> 0.2). The “acceptable” and “good” values have been selected on the basis of visual fit to the regression line. The majority of the plots with a “good” correlation value have positive correlations, while 11/86 plots have negative slopes for the regression lines. Interestingly, two topographic profile shapes were clear in swath profiles: Concave-up (e.g., the central-western Himalaya and the northern Bolivian Andes) and concave-down or straight (e.g., the eastern Himalayas and the southern Bolivian Andes). On the orogen scale, the concave-up shape is often related to relatively high precipitation and erosion rates on the slopes of steep topography. The concave-down/straight profiles seem to occur in association of low rainfall and/or erosion rates. Though we cannot say with confidence, the lack of a clear correlation between long-term erosion rates and climate or topography may be due to the difference in their respective timescales as climate can vary over shorter timescales than 10 -10 years. In that case, variations between fluvial and glacial erosion may have overprinted the erosional effects of one another. Avainsanat – Nyckelord – Keywords Swath profile analysis, erosion, climate, tectonics, thermochronometry Säilytyspaikka – Förvaringställe – Where deposited University of Helsinki Digital Theses Muita tietoja – Övriga uppgifter – Additional information Tiedekunta/Osasto Fakultet/Sektion – Faculty Laitos/Institution– Department Matemaattis-luonnontieteellinen tiedekunta Geotieteiden ja maantieteen laitos Tekijä/Författare – Author Niclas Blomqvist Työn nimi / Arbetets titel – Title Global variations in erosion of young orogens: Swath profile comparison of climatic, erosional and topographic metrics to long-term exhumation rates Oppiaine /Läroämne – Subject Geologia Työn laji/Arbetets art – Level Aika/Datum – Month and year Sivumäärä/ Sidoantal – Number of pages

Pro gradu -tutkielma 02/2016 185 Tiivistelmä/Referat – Abstract Maapallon pinnanmuotoihin vaikuttavat tektoniikka, eroosio ja ilmasto. Täten topografian pitäisi säilyttää merkkejä näistä prosesseista, jotka voidaan havaita myös topografisen analyysin avulla. Tämän tutkimuksen ongelmana on, edustaako nykypäivän pinnanmuodotkin muokannut paikallisesti vaihteleva eroosio pitkän aikavälin eroosiota vuoristoalueilla. Vertailemme pitkän aikavälin eroosionopeuksia, jotka ovat peräisin alhaisen lämpötilan termokronometriasta, erosionaalisiin arvoihin, joita on laskettu topografisten ja ilmastollisten aineistojen analyyseista. Tutkimus on tehty maailmanlaajuisessa mittakaavassa valitsemalla kuusi orogeenia: Himalaja, Andit, Taiwan, Olympicin vuoristo Luoteis- Yhdysvalloissa, Eteläiset Alpit Uudessa Seelannissa ja Euroopan Alpit. Tutkimus suoritettiin uudella swath profile -analyysityökalulla (ArcSwath), joka toimii ArcGIS -ohjelmassa, ja jonka avulla määritimme riippuvuussuhteita pitkän aikavälin eroosion ja nykypäivän korkeusaseman, rinteen jyrkkyyden, reliefin 2.5 km- ja 5 km halkaisijaltaan olevan ympyrän sisällä, eroosiopotentiaalin, normalisoidun uomanjyrkkyys indeksin (normalized channel steepness, ksn), ja vuotuisen sademäärän välillä. ArcSwath on kirjoitettu Python-koodiin, joka on sisällytetty ArcMap 10.2 add-in -työkaluun. Työkalu eristää swath-profiileja noin kymmenessä sekunnissa, kun aikaisemmin yhden swath-profiilin tekeminen on saattanut kestää jopa tunnin. Swath profile -analyysi on suhteellisen yleinen menetelmä geomorfologiselle tutkimukselle. Swath profiili on suorakulmainen eristys digitaalisesta mallista poikkileikkaukseen, jossa statistiset parametrit (minimi, keskiarvo ja maksimi) esitetään pitkin profiilin pituutta. Aikaisemmissa tutkimuksissa swath profiileja on käytetty tunnistamaan suhteita pinnanmuotojen ja merkittävien rakenteiden välillä, vertaamaan eri ilmastollisia, eroosionaalisia ja topografisia tietoja niin orogeenin kulkua vasten kohtisuoraan kuin myös kulkua pitkin, ja tunnistamaan fluviaalisesti ja glasiaalisesti erodoituneita muodostumia.

Yksiselitteistä riippuvuussuhdetta topografisten tai ilmastollisten mittarien ja pitkän aikavälin eroosionopeuksien välillä ei ollut. Lineaaristen regressiosuorien asettaminen topografisten/ ilmastollisten aineistojen ja pitkän aikavälin eroosionopeuksien välille osoittaa, että 86/288 (30%) diagrammista on ”hyvä” R2 arvo (>0.35) ja 135/288 (47%) on ”hyväksyttävä” R2 arvo (>0.2). ”Hyväksyttävä” ja ”hyvä” arvot on valittu pisteiden visuaalisen sopivuuden perusteella regressiosuoraan. Suurin osa diagrammeista, joilla on ”hyvä” riippuvuusarvo, ovat positiivisia riippuvuussuhteiltaan, kun 11/86 diagrammilla on negatiivinen regressiosuoran kulmakerroin. Topografisille profiileille tunnistettiin kaksi erilaista muotoa: Kovera (esim. Keski- ja Länsi-Himalaja ja Bolivian Andien pohjoisosa) ja kupera/suora (esim. Itä-Himalaja ja Bolivian Andien eteläosa). Orogeenin mittakaavassa kovera muoto liittyi usein verraten korkeaan sademäärään ja eroosionopeuteen jyrkän topografian rinteillä. Kupera/suora muoto esiintyivät usein alhaisen sademäärän ja eroosionopeuden yhteydessä. Selkeän riippuvuussuhteen puuttuminen pitkän aikavälin eroosionopeuksien ja ilmaston tai pinnanmuotojen välillä voi mahdollisesti johtua keskinäisestä aikaskaala erosta, koska ilmasto muuttuu lyhyemmässä skaalassa kuin 10 -10 vuotta. Tässä tapauksessa fluviaalinen ja glasiaalinen eroosio ovat voineet vuorotella ja näin poistaneet toistensa erosionaaliset jäljet. Avainsanat – Nyckelord – Keywords Swath profile, eroosio, ilmasto, tektoniikka, termokronometria Säilytyspaikka – Förvaringställe – Where deposited Helsingin yliopiston digitaalinen opinnäytearkisto Muita tietoja – Övriga uppgifter – Additional information Contents 1. INTRODUCTION ...... 3-4 2. GEOLOGIC SETTINGS ...... 4-24 2.1. The orogens ...... 4-5 2.2. Himalaya and Tibet ...... 5-9 2.3. Andes ...... 9-12 2.4. Taiwan ...... 12-15 2.5. Olympic Mountains ...... 15-18 2.6. The Southern Alps of New Zealand ...... 18-20 2.7. The Alps ...... 20-23 2.8. Low-temperature thermochronometry ...... 23-24 3. MATERIALS ...... 24-25 4. METHODS ...... 25-34 4.1. The SRTM and rainfall raw data processing methods ...... 26-28 4.2. How to use the ArcSwath tool? ...... 28-30 4.3. Correlation plots ...... 31 4.4. The theory of normalized channel steepness and the settings . . . . . 32-34 used with Stream Profiler tool 5. RESULTS ...... 34-60 5.1. Overall ...... 34 5.2. Himalaya ...... 35-41 5.2.1. Marsiyangdi and Bhutan (Mb swaths) ...... 35-39 5.2.2. Sutlej ...... 39-41 5.3. Central Andes (Bolivia) ...... 42-45 5.4. Taiwan ...... 46-48 5.5. Olympic Mountains ...... 49-51 5.6. Southern Alps of New Zealand (NZ) ...... 52-55 5.7. The Alps ...... 56-60 6. DISCUSSION ...... 60-66 6.1. Global variations in erosion of young orogens ...... 60-63 6.2. Swath profile analysis implications for tectonics ...... 63-64 6.3. Recognizing glacial erosion ...... 65-66 7. CONCLUSIONS ...... 66-67 8. REFERENCES ...... 68-74 9. APPENDIX A – Complete set of swath profile and correlation plots for all orogens 75-181 3

1. INTRODUCTION

On the global scale, the location and type of plate boundaries determine whether mountain ranges exist as tectonics is responsible for rock uplift. The evolution of an individual mountain range, however, is also controlled by other factors such as erosion, climate and the deep structure of an orogen (Montgomery et al. 2001). Erosion alone does not uplift the Earth’s surface, but it surely influences the size, shape and evolution of active orogens (Whipple 2004). Furthermore, the role of erosion is critical in linking the dynamic coupling between climate and tectonic activity that was proposed in the late 1980s (Beaumont et al. 1992, Koons 1989 , Willett 1999). These early numerical models showed that mass removal by erosion from a certain part of an orogen could cause faulting and in that sense affect the tectonic evolution of an orogen. Thus, one of the most provocative hypotheses in the last 25 years of geoscience is the idea that the distribution of rainfall on the Earth’s surface can influence tectonic activity in active mountain systems. Although the connection between climate, erosion and tectonics has been predicted in numerical models, it has been difficult to verify the connection in field studies. If it is assumed that climate affects the geomorphology and kinematics of an active orogen, there should exist observable changes in deformation, erosion rates, uplift and landscape patterns where climate changes occur (Grujic et al. 2006). That is to say, the topography should preserve a record of the interactions between climate, tectonics and erosion. The study of this coupled system of climate, erosion and tectonics is, however, complicated by uncertainties in the temporal and spatial evolution of an orogen in response to climatic gradients and the variable impact of fluvial and glacial erosion, for example (Hoth et al. 2006, Roe et al. 2006, Willett et al. 1993, Zeitler et al. 2001).

In this study long-term erosion rates are compared to modern topography and rainfall patterns on a global scale to determine correlations between various modern topographic or climatic metrics and long-term rates of erosion in active orogens. Previous studies have been limited as the scope has been on a scale of an individual orogen. Therefore, six orogens have been selected: The Himalaya, Andes, Taiwan, Olympic Mountains, Southern Alps in New Zealand and the European Alps. These are all active orogens and they differ in various aspects (e.g., size, climate, bedrock lithology, tectonic setting, etc.). The long-term erosion rates are derived from low- 4 temperature thermochronology integrating tectonic and erosional activity over 10 -10 years, whereupon many geological processes or feedbacks operate (Reiners and Brandon 2006). The long-term erosion rates, topographic metrics and climate data are compared using swath profile analysis, which can be helpful in identifying relationship between topography and major structures (Wang et al. 2009), and in comparing various climatic, erosional and topographic data across a given orogen and along strike. The goal is ultimately to relate fluvial or glacial topographic forms to long-term erosion rates and modern rainfall patterns. Swath profiles have been produced using a new tool for GIS analysis (ArcSwath) to make creation of swath profiles easier and more efficient. One goal of this work is to identify significant correlations between the topographic, climatic and erosional variables and link local observations to global-scale patterns.

2. GEOLOGIC SETTINGS

2.1. The orogens

In this chapter I provide the geological and climatic background of the orogens that are the focus of this study, and present the theory of low-temperature thermochronometry, which is the basis for the long-term erosion rates used in the swath profile analysis. As mentioned in the previous chapter, this study considers six young active orogens: The Himalaya, Andes, Taiwan, Olympic Mountains, Southern Alps of New Zealand and the European Alps (Figure 2.1.1). These are all active orogens, but span a wide range in size, altitude, topography, latitude, (longitude), tectonic settings, lithology, climate, continentality, erosion rates and initiation age. In addition, glaciation and erosion rates have not been constant in the orogens. All the previous features combine to create unique and distinctive orogens. For example, according to Barry (1994) mountain climates are governed by four major factors: continentality, latitude, altitude and topography. The main idea for the study is to compare modern topography and climate to long-term erosion rates in swath profiles and correlation plots first on a local scale and then make interpretations for global scale patterns. 5

Figure 2.1.1. Locations of the study orogens indicated by yellow stars. Base map created by following sources: National Geographic, Esri, DeLorme, HERE,UNEP-WCMC, USGS, NASA, ESA iPC, NRCAN, METI, GEBCO and NOAA.

2.2. Himalaya and Tibet

The Himalayan-Tibetan orogen is, in addition to being the largest, also Earth’s most distinctive collisional system. Mountains in the Himalaya reach the highest elevations on the Earth. The Himalayan-Tibetan orogen initiated at circa 50 Ma as a result from the collision between Eurasian and Indian plates (Hodges 2000). The main focus of the research is in the south at the highest topographic section of the Tibetan Plateau, the Himalaya, which is considered as an arc of mountains expanding 2500 km between two distinct peaks: Namche Barwa (7782 m) in the east and Nanga Parbat (8125 m) in the west (Hodges 2000).

It is well known that Himalayan orogeny is the result of collision between Eurasia and India and has evolved to its modern form in the ~50 Ma following collision. Prior to collision, the southern tip of Eurasia was characterized by a continental arc that evolved as the Neo-Tethys oceanic crust subducted northward (Dewey and Bird 1970, Tapponnier et al. 1981). The Transhimalayan zone in the north represents the volcanic 6 and plutonic rocks of this continental arc, their metamorphosed Precambrian-Mesozoic country rocks. Some Cretaceous-Tertiary forearc basin sequences are also found in the Transhimalayan zone (Burg et al. 1983, Searle 1991). Between the Transhimalayan and Tibetan zone is Indus-Tsangpo Suture Zone which represents Neo-Tethyan Ocean basin and its northern and southern continental margins. The Tibetan zone (Tethyan zone) to the south has preserved a nearly complete stratigraphic record of the northern continental margin of India (Figure 2.2.1; Gaetani and Garzanti 1991).

Figure 2.2.1. Tectonic units and major faults in the Himalayas. Modified after Searle (2007).

The South Tibetan Fault System separates the Tibetan zone from the Greater (High) Himalayan zone with normal faults that are considered to have been active since the Miocene (Figure 2.2.1; Hodges 2000). The Greater Himalayan zone is a continuous belt of high-grade metasedimentary and meta-igneous rocks and associated leucogranites. The Main Central thrust system, situated at the boundary of Greater Himalayan and Lesser Himalayan zones, is the oldest (early Miocene) of the major south-vergent thrust faults. The Lesser Himalayan zone in its footwall consists mainly of lower-greenschist to lower-amphibole-facies clastic metasedimentary units that form a complex system of fold-and-thrust nappes and duplex structures at the foothills of the Himalaya. Deposition of this unit is considered to have happened on the north Indian passive margin (Gansser 1964). The second major thrust system, the Main Boundary Thrust 7

System, separates Lesser Himalayan units from the Subhimalayan Zone. Initiation ages of late Miocene–Pliocene have been considered for this north dipping fault system (Hodges 2000), but the onset of faulting is not well constrained. The Subhimalayan zone comprises sequences of clastic sedimentary rocks between the Lesser Himalayan Zone and the present-day thrust front. This thrust, the Main Frontal thrust system, is considered to have become active during the Pliocene or Holocene (Hodges 2000).

The climate in the area of the Himalayan-Tibetan collisional system is highly variable because of the monsoonal winds and rainfall, and the strong rain shadow effect created by the high mountains. The central Himalayan river systems are dominated by rivers that flow directly off the southern plateau. In the western and eastern sectors of the Himalaya, the average relief is not as large as in the central sector and the river networks are also more complex. Monsoon rainfall is heaviest in the eastern Himalaya before losing energy and decreasing while traversing to the west along the mountain front (Hodges 2000). The climate of the Himalayan foreland varies from wet tropical in the east to semiarid conditions in the far west: In the eastern part of the Himalayas, rainfall is distributed across a wider band in the north-south direction, with up to 8000 mm/yr of precipitation while the maximum annual rainfall is around 6500 mm in the west. The eastern Himalaya also do not create a complete rain shadow, but in the central sector the peaks are high enough to block the precipitation effectively. Thus, eastern Tibet experiences abundant precipitation, and on the other hand the central and western Tibet can be defined as high-altitude deserts (Hodges 2000). The Tibetan Plateau central regions are drained internally with low relief (Fielding et al. 1994).

Erosion is focused on the edges of the Tibetan-Himalayan plateau because of the orographic precipitation and efficient river incision that is also centered at the borders of the plateau, while the interior parts of the orogen are protected from rapid erosion in spite of their high elevation (Fielding 1996). Erosion of the broad plateau has been unable to match the uplift in Tibet, in contrast to narrow mountain belts. According to Fielding (1996), the Himalaya have experienced more than 25 km of denudation since the Miocene, whereas central Tibet does not seem to have experienced major erosion since that time. Comparing the flux of suspended load in Ganga and Brahmaputra rivers, Galy and France-Lanord (2001) found erosion rates of 2.9 mm/yr for the eastern Himalaya and 2.1 mm/yr in the western Himalaya. Long term erosion rates derived 8 from bedrock thermochronologic ages are approximately the same, with the difference of higher erosion rates in the western Himalaya (Herman et al. 2013).

The Himalaya-Tibetan plateau hosts the largest area of glaciers outside the polar regions (Owen et al. 2002). The distribution and style of glaciation in central Asia is strongly affected by the SW Indian Monsoon, the mid-latitude westerlies and the El Niño Southern Oscillation (Owen et al. 1998). In addition, the rapid uplift of the Himalaya during late Tertiary and Quaternary has a major role in controlling glaciation (Fort 1996). The Last Glacial Maximum (LGM) has been dated at circa 18 ka (Owen et al. 2002). The interpretation of glacial extent is often confused as mass-wasting-derived sediments and glacial landforms are very similar (Owen et al. 2002). Owen and Benn (2005) have summarized the modern and LGM equilibrium-line altitudes (ELAs) in different locations along the Himalaya (Table 2.2.2): In Garhwal (the western Himalaya) the ELA at LGM was at 4300 m elevation and the modern ELA is 4500- 5400 m above sea level (Barnard et al. 2004, Sharma and Owen 1996). In Langtang (the central Himalaya) the LGM and modern ELAs are at 5920 m and 5320 m (Fort 1995, Shiraiwa 1993, Shiraiwa and Watanabe 1991). In Kanchenjunga (the eastern Himalaya) the LGM and modern ELAs are at 4200 -5200 m and 5000 -6000 m (Asahi and Watanabe 2000, Tsukamoto et al. 2002). Also, the lowest-elevation terminal moraines occur at 2300 m in Garwhal, at 2600 m in Langtang and at 2700 m in Kanchenjunga. There is also a seasonal contrast in precipitation and snow accumulation along the Himalaya; in the east most of the snowfall comes during the summer and melts immediately whereas the dominant input of snow into the glaciers occurs in the winter in the western Himalaya and Karakoram. In the west, the accumulation zone is also much broader creating larger glaciers (Owen et al. 1998). Of these, two main types of glaciers exist: continental and maritime. In continental glaciers, the mean snowline air temperatures are between -16°C and -15°C and the annual precipitation is between 200 and 1000 mm. Most of these glaciers are cold-based and ice flow velocities are low. Such glaciers are mostly situated on the Tibetan Plateau and on the northern slopes of the Himalaya. Maritime glaciers, on the other hand, occur on the monsoon-influenced slopes of the Himalaya where precipitation surpasses 2500 mm/yr. Mean annual temperatures at the snowline are higher than in continental glaciers (-2°C to -4°C), the ice is warm based and ice velocities are comparatively high (Owen et al. 1998). 9

Table 2.2.2. LGM and modern ELAs for the locations analyzed in Himalayan swath profiles. Geologic evidence and methods used to calculate ELAs are also given. Modified after (Owen and Benn 2005). AAR = the accumulation area ratio to the total glacier area and THAR = the toe-to-headwall altitude ratio is the ratio between the maximum and minimum altitude of a glacier. LGM Langtang Kanchenjunga Garhwal Himalaya snowline locality Swath West Mb East Mb Sutlej analysis area Country Nepal Nepal India Lat °N 28 27 31 Lon °E 85 88 79 Glacier type Valley Valley Valley ELA at LGM 5920 4200–5200 4300 (m) LGM method AAR = 0.6 Not known AAR = 0.6 and 1.3, THAR = 0.4 and 0.5 Modern ELA 5320 5000–6000 4500–5400 (m) Modern ELA AAR = 0.4, changes in surface Not described AAR = 0.6 and method contours, highest lateral moraines 1.3,THAR = 0.4 and 0.5 Modern ELA 1980s 1992 1994 date References Shiraiwa and Watanabe 1991, Asahi and Watanabe Sharma and Owen Shiraiwa 1993, Fort 1995 2000, Tsukamoto et al. 1996, Barnard et al. 2002 2004

2.3. Andes

Other than the Himalayan-Tibetan orogen, the Andes are the largest tectonically active orogen. Unlike the Himalayan collision zone, the Andes extend for a distance much greater than their width, with notable variations in climate, topography and crustal structure (Dewey and Lamb 1992). The Andean mountain building processes are believed to have started at about 70 Ma (McQuarrie et al. 2005), though subduction beneath South America has been on-going since the late Jurassic (e.g., Maloney et al. 2013). The main focus of this study is on the central Andes in Bolivia, which is defined by a region at approximately 14 -28°S in western South America (Allmendinger et al. 10

1997). In this region, the normal, east-dipping (~30°) subduction of the Nazca plate beneath the central Andean Plateau is accommodated between zones of flat slab subduction. Over the Cenozoic, normal-to-oblique subduction has produced a notable magnitude and latitudinal variation in shortening (~530-150 km) that has bent the Bolivian part of the orogen (i.e., the Bolivian orocline) and resulted in the uplift of the Andean Plateau (McQuarrie 2002a). According to de Silva (1989), Neogene magmatism was distributed throughout the plateau flanks but from Pliocene to recent the magmatism is present only at the western flank and does not exist in the flat slab zones in the north and south.

The central Andes include of several tectonomorphic zones (Figure 2.3.1) that have developed over the past ~50 Ma. From the west, the Precordillera and Western Cordillera form the western flank of the Andean Plateau, which is a faulted monocline of west-dipping Neogene sediments (Hoke et al. 2007, Muñoz and Charrier 1996, Victor et al. 2004). The Western Cordillera includes the modern volcanic arc and represents the western drainage divide of the Altiplano-Puna basin (Barnes and Ehlers 2009). The Altiplano region is a sedimentary basin between the uplifting mountain ranges of the Western and Eastern Cordilleras (Lamb and Hoke 1997). McQuarrie (2002b) describes the eastern Andean Plateau margin as a thick-to-thin-skinned central Andean fold-thrust belt (Altiplano, Eastern Cordillera, Interandean zone and Subandes; Figure 2.3.1). The Eastern Cordillera is also the eastern Altiplano-Puna drainage divide and the highest relief zone with Paleozoic sediments and Precambrian metamorphic rocks with overlying Cenozoic volcanics. In the Interandean zone and the Subandes to the east and north of 23°S, the topography steps steeply downwards, as eastward the structural depth shallows to expose Devonian and Carboniferous rocks through to Mesozoic and Cenozoic rocks (Kley 1996, McQuarrie 2002b). The eastward- propagating thrusting in the fold-thrust belt waned in Eastern Cordillera from 40 to 15 Ma. The shortening started to transfer from the Eastern Cordillera and Interandean zone to the Subandean zone at circa 20 Ma (McQuarrie et al. 2005). 11

Figure 2.3.1. Tectonic units and major structures in the Central Andes. Location of the study area is indicated in top right corner. Modified after Charrier et al. (2013)

The climate is highly variable in the central Andes. At about 18°S the precipitation and climatic conditions are divided; in the south, subtropical conditions prevail with small mean rainfall amounts (2000 to less than 1000 mm/yr) and in the north subequatorial wet conditions result in heavier rainfall with a larger variation as Hadley-cell-driven winds bring precipitation to the mountains (4000 to less than 1000 mm/yr; Barnes et al. 2012). The El Niño Southern Oscillation has also an important effect on the annual precipitation north of 18°S (Bookhagen and Strecker 2010).

Modern glaciation of the central Andes is not significant in comparison to many of the orogens studied in this work. Glaciers north of 18° 23’ S in Bolivia cover more than 560 km2. The Bolivian Andes become increasingly arid to the south, causing the peaks that exceed the altitude of the 0°C isotherm in southern Bolivia to remain unglaciated for lack of precipitation (Jordan 1998). The fact that glacier coverage has previously been 12 significantly larger is indicated by the presence of moraines and other glacial deposits in the regions of the Bolivian Andes that currently have either no glaciers or glaciers terminating far up-valley (Smith et al. 2008). According to Klein et al. (1999), the modern snowline in Bolivia rises steeply southwestward up the Eastern Cordillera from 4400 to 5000 m and reaches a maximum of 5800 m in the Western Cordillera. The LGM snowline altitudes were consistently below 3400 m along the Eastern Cordillera (Klein et al. 1999).

The relative roles of climate and tectonics in the evolution of the Andean orogen are difficult to determine because of the lack of detailed and extensive observations for both the long-term deformation and erosion history of the Andean fold-thrust belt. Wittmann et al. (2011) calculated a cosmogenic nuclide-derived denudation rate of 0.45 ± 0.07 mm/yr for the floodplain part of Beni drainage basin in northern Bolivia and that the Andean source area erodes at a similar rate of 0.37 ± 0.06 mm/yr. The erosion rates derived from low-temperature bedrock thermochronometry vary from 0.4 to 0.7 mm/yr in the area of the north central Andes in Bolivia (Herman et al. 2013).

2.4. Taiwan

The Taiwan orogen is relatively young even compared to the other orogens of this study, and has formed in a complex tectonic setting. According to sedimentation rates and paleomagnetism, the mountain building processes started at about 4 Ma (Suppe 1984). The Taiwan orogen is considered as a collision system of two active subduction systems. In the northeast, the Philippine Sea plate is subducting at a rate of 80 mm/yr beneath the Eurasian plate along Ryukyu Trench, thus creating the Ryukyu Arc and the Okinawa Trough backarc basin (Figure 2.4.1). In the south, the South China Sea subplate is subducting beneath the Philippine Sea plate resulting in formation of the Luzon Arc (Kim et al. 2005). Continuous subduction has brought the Luzon Arc closer to the leading edge of the Eurasian continent and caused their collision in the late Miocene. The arc-continent collision has been oblique and the point of collision has migrated southward because the Asian continental margin trends differently from the Luzon Arc (Suppe 1984). The northward subduction of the Philippine Sea plate also results in the northern Luzon Arc underthrusting the Ryukyu arc-trench system and thus causing tectonic exhumation (Huang et al. 2000). 13

The Taiwan collision system can be subdivided by the Longitudinal Valley into two major tectonic provinces (Figure 2.4.1). The Longitudinal Valley is characterized by NNE-SSW trending linear seismicity and topographic features and it is considered as the surface expression of the suture zone of the collision system. The western province can be subdivided petrologically into a few NNE –SSW trending structural belts (Kim et al. 2005). Starting from west, the Coastal Plain, Western Foothills and Hsuehshan Range are part of the passive Asian continental margin deposits now deformed into a fold-and-thrust belt. Parts of the western province also include the western Central Range –Hengchun Peninsula and eastern Central Range. The western Central Range– Hengchun Peninsula represents the pre-collision accretionary prism (Huang et al. 2000). The eastern Central Range is considered the underthrust Eurasian continent. The eastern province consists of the Coastal Range and a few small islands. The Coastal Range is considered a remnant of a Neogene island arc, which also represents the leading edge of the Philippine Sea plate (Kim et al. 2005). Basically, all the units are separated by faults or other structural discontinuities and in the western province the grade of metamorphism increases with proximity to the suture zone (the Longitudinal Valley). The Lishan-Laonung Fault, between western Central Range and Hsuehshan Range– Western Foothills, is the former boundary between the Eurasian plate and the Philippine Sea plate (Huang et al. 1997). 14

Figure 2.4.1. Major geological units and structures in Taiwan. In top left corner: EP= Eurasian plate, PSP= Philippine Sea plate, MTr= Manila Trench, RTr= Ryukyu trench, OT= Okinawa trough. Modified after Dadson et al. (2003)

The subtropical climate of Taiwan has a mean precipitation of 2500 mm/yr. The northeasterly monsoon during the cold season and the southwesterly monsoon during the warm season bring precipitation to Taiwan. The spatial distribution of precipitation is affected significantly by the orographic effect of the Central Mountain Range and the prevailing wind direction (Chen and Chen 2003). The island encounters also on average four typhoons a year, which bring heavy rainfall (up to 1798 mm of 24-h rainfall was calculated in the Central Range during the Typhoon Herb between 31 July and 1 August 1996; Wu and Kuo 1999).

The climate factors together with frequent earthquakes and weak substrate drive rapid mass-wasting and fluvial incision, resulting in high erosion rates despite low 15 topographic relief (Dadson et al. 2003). The suspended sediment erosion rates indicate that the metamorphic core of the orogen is eroding at approximately 6 mm/yr. To the south of the Western Foothills thrust belt, the erosion rate around active faults can be up to 60 mm/yr (Figure 2.4.1). North and west Taiwan experience relatively lower erosion rates from 1 to 4 mm/yr (Dadson et al. 2003). Also, the long-term erosion rates derived from the bedrock thermochronology show a similar pattern as rates of erosion recorded by shorter term measurements (Herman et al. 2013).

The glacial history of Taiwan has been recorded in few locations: Mt. Xue Shan, Mt. Nanhuta Shan and Mt. Yu Shan. Ono et al. (2003) summarized the LGM ELAs to range from 3500 to 3700 m. According to Ono and Naruse (1997), the most extensive phase of glaciation took place during Marine isotope stage (MIS) 3 to 4 (57-71 ka) and the extent of glaciation during MIS2 (29 ka, near the LGM) was lower. The Equilibrium line altitude is proposed at 3050 m for MIS3/4 (Hebenstreit and Böse 2003). In the cirque of Mt. Xue Shan, the youngest moraine has been dated at 18.3±1.5 ka (Cui et al. 2002). This terminal moraine is located at nearly the lowest end of the U-shaped valley at 3350 m elevation (Ono et al. 2005). The modern ELAs have been estimated by many authors (e.g. Benn and Lehmkuhl 2000, Cui et al. 2002, Hebenstreit and Böse 2003) and the results vary from 3900 to 4600 m.

2.5. Olympic Mountains

The Olympic Mountains are situated in the northwestern corner of the state of Washington, USA. They comprise an accretionary wedge formed by the subduction of the Juan de Fuca Plate under the North American Plate at a rate of subduction of approximately 30 mm/yr. The surface trace of the subduction thrust is marine and is situated about 80 km off the coast at 2500 meters below sea level (Hellwig 2010). In contrast, the highest peaks rise to almost 2500 meters (Mt. Olympus is the highest with an elevation of 2432 m). Further inland, the Cascade volcanic arc is parallel to the subduction zone and situated where the subducting slab lies at approximately 100 km depth and melting begins. The dip of the subducting slab is shallowest in the vicinity of the Olympic Mountains and thus the distance between the subduction zone and the arc is the greatest. The shallow dip caused the Olympic Mountains to reach elevations 16 above sea level earlier than any other section of the forearc high, about 15 Ma, implying it is the most evolved section of the forearc high (Brandon 2004).

Figure 2.5.1. Lithological assemblages and structural formations of the Olympic Mountains. Modified after Stewart and Brandon (2004)

Two lithological assemblages can be found in the Olympic Mountains related to subduction and the accretionary wedge: Peripheral and core rocks. The peripheral rocks, or Coastal Range terrane, consist of an oceanic crust unit that dips landward within the Cascadia wedge with the accreted sediments extending landward beneath the terrane (Figure 2.5.1; Brandon et al. 1998, Pazzaglia and Brandon 2001). There are different theories for the obduction of the oceanic crust on to the continent: collision of an intra- Pacific seamount province or back- or forearc rifting at the North American plate margin (Brandon et al. 1998). The Coastal Range comprises two basal units: The Crescent Formation and the overlying Peripheral Sequence. The Crescent Formation consists of pillowed massive basalt flows that are cut by dikes and interbedded with pelagic limestones and mudstones. The Peripheral Sequence is marine clastic strata from Eocene to lower Miocene (Brandon 2004, Brandon et al. 1998). The Hurricane 17

Ridge and Calawah faults separate the Coastal Range Terrane from the core rocks. The terrane is also folded into a horseshoe shaped belt that rings the core rocks from the northern, eastern and southern sides of the Olympic Peninsula (Hellwig 2010). The core rocks, or Olympic Structural Complex, are slightly metamorphosed accretionary wedge sediments. The Olympic Structural Complex is subdivided into Upper, Lower and Coastal units. The Upper unit contains Eocene clastic sediments, mostly shales and sandstones from turbidite sequences, and pillow basalts of the same age. The Lower unit is composed of late Oligocene and early Miocene clastic sediments, for the most part turbidite sandstones. The Coastal unit consists of an accretionary unit of turbidite sequences from the Miocene with bodies of Eocene pillow basalts and sedimentary rocks occurring within it (Brandon 2004, Brandon et al. 1998). According to Brandon et al. (1998) and Brandon (2004), the subducting plate is not removed completely, but rather the Juan de Fuca Plate is covered by approximately 2500 m of sediment and geophysical and geochemical footprints indicate that all of the incoming sediment is accreted into the accretionary wedge, the Cascadia Wedge, as the underlying crust and mantle are subducted with minor amounts of underplating occurring.

A maritime climate is predominant in the Olympic Mountains and the precipitation rate decreases significantly across the range from west to east. 80% of the annual precipitation is delivered by the southwesterly and westerly winds during fall and winter (Thackray 2001). The total amount of annual precipitation rises from the coastal areas (1780-2540 mm/yr) to the valleys of western side (>3175 mm/yr) (Spicer 1986). On the eastern side as clouds start to descend, air is already deprived of the highest amount of moisture and goes through adiabatic warming and drying creating a rain shadow (Hellwig 2010). On northeastern side of the Olympic Mountains, Port Angeles and Sequim receive less than 300 mm/yr of rainfall (Spicer 1986).

The Olympic Mountains have been extensively glaciated during the Quaternary. Glaciers exist at relatively low altitudes owing to the modern orographic precipitation pattern and maritime climate. The Olympic Mountains are currently home to over sixty active glaciers (Spicer 1986). These glaciers are for the most part cirque glaciers, but the wetter portions of the mountains accommodate also a few valley glaciers. The Olympic Mountains have not been overrun by continental ice sheets. The Cordilleran 18

Ice Sheet has only reached the eastern and northern sides of the Olympic range flowing into the Strait of Juan de Fuca and Puget Sound (Figure 2.5.1; Hellwig 2010).

Apatite fission-track data has been used to estimate an average long-term exhumation rate of 0.28 mm/yr for the entire Olympic Mountains Peninsula. Modern erosion rates (0.18 to 0.32 mm/yr) from sediment yield data for two major rivers of the Olympic Mountains are well comparable with the previous results (Brandon et al. 1998). Most of the eroded sediment is transported into the Pacific Ocean, where it is recycled either by tectonic accretion or by sedimentary accumulation back into the accretionary wedge. The ratio between influx of accreted sediments and outflux of eroded sediments indicates that the Olympic segment of the Cascadia margin is currently in an erosional and topographic steady state (Brandon et al. 1998).

2.6. The Southern Alps of New Zealand

New Zealand is situated in the collision zone of the Australian and Pacific plates. To the north of New Zealand and beneath the eastern North Island, the Pacific plate is subducting beneath the Australian plate. Along the Alpine Fault of the South Island, the plate bounding fault is eastward dipping with dextral oblique-slip (Figure 2.6.1). South of New Zealand, the Australian plate subducts below the Pacific plate. The Alpine Fault has experienced approximately 500 km of strike slip motion and more than 20 km of vertical motion since the Miocene, but erosion has limited peaks to up to 4000 meters above sea level (Adams 1981, Norris et al. 1990). During the last 5 Ma, the Alpine Fault has undergone a total of 90 ± 20 km of convergence between the Australian and Pacific plates (Walcott 1998). The collision has formed the Southern Alps, which extend about 500 km in a northeastern trend along the Alpine Fault.

The main rock type of the Southern Alps is Mesozoic greywacke and its metamorphic equivalent, high-grade schist. The metamorphic grade decreases from amphibolite facies at the Alpine Fault west of the Main Divide, through pumpellyite-actinolite and greenschist facies, to prehnite-pumpellyite facies east of the Main Divide. The Torlesse Terrane west of the Main Divide consists of quatzofeldspathic greywackes which contain Permian-Jurassic undifferentiated quartz, albite, microcline, and biotite mineral fragments derived from the erosion of granite, gneiss, and schist (Suggate 1978). 19

Volcanic fragments (calcic plagioclase, amphibole and pyroxene) in Torlesse Terrane rocks are rare. The Haast Schist Group is formed of metamorphosed greywackes with vacillating schistosity to extreme segregation banding and gneissic texture. The Haast Schist, east of the Main Divide and south of the Torlesse Terrane, is referred to as Otago Schist, and with a chlorite zone metamorphic grade. On the western side of the Main Divide, the Haast Schist is called the Alpine Schist and lies parallel to the Alpine Fault. The metamorphic grade within about 20 km of the Alpine Fault is of the biotite and garnet-oligoclase zones (Figure 2.6.1; Suggate 1978). Northwest of the Alpine Fault are distinct Devonian granitic plutons of the Karamea batholith, which contain quartz, oligoclase, microcline, biotite and muscovite with some accessory apatite and zircon (Suggate 1978). In the uplift of the Southern Alps, intense hydrothermal activity has taken place and deposited large quartz-calcite veins with small amounts of pyrite, galena, chalcopyrite and chlorite (Koons and Craw 1991, Smith et al. 1996, Templetona et al. 1998). Veins contain calcite at all metamorphic grades, but it is the most prevalent in the biotite and garnet zones. Overall, greywacke and schist basement rocks can contain up to 3% calcite (Jacobson et al. 2003).

Figure 2.6.1. Tectonic setting and structural formations in the Southern Alps of New Zealand. Modified after Little et al. (2005). 20

The Southern Alps have prevailing westerly winds that dominantly supply moisture to the western side of the mountain belt, creating a large climatic and hydrologic difference between the western and eastern sides of the Main Divide. The Main Divide separates the western and eastern coast water catchments from each other (Jacobson et al. 2003). Up to 12 m/yr precipitation can be experienced at the western coast of the South Island due to the moisture rich air masses arriving from the Southern Ocean (Henderson 1993, Mosley and Pearson 1997). Because of orographic effects, the leeward (eastern) side of Main Divide receives less than 1 m/yr of precipitation (Chinn 1979). Glaciers on the South Island occur along the crest of the Southern Alps from the Spencer Mountains in the north to southern Fiordland. The glaciers cover an area of 1159 km2 with total number of glaciers larger than 0.01 km2 is ~3150 (Chinn 1988). Suggate (1990), Nelson et al. (1993) and Suggate and Almond (2005) have documented nine glaciations during the last 0.7 Ma. ELA positions are considered to change up to 900 ± 50 meters between LGM and today, which has been concluded from studying few glaciers in the eastern part of the orogen and in close proximity to Mount Cook (Porter 1975).

Present uplift on the Alpine Fault is partly tectonic and partly glacio-isostatic, and uplift and erosion are roughly in balance in the Southern Alps (Adams 1980). The current rate of crustal shortening across the Southern Alps is 22±2 mm/yr. The shortening is converted into uplift as the crust of Pacific plate is decoupled from the mantle lithosphere along a curved fault plane (Adams 1980). Low-temperature thermochronometry indicates erosion rates of almost 6 mm/yr in the close vicinity to Alpine Fault. The erosion rates are lower in Fiordland and moving away from the Alpine Fault (Herman et al. 2013). Hovius et al. (1997) determined denudation rates to be 9 ± 4 mm/yr in the montane region from the sediment discharge caused by landsliding.

2.7. The Alps

The Alps are situated in south-central Europe, and the fact that the European Alps are situated on the “old continent” has caused them to be focus of research for centuries. 21

The Alps form an arcuate geometry that extends first in a north-south direction from the Mediterranean in France, then bends eastward through Switzerland and finally reaches eastern Austria in an east-west direction. The mountain belt is the result of collision of the African and European plates involving closure of ocean basins situated in the Mediterranean region. The Alps are proposed to be the product of two orogenies, which occurred in the Cretaceous and Tertiary. In the Cretaceous orogeny the Meliata Ocean closed into Apulia and in the Tertiary orogen the Alpine Tethys closed between Apulia and Europe (Schmid et al. 2004). Certain oceanic domains are kinematically related to the opening of Atlantic Ocean and called “Alpine Tethys” (the Piedmont-Liguria and Valais oceans). The “Tethyan” oceanic domains (Neotethys, Meliata Ocean and Vardar Ocean) located further to the east are not related to the previous process (Stampfli 2000).

The tectonic units of the Alps can be related to the paleogeographical domains to clarify the history of tectonic development (Figure 2.7.1) (Schmid et al. 2004). The main focus of this study is the so-called internal parts of the orogen: The Austroalpine nappes, Southern Alps, Penninic nappes and Helvetic nappes. Together with the external Dinarides, the Southern Alps represent Apulian plate south of the Periadriatic line, which is a major E-W–directed dextral strike-slip fault in the central and eastern part of the Alps (Figure 2.7.1). The Apulian plate is considered as all the continental paleogeographic regions located south of the Alpine Tethys and north of the Neotethys. According to Schmid et al. (2004), the Southern Alps are best described as a south- verging fold-and thrust belt of upper crustal slices. The Austroalpine nappes are a system of preserved basement and cover slices of the Piedmont- Liguria Ocean (Alpine Tethys) situated north of the Periadriatic line. The Austroalpine nappes are overlying the Penninic units, which consist of fragments from the distal European margin and Alpine Tethys, around the Tauern window along a thrust formed during Tertiary orogeny (Schmid et al. 2004). The Piedmont-Liguria Ocean was located directly adjacent to the Apulian margin and south of the Briaçonnais ribbon continent, with relicts that can now be observed on top of the Penninic nappe stack (Upper Penninic nappes) (Bucher et al. 2003). The Briaçonnais terrane constitutes for its part the Middle Penninic nappes. The Valais Ocean is suggested to be a more northerly located branch of Alpine Tethys. Remnants of this oceanic domain and its adjacent distal margin form the Lower Penninic nappes (Fügenschuh et al. 1999). The northern and western foreland 22 of the Alps is composed of the European margin. The Helvetic nappes are part of a detached sedimentary fold-and-thrust belt derived from the European margin which is now situated in the Swiss and westernmost Austrian Alps (Schmid et al. 2004).

Figure 2.7.1. Tectonic units and main structures of the European Alps. Modified after Stüwe and Schuster (2010).

The European Alps act as a boundary between Mediterranean-type, Atlantic, and continental climates. The alpine climate is also occasionally influenced by Saharan air masses. The Alps are situated in one of the warmest areas of the Northern Hemisphere mid-latitudes thanks to the proximity to the influence of the Atlantic Ocean and the heat reservoir contained in Mediterranean Sea (Beniston 2006). Precipitation is the highest on the northern side of the Alps, and the amount of rainfall decreases from west to east as the distance to Atlantic source regions increases (Keiler et al. 2010). The North Atlantic Oscillation causes variations in alpine climate because it determines the tracks of precipitation-bearing storms across Europe, as well as temperature anomalies (Beniston and Jungo 2002, Casty et al. 2005). During the last glacial maximum (from 21 to 22 ka) the snowline (i.e. altitude above which snow does not melt in average climate years) was at approximately 1100 m above sea level in the northern Alps and at about 2000 m above sea level in the central parts of the eastern and southwestern Alps (Schönswetter et al. 2005). During the last glacial maximum (LGM) parts of the 23 foreland basins were glaciated, but after this the glaciers never reached the same extent (Wittmann et al. 2007).

Large volumes of sediment provide clear evidence that erosion is an important exhumation process in the Alps (England 1981). Long-term bedrock thermochronometric ages show that erosion rates vary between 0.1 to 1.14 mm/yr from 2 Ma to present (Herman et al. 2013). The northern parts of the central and western Alps have the highest erosion rate values. Modern denudation rates derived from river loads and delta surveys vary between 0.03 to 0.36 mm/yr (Hinderer 2001). In the Alps erosion and tectonic uplift are operating together in an approximately steady-state fashion (Bernet et al. 2001).

2.8. Low-temperature thermochronometry

Thermochronology provides information of a rock’s path from depths towards the Earth’s surface. The exhumation can be due to the removal of rock or soil by tectonic and/or surficial processes (Ring et al. 1999). Many geologic processes operate on timescales too long to be measured with erosion rates calculated for short timescales, such as sediment fluxes in rivers that provide information from the past 1-10 years, or cosmogenic nuclide measurements made on fluvial sediments or in-situ materials that indicate the erosion rates during the last 103 -104 years. To confront this problem, there is radioisotopic thermochronometers which are sensitive to temperatures between 60-

550⁰C, typically corresponding to 2 to >10 km depths in the crust, and constraining erosion rates on timescales of 105-107years (Reiners and Brandon 2006). The basis for thermochronology is the production of an isotope or radiation damage track by nuclear decay and the thermally controlled retention of the decay products. Every thermochronometric system has an effective closure temperature at which the diffusion of noble gases through the crystal or annealing of spontaneous U238 –fission-produced damage trails becomes negligible. The effective closure temperature is specific (assuming a steady monotonic cooling history) for a thermochronologic cooling age, which tells us the duration of time that the dated mineral has been cool enough to retain the decay product. Collecting information of many thermochronometers gives us the cooling history of a rock through temperature field near the Earth’s surface, which can 24 be converted into a long-term average erosion rate over the timescale provided by the measured ages of the thermochronometric systems (Reiners and Brandon 2006). To extract exhumation rates from low-temperature thermochronometric data, a thermal model of the crust is required. Herman et al. (2013) use a sophisticated formal inverse procedure in which the depth to the so-called closure isotherm can be described as the integral of exhumation rate from the cooling age to present day. The linear Gaussian inversion method uses the implicit correlation between travel-time from closure depth to the Earth’s surface and the age-elevation relationship (see Fox et al. 2013 for details). The long-term erosion rates vary in the studied orogens from 0.1 to ~6 mm/yr (Herman et al. 2013).

3. MATERIALS

The comparison between long-term erosion rates, modern topography and climate has been accomplished with a combination of digital models of topography and annual rainfall, and estimates of long-term rates of rock exhumation from an inversion of low- temperature thermochronometer data (Herman et al. 2013). Digital topography is from the Shuttle Radar Topography Mission (SRTM version 4) with a 3 arc-second spatial resolution (Reuter et al. 2007). The SRTM data is freely available for download at http://srtm.csi.cgiar.org/SELECTION/inputCoord.asp. Additional information about the precipitation and erosion rate data is given below.

The annual precipitation data for the different orogens have been collected from several sources because no single source can provide high-resolution precipitation data for all the orogens considered in this study. Rainfall data for the Himalaya, Andes and Taiwan is from the Tropical Rainfall Measurement Mission (TRMM) data product 2B31. TRMM is satellite-derived data and has been collected from 1998 to 2009. The spatial resolution of the TRMM data is 5 km, and it has been processed by Bookhagen (in review). The processing is done in three steps: First, single TRMM orbits are projected to equally spaced 5×5 km grids using a bilinear interpolation. Then a footprint- resolution correction associated with the boosting of TRMM platform during several maneuvers in August 2001 is executed. Lastly, rainfall intensities are calibrated using mean annual and monthly rain gauge data (Bookhagen in review). The National Institute of Water and Atmospheric Research (NIWA) provided precipitation data in the 25

Southern Alps of New Zealand and it has been calculated using a spatial interpolation of meteorological data from irregularly spaced climate stations by Wratt et al. (2006). NIWA rainfall data has been collected from 1981 to 2010 and it has the spatial resolution of 500 m. The Olympic Mountains precipitation data is provided by the PRISM Climate Group (Oregon State University, http://prism.oregonstate.edu, data created 12.11.2014). PRISM data has been developed with help of climate stations and a climate-elevation regression interpolation (Daly et al. 2008). Data has been collected between 1981 and 2010 and the spatial resolution is 800 m. The precipitation in European Alps is from the EURO4M dataset. EURO4M precipitation data is also collected from a multitude of stations and then interpolated with a linear precipitation– elevation relationship algorithm similar to the one used in PRISM (Isotta et al. 2014). The EURO4M dataset is provided from 1971 to 2008. The grid spacing in this dataset is 5 km.

All of the orogens are covered with data points for long-term erosion rates, but some orogens have unfortunately sparser set of points than others. Herman et al. (2013) provides calculated erosion rates using 17833 thermochronometric ages around the world from four low-temperature thermochronometric systems: apatite (U -Th)/He, apatite fission track, zircon (U -Th)/He and zircon fission-track. The previous thermochronometers have effective closure temperatures of about 70 C, 110 C, 180 C and 250 C. Herman et al. (2013) have estimated erosion rates for the ⁰past 8 Ma,⁰ but ⁰the erosion rate data used in this study is from the last 2 Ma during which glaciation in the ⁰ northern hemisphere has been extensive and a major contributor to erosion.

4. METHODS

This chapter provides the SRTM and rainfall raw data processing methods, the instructions for using the ArcSwath tool, the theory of normalized channel steepness and the settings used for calculating the normalized channel steepness with the Stream Profiler tool. 26

4.1. The SRTM and rainfall raw data processing methods

The SRTM models for each orogen have been created by merging 5°×5° mosaic raster files together in ArcMap 10.2.1. I have executed a correction of elevation for locations below sea level and no data (voids) with the Raster Calculator tool, because no data (voids) are assigned a value of 32768 and elevations below sea level are reported as decreasing values from 65536 (e.g. an elevation of -3 m has a value of 65533 m). The steps are as follows:

Step 1: SetNull("inputraster" == 32768,”inputraster”)

Step 2: (use output file from Step 1): Con("inputraster" 32768,"inputraster" - 65536,"inputraster")

> After this procedure, each SRTM file has been projected from geographic coordinates with the WGS 1984 datum to the local UTM zone: The Alps are in zone 32N, the Bolivian Andes in 20S, New Zealand in 59S, the Olympic Mountains in 10N, Taiwan in 51N and the Himalayas are divided in two zones, 43N in the Sutlej river region and 45N in the central part of the Himalayas named after the Marsyangdi river and the country of Bhutan. The annual rainfall models and erosion rate point data have also been projected into local UTM zones. The original EURO4M rainfall dataset is in NetCDF format with each file containing daily rainfall for one month. To compile the annual rainfall, I used a Python script to sum the total rainfall from the whole time range and then divided the total rainfall by the number of years using Raster Calculator tool. 27

Figure 4.1. Digital elevation model (SRTM) of New Zealand.

To better understand and compare the factors affecting and connecting topography, precipitation and erosion rates, I have calculated several derived datasets from the original topography and precipitation data. Local relief and slope maps have been created from the SRTM digital elevation models with the ArcMap 10.2.1 Focal Statistics and Slope tools (Figure 4.2). The local relief has been calculated inside 2.5- km and 5-km diameter circles. Relief is locally the difference in elevation between the highest and lowest points within the circle. Both tectonic activity and denudation create differences in height, thus local relief is an important numerical value to describe the competition between erosion and tectonics (Kühni and Pfiffner 2001). The gradient in slope maps are calculated in degrees (°). I have also used the Raster Calculator tool in ArcMap 10.2.1 to create a hypothetical erosion potential map, where I have simply multiplied precipitation models with slope angle maps. 28

Figure 4.2. Slope angle (left) and local relief (right) maps of New Zealand. Slope angle is calculated in degrees (°). Local relief has been calculated inside a 5 km diameter circle.

4.2. How to use the ArcSwath tool?

The main tool that I have chosen to use for this study is swath profile analysis. Swath profile analysis is widely used, particularly in geomorphology, to display relationships between point and continuous spatial datasets, such as elevation, rainfall, hillslope angle or geochronological age. Swath profile analysis is useful in recognizing effects of vertical surface movements in large orogens and to identify topographic forms affected by glacial or fluvial erosion (Telbisz et al. 2013). It has also been used earlier to differentiate uplift between tectonic-geomorphic units (Wang et al. 2009). Swath profiles are rectangular extractions from digital map data into “extended” cross-sections. Cross-sections usually present minimum, mean and maximum values of various data in relation to distance along the profile with a self-selected interval.

Figure 4.3. ArcSwath toolbar in ArcMap. 29

In this study swath profiles are calculated using a Python script that has been incorporated into a new ArcMap 10.2 add-in tool named ArcSwath (https://github.com/HUGG/ArcSwath). ArcSwath allows users to create swath profiles of gridded and point data in ArcMap, reducing the time needed to create a single swath profile from hours to less than one minute. To use ArcSwath, a user first selects a raster file that has been loaded into the ArcMap table of contents (Figure 4.3, Input raster) and then click two points in the ArcMap Data View (Point 1 and 2 in Figure 4.3). The code then calculates geometrically corner points in a rectangular polygon according to the selected swath width (Figure 4.3). Currently, UTM projection is necessary for the tool to be usable because of the geometrical calculations. The polygon works as a frame when the raster file is clipped. From the clipped raster file, the code separates the length of the swath into sections from the first selected point using the distance of the parameter Increment (Figure 4.3). In each section, minimum, mean and maximum values are then calculated and the values are appended to arrays. After all sections have been processed, the values from each section are written to a CSV file and a plot of the swath data can be displayed using the Python Matplotlib library, if that option is selected (Figure 4.4. right top). Point data, such as erosion rates, can also be projected to the swath profiles in relation to the distance from the profile start, making comparison between erosion rates and modern topography and rainfall rates easier. All point data files can be selected as Input points in the toolbar (Figure 4.3). Similar to the gridded data, the ArcSwath script extracts all the points from inside the swath polygon. Then the points are projected (again, this requires UTM projection) to match the cross-section line between Points 1 and 2. The Point Feature option allows users to select variables of the point file attribute table. Again, a CSV file is written with data related to the points: distance to profile start (Point 1), point data value and the location on the profile line. Point data can also be plotted if so preferred (Figure 4.4). 30

Figure 4.1. An example from New Zealand of swath profiles (nz swath 4). On the left is the location of the swath profile on the map. Profile starts from southeast. The upper right shows an example of the profile plotted with the matplotlib library. Red line = maximum elevation (m), black line = mean elevation (m), blue line = minimum elevation (m) and red circle = erosion rate (mm/yr), which is plotted on the right y-axis. At the lower right is the same swath profile plotted with the Tecplot 360 EX 2015 R1 program using the output CSV file. Grey line = maximum value, black line = mean value, dashed grey = minimum value and red circle = erosion rate, which is plotted in the right y-axis.

For the swath profile analysis in this work (Figures 4.5, 4.6), I have tried to create swaths perpendicular to the orogen trend. The rectangular swath profiles are 50 km wide in most of the orogens, but in Taiwan and the Olympic Mountains I have also made swaths that are 25 km wide because the orogens are relatively small and the swath profiles were more informative when only 25 km wide. The interval of sampling data is normally 720 m, except in the EURO4M and TRMM data, and the erosion potential data derived from the two previous, where the spatial resolution of the input data is 5 km requiring use of a 5 km sampling interval. The lengths of the profiles vary according to the width of the orogens. The swath profiles are oriented to start from the foreland side and advance to hinterland. 31

4.3. Correlation plots

The swath profile data interpretation has been taken further by interpolating erosion rate points inside the swath along distance to find the closest mean values of topography, slope, relief, precipitation and erosion potential in the swaths (Figure 4.7). The interpolation has been executed with Python scripts for all the orogens. Results have been plotted using Tecplot 360 EX 2015 R1 and linear least squares regression lines have been calculated for the data points. Also, the equation for the line and an R2 value have been calculated. The R2 value, the Pearson correlation coefficient, expresses how well the data fits a statistical model, a line in this case. If R2 is equal to 1, the regression line fits the data perfectly, and if R2 is near 0, the line does not fit the data at all. The goal has been to recognize correlations between the modern topographic and precipitation data and erosion rates from the past 2 Ma.

Figure 4.7. An example of the correlation plots. In all subplots erosion rate is on the y-axis and on the x- axis are elevation (left top), slope angle (right top), relief 2,5 km (left middle), relief 5 km (right middle), annual precipitation (left bottom) and erosion potential (right bottom). 32

4.4. The theory of normalized channel steepness and the settings used with Stream Profiler tool

The last step of this research is calculating the normalized channel steepness index ksn and comparing it to the long-term erosion rates, rainfall data and other topographic metrics. In the following paragraphs, I explain the background for calculating the normalized channel steepness and the settings that I used in this study to calculate the normalized channel steepness indices.

The rock uplift and bedrock channel incision rate are considered to control river channel gradients and hillslope evolution in tectonically active regions (Howard 1994). Rapid tectonic uplift increases river incision and causes propagation of knickpoints along a river profile, also causing increasing local relief. Thus the long term interaction between tectonics and climate is expressed in the topography of mountainous landscapes (Castillo et al. 2014). River incision is often modelled as a power law function of contributing drainage area and local channel gradient (Howard and Kerby 1983),

ksA , -Ɵ S = where S stands for local channel slope, A is the upstream drainage area, and ks and Ɵ are the channel steepness and concavity indices, respectively. Previous equation holds above a critical threshold drainage area Acr, which can be considered as the limit between divergent and convergent topography or where debris-flow changes into fluvial processes (Montgomery and Foufoula-Georgiou 1993, Tarboton et al. 1989).

For a given river, the concavity index (Ɵ) and channel steepness index (ks) can be determined from the linear regression of the channel gradient against drainage area on logarithmic plot. A normalized channel steepness index (ksn) has also been developed to handle the problem of having small variations in the concavity index and wide variations in the steepness index. For this, a reference concavity ( ref) is required. ref is usually determined as the regional mean value in the natural undisturbed channel Ɵ Ɵ 33

segments (Wobus et al. 2006). An approximate ksn can be calculated with following equation:

ref ) ksn ks cent , (Ɵ Ɵ = where ks and Ɵ are determined from the regression, and Acent is the midpoint value between Amin and max segment bounds. Mathematically, the value of Acent can be

(log Amax+log Amin)/2 found as Acent 10 . By calculating the ksn value, we have a reference framework for evaluating the deviation of channel steepness from an ideal condition = where a river is in equilibrium because rivers that undergo tectonic forcing increase their channel gradient (Ouimet et al. 2009, Whittaker 2012). Furthermore, high ksn values along a stream profile help us detect zones that have been influenced by tectonic forcing where rivers have steep gradients and where the erosion rates can also be expected to be high (DiBiase et al. 2010, Kirby and Whipple 2012).

The channel steepness calculations have been performed using the Stream Profiler tool provided at http://www.geomorphtools.org/Tools/StPro/StPro.htm. The tool works with ArcMap10.2 and in addition MATLAB R2012a has been used to run the required geomorphtools profiler scripts. Normalized channel steepness values are calculated by first creating flow accumulation maps for the whole orogen and then each swath profile area has been extracted from the flow accumulation grids using the DEM of the swath as a mask. The profiler51_batch code has been used to perform the analysis. The batch mode settings are shown in Table 4.8. The profiler tool can only import the analyzed channels as shapefile lines and therefore lines must be converted into points with Feature Vertices To Points tool in ArcMap 10.2.1. After this, these points can also be plotted on the swath profiles. 34

Table 4.8. The Stream Profiler settings for different swath profiles are shown in the table.

Standard Exceptions in : Taiwan, Olympic Mountains, Sutlej Mb swaths 1, 5-7 and Alp swaths 6 and 11 swath 3 1e8 m2 1e7 m2 2e8 m2

ref 0.45 smoothing window 2000 m contourƟ sampling interval 20 m auto ksn 2000 m search distance 10 pixels minimum accumulation 10 pixels spike remover employed smooth profile employed

5. RESULTS

5.1. Overall

Total number of analyzed swath profiles is 48, not including the 50-km-wide swaths from Taiwan and the Olympic Mountains. From the 48 swath profiles, six variables (elevation, slope, relief in a 2.5-km-diameter circle, relief in a 5-km-diameter circle, annual precipitation and erosion potential) are compared to long-term erosion rates with a total number of 288 plots that have been examined. Correlations between the erosion rates, topographic metrics and climate are present, but not consistent among the variables analyzed. Thus, an unambiguous “good correlation” between one of the variables and the erosion rates is hard to define because they vary in different swath profiles along a given orogen and between orogens. This means that all the orogens have to be treated and studied individually. Furthermore, the R2 values for the correlations are generally low, but are still useful for comparison of the correlations within and between orogens. R2 values of >0.35 are considered to be a “good” fit and R2 >0.2 is an “acceptable” fit. These values have been selected on the basis that plots with values of R2 below 0.2 are highly scattered and those with an R2 value over 0.35 have a good visual fit to regression line. The R2 limits (0.2 and 0.35) have been selected particularly for this study and are not considered as extensively used standards. 35

5.2. Himalaya

Figure 5.2.1. The swath profile locations in the Himalayas on the SRTM topographic map.

5.2.1. Marsiyangdi and Bhutan (Mb swaths)

Erosion rates show a large variation along the front of the Himalaya in the ten Mb swaths. While the western swath profiles (Mb swaths 1 -5) have thermochronometric erosion rates from 0.15 to 3.36 mm/yr, the eastern profiles (Mb swaths 6-10) have a more limited variance in erosion rate (0.3-1 mm/yr). In addition, erosion rates vary along the distance of the swath profiles in the western profiles, but in the eastern profiles erosion rates are invariable. This may be related to the shape of the rising topography, which is different on the western and eastern sections of the mountain belt; mean elevations in the western swath profiles form a concave-up (concave) shape compared to the concave-down (convex) or straight/linear shape in the eastern swath profiles (Figure 5.2.2). 36

Figure 5.2.2. Comparison between Mb swath profiles 3 (western) and 7 (eastern). In Mb swath profile 3 the erosion rates are more variable and topography is concave up whereas in Mb swath profile 7 the topographic shape is concave down and erosion rates exhibit less variance.

In general, the correlation between long-term erosion rates and topographic/climatic variables are good in the western Mb swaths and poor in the eastern Mb swaths. Relief has the best correlation with erosion rates, with a positive slope in the western Himalaya where only one plot has an R2 value of < 0.55. In total, 24 of 60 plots have an acceptable R2 value and 19 of these are located in the western section of swath profiles (Figure 5.2.3). In the eastern section, the relationships between slope, relief, and erosion potential to erosion rate have a nearly flat regression line slope. These variables also plot together in a tight bunch in the correlation plots and the nearly horizontal regression line suggests that the erosion rates are independent of the different metrics. In the eastern Himalaya the regression line for rainfall has the highest R2 value (0.38 for Mb swath 6 and 0.27 for Mb swath 9), but the slope is close to horizontal (Figure 5.2.4). 37

Figure 5.2.3. Correlation plots for the western part of the Himalaya (Mb swath 1) have good R2 values and linear-fit trend lines with positive slopes. 38

Figure 5.2.4. Correlation plots from the eastern section of Himalaya (Mb swath 9) have flat trend lines and thus erosion rates are suggested to be independent of the other variables though R2 values are close to acceptable in many cases.

The most prominent difference in the topographic profile between the western and eastern Mb is the concave-up shape in the west and concave-down shape in the east. In addition, the maximum elevations in the west are slightly higher than in the east. A comparison in rainfall can also be performed. In the western swath profiles abundant (mean rainfall > 2000 mm/yr) rainfall reaches mean elevations as high as 3000 – 4000 meters and a strong rain shadow leads to quick decrease in the rate of precipitation along the distance of a swath profile. Whereas in the east abundant (mean rainfall > 2000 mm/yr) rainfall reaches only ~2000 meters mean elevation with a more gentle decrease along distance as the elevation rises (Figure 5.2.5). In the western Mb swaths, the maximum long-term erosion rates occur on the rising slopes where also the maximum rainfall occurs. The maximum long-term erosion rates and maximum annual rainfall are offset about 20 km at most along swath distance with maximum rainfall at 39 slightly lower elevations than the peak in erosion rates. In the eastern Mb swaths the long-term erosion rates are invariable, which is why the relation to the rainfall is confusing (Figure 5.2.5).

Figure 5.2.5. Mean elevation and mean precipitation on the left y-axis, and erosion rates on the right y-axis versus distance along the swath profile. Mb swath profiles 2 and 8 show that in the western Himalaya (Mb swath 2) rainfall occurs on average at higher elevations than in the eastern Himalaya (Mb swath 8). Also the shapes of the orogen profiles and erosion rate patterns are very different between western and eastern swath profiles.

5.2.2 Sutlej

The four swath profiles in Sutlej region have erosion rates between 0.1-1.8 mm/yr with maximum erosion rates occurring at the mean elevation of 3000 to 3500 meters. In this way, the swath profiles in the Sutlej region have many similar features to the western section of Marsiyandi and Bhutan (Mb swaths): The topographic profile has the same 40 concave-up shape, the amount of rainfall has decreased significant at around 3000 meters elevation (rain shadow) and relief has the best correlation with erosion rate (5.2.6). Two out of four correlation plots for relief in 2.5 and 5 km windows have good R2 values (> 0.4).

Figure 5.2.6. Correlation plot for Sutlej swath 2 show good correlation between the topographic metrics and the long-term erosion rates.

On average the amount of precipitation in Sutlej is lower than in the western Mb swaths and the decrease in amount of precipitation does not occur as fast in the Sutlej region as in the western Mb region when moving to higher elevations along the swath profile. The heaviest rainfall is experienced at ~1500-2000 m elevation. Many features seem also to be part of a continuing change moving to the west along the mountain belt from the western section of Mb swaths; erosion rates are variable, but the maximum of erosion rate is already lower than in western Mb swaths. The previous observations can be 41 explained in part, if we consider that the monsoon precipitation along the length of the Himalaya decreases from east to west (Hodges 2000). Also, Sutlej swath 4 is different from the rest of the regions swath profiles: The topographic shape is not as clearly concave up and erosion rates do not vary along the distance in same pattern as in the other Sutlej swath profiles (Figure 5.2.7). This does not appear to be related to a decrease in rainfall in westernmost Sutlej swath, as no such decrease is observed.

Figure 5.2.7. Mean elevation and precipitation on the y-axis on the left, and erosion rates on the right y- axis versus distance along the swath profile. Comparison of Sutlej swath profiles 2 and 4. Erosion rate points form an obvious peak in the Sutlej swath 2, but in the swath 4 the peak is not as notable and the erosion rates show less variation. 42

5.3. Central Andes (Bolivia)

Figure 5.3.1. The swath profile locations in the Central Andes on the SRTM topographic map.

Out of the seven swath profiles in the Central Andes only four swaths have long-term erosion rate points available (Bolivia swaths 1, 2, 3 and 5) with limited spatial distribution. In general, the total number of erosion rate points is small compared to the other mountain ranges. The points are also not well distributed across the whole area of swath profiles; they are instead located not much more than 50 km apart from each other in the swaths. Furthermore, the low-temperature thermochronometer-derived erosion rate does not show much variance (0.4 -0.7 mm/yr). In Bolivia swaths 1 and 2 the maximum erosion rate values are at the peaks of the Eastern Cordillera whereas in Bolivia swaths 3 and 5 erosion rates are offset from the peak elevations. 43

For all of the Bolivia swaths with erosion rate data, there is no consistent good correlation between any of the variables and long-term erosion rates. In Bolivia swaths 1 and 5, erosion potential has good R2 values, but the regression line slope is negative in Bolivia swath 1 and positive in Bolivia swath 5. Correlation plots of Bolivia swaths 1 and 3 have the same pattern, but for elevation and precipitation (Figure 5.3.2). In swath profiles 1-3, all erosion rate points are situated in the Eastern Cordillera, but in the swath profile 5 the points are located in the Subandes, which is the present tectonically active part of the orogen (McQuarrie et al. 2005). Lastly, the maximum erosion rates on swath 5 are at the same location as the peak of precipitation (Figure 5.3.3). Nonetheless the maximum erosion rate is no more than 0.1 mm/yr higher than in the other swath profiles.

Figure 5.3.2. The correlation plots in Central Andes do not have consistency though some of the plots have good R2 values. 44

Figure 5.3.3. In the Bolivia swath 5 mean elevation (m) and mean precipitation (mm/yr) are plotted to the left y-axis. Erosion rate points are situated in the Subandes, which is considered to be recently tectonically active. Nevertheless the erosion rates are not significantly higher than in the other swath profiles where the erosion rate points are situated in the Eastern Cordillera or Interandean zone.

The rain shadow effect is strong in the Central Andes as less than 500 mm of annual rainfall reaches the Altiplano plateau in the northern Bolivian Andes. Further south on the Puna plateau precipitation is even scarcer, but the south receives much less precipitation than the north in general. Like the Himalaya, the distribution of rainfall in the Andes appears to relate to the profile of topography. Comparison between the Bolivia swaths 1-5 versus 6 -7 shows different shapes for the topography and rainfall. The northern profiles (Bolivia swaths 1-5) have concave-up topographic profiles with peak rainfall amounts at elevations of ~1500 -2000 m similar to the western Himalaya. The southern swaths (Bolivia swaths 6 -7) have linear/straight to concave-down profiles with minor rainfall throughout the swaths as maximum annual rainfall is around 2000 mm and less than 500 mm of annual rainfall receives elevations above circa 3000 m (Figure 5.3.4), comparable to the Eastern Himalaya. 45

Figure 5.3.4. Mean elevation and mean precipitation are plotted to the y-axis on the left for Bolivia swaths 1 and 6. Comparison of the topographic shape between swath profiles north of 18°S (Bolivia swath 1) and south (Bolivia swath 6) shows a notable difference in topographic profile. There is also significant difference in the amount of rainfall on the slopes in Bolivia swaths 1 and 6. 46

5.4. Taiwan

Figure 5.4.1. The swath profile locations in Taiwan on the SRTM topographic map.

There are a total of five swath profiles with both 25 and 50 km widths and the same starting and finishing points for the island of Taiwan (Figure 5.4.1). In general, the area considered in Taiwan is smaller than in the previous regions, and 25 km swaths were calculated because the correlation plot R2 values are better and there is less scatter in the long-term erosion rates at the northern and southern ends of the orogen. Taiwan has the biggest variance together with the Southern Alps of New Zealand in erosion rates of all the studied orogens. Thermochronometric erosion rates vary from 6 mm/yr in the southeast to 0.1 mm/yr in the west (Figure 5.4.2). It is easy to notice that erosion rates 47 are remarkably higher on the eastern side of the island than on the western side (e.g., Figure 5.4.3), and erosion rates also decrease slightly in the north.

Fourteen out of 30 regression lines in the correlation plots have negative slopes and 6 of the negative regressions have an acceptable R2 value. In total 13 plots have acceptable R2 values of which slope angle showed the most common acceptable fit to the erosion rate data with 4 acceptable fits and, only one negative slope for the correlation regression line. Four of 5 correlation plots between rainfall and long-term erosion rates have negative regression line slopes suggesting there is an anti-correlation between increased rainfall and erosion rate, and 2 of these have good R2 values (> 0.4; Figure 5.4.2).

Figure 5.4.2.Example correlation plots for Taiwan swaths 2 and 4, which both have negative regression line slopes and good R2 values. 48

Rainfall is abundant across the island with a maximum around 5500 mm/yr. The Western side of the mountain range receives more precipitation than the eastern side (Figure 5.4.3). On the western side of the orogen, the maximum rainfall occurs at ~2000 m elevation. The mountain topography is also asymmetric. Elevations and relief increase more rapidly along the swath distance on the eastern side of mountain range, whereas in the west they increase more gradually.

Figure 5.4.3. Mean elevation (m) and mean precipitation (mm/yr) on the y-axis on the left versus distance along the swath profile for Taiwan swaths 3 and 4. Taiwan swath profiles 3 and 4 show that erosion rate is significantly higher on the eastern side of the orogen. Also notable is that precipitation is abundant across the swath profiles but still slightly higher on the western side of the mountains. E = east and W = west at the ends of the swath profiles. 49

5.5. Olympic Mountains

Figure 5.5.1. The swath profile locations in Olympic Mountains on the SRTM topographic map.

There is in total three 25-km-wide and three 50-km-wide swath profiles in the Olympic Mountains, the smallest region considered. Again the narrower swath profiles have better regression line R2 values and less scatter between maximum and minimum variable values for a given distance. Erosion rates have a similar pattern from south to north along the swath profiles with a range of long-term erosion rates that varies between 0.3 to 1 mm/yr. At the windward side of the orogen in the west the erosion rate ascends to a maximum similar to both precipitation and elevation. The precipitation is the highest (~6500 mm/yr) at ~1000-1500 m elevation. At the leeward side erosion rates are surprisingly high compared to the heavy decrease in the amount of precipitation 50

(~0.6 mm/yr with rainfall of ~2000 mm/yr). In addition, the leeward side relief and slope angle stay high and correlate very well to the erosion rate throughout the orogen (Figures 5.5.2 and 5.5.3). The high relief and slope angles can result in part from the many reverse faults located east of the topographic maximum along the horseshoe- shaped wedge. The topographic metrics correlate very well to the long-term erosion rates as all have positive regression lines slopes with acceptable R2 values (> 0.25).

Figure 5.5.2. 25-kilometer wide Olympic Mountains swath profile 1 with all six variables. 51

Figure 5.5.3. 25-kilometer wide Olympic Mountains swath profile 3 with all six variables. 52

5.6. Southern Alps of New Zealand (NZ)

Figure 5.6.1. The swath profile locations in the Southern Alps on the SRTM topographic map.

The total number of swath profiles in the Southern Alps is ten, spanning the entire southern island of New Zealand (Figure 5.6.1). The low-temperature thermochronometer derived erosion rates have a large variance along the mountain belt in north-south direction. The southern part (NZ swaths 1-4) of the Southern Alps has experienced long-term erosion rates of less than 1.5 mm/yr (Figure 5.6.2), while the central section (NZ swaths 5-8) has erosion rates up to 5.6 mm/yr (Figure 5.6.3). Erosion rates decrease again further north with erosion rates up to 3.5 mm/yr (NZ swaths 9 -10; Figure 5.6.5). The long-term erosion rate maximum occurs on the western slopes of the Southern Alps. 53

The R2 values for the correlation plots are the best for precipitation and erosion potential in the central section (NZ swaths 5-8 have positive regression line slopes and good R2 value of which only one value is only acceptable). Neither elevation nor the topography metric values show good correlation with erosion rates in the central section. In the southern section, the variation in long-term erosion rates is small leading to nearly flat regression line slopes for correlation plots. Furthermore, the R2 values are poor (< 0.2). In NZ swaths 9 and 10, the topographic metrics have consistently acceptable R2 values.

Figure 5.6.2. The NZ swath profile 2 in the south has low erosion rates though abundant precipitation.

The maximum rainfall occurs on the western side of the Southern Alps similar to the maximum long-term erosion rates. The maximum rainfall in the southern section (8000 to 10000 mm/yr) is almost the same magnitude as in the central section (10000 -12000 54 mm/yr) where the Alpine Fault is present, though the Alpine Fault is located beneath the sea to the west of Fiordland in the south (Figures 5.6.2 and 5.6.3). In the northern section, the precipitation maximum is lower, around 6500 mm/yr. On the eastern side of the entire mountain range the precipitation magnitude is much less, and the maximum annual rainfall is around 2000 mm/yr.

Figure 5.6.3. NZ swath 6 from the central section of the southern Alps. In the vicinity of the Alpine Fault the erosion rates and precipitation are the highest, suggesting precipitation, erosion and tectonic activity may be related here. 55

Figure 5.6.4. Mean elevation and precipitation on the y-axis on the left versus distance along the swath for NZ swath profile 5. It is clear that the highest erosion rates are on the western slopes of the orogen.

Although most of the rainfall occurs at elevations of 1000 -1500 m on the western side of the orogen, the peak in erosion rates does not always correlate with the peak in rainfall. In the central section, the long-term erosion rate peaks are situated at lower elevations than the maximum precipitation (Figure 5.6.3). As noted above, the erosion rates are the highest in the central section on the western side where the Alpine Fault is also located (Figure 5.6.4). In contrast, in the northern part of the South Island the Marlborough fault zone and the peak of erosion rate are closer to the peak of the mountain elevations (Figure 5.6.5).

Figure 5.6.5.Mean elevation and precipitation on the y-axis on the left versus distance along the swath for NZ swath10. NZ swath profile 10 in the vicinity of Marlborough fault zone. 56

5.7. The Alps

Figure 5.6.1. The swath profile locations in the Alps on the SRTM topographic map.

In the twelve swath profiles of the European Alps (Alps swaths 1-12) long-term erosion rates vary from 0.12 to 1.14 mm/yr, with western Alps having more variance and higher erosion rates than the eastern Alps. In the western Alps, the highest erosion rates can be observed on the northwestern side of the mountain front. In the eastern part of the orogen, the erosion rate points are fewer in number and less spread, in addition erosion rates along swath profile length have little variation (Figure 5.7.2). The topographic profile of the orogen also changes along strike in the Alps. The swath profiles show that in the western part (Alp swaths 1-6) the elevation decreases more rapidly from the 57 peaks on the southeastern side of the orogen than on the northwestern side. To the east the shape of the mountain range is more symmetrical on both sides of the orogen and elevation decreases steadily (Figure 5.7.2). Interestingly, high thermochronometric erosion rates are not coupled to variations in elevation over shorter distances, as can be seen in the slope and relief swath profiles (e.g., Figure 5.7.3).

Figure 5.7.2. Elevation and erosion rate comparison between Alps swath profile 3 and 8. Notable is the difference in topographic profile and variance in erosion rates. Maximum elevation is grey thin line, mean elevation is black thick line, minimum elevation is grey dashed line and erosion rates are red dots.

In the correlation plots, elevation has the strongest relation to erosion rate, especially in the western part of the orogen. In the west, 7 out of 12 swaths have positive regression line slopes with acceptable R2 values and 5 swaths have good R2 values. Other topographic values have only slightly worse correlations to the erosion rates than elevation (Figure 5.7.3 and 5.7.4). In contrast to elevation, precipitation and erosion potential do not have good correlation to the erosion rates with only two plots having positive regression slopes and good R2 values. Interestingly, out of 72 correlation plots 58 for the Alps, 35 have acceptable R2 values of which only two plots have negative regression line slopes. The total number of negative slopes among all the plots is 20, indicating that most negative slope occur for poor fits to the erosion rate data.

Figure 5.7.3. Correlation plots for the topographic metrics from Alps swath profile 1 in the western Alps. 59

Figure 5.7.4. Correlation plots for the topographic metrics from Alps swath profile 5 in the western Alps.

In the western part of the orogen, precipitation is slightly focused on the north side of the mountains. In the western Alps, the maximum rainfall (~2000 mm/yr) occurs at ~1000-1500 m elevation on the northwestern side of mountains whereas close to the same amount of rainfall is experienced at ~1000 m elevation on the southeastern side. In the eastern Alps, the maximum rainfall occurs at ~1000-1500 m elevation on the northern side of the mountains while most rainfall is experienced at ~500-1000 m elevation on the southern side (Figure 5.7.5). In the east the total amount of rainfall is close to similar on both sides of the orogen. 60

Figure 5.7.5. Elevation, precipitation and erosion rates from Alps swath profiles 4 and 9. Elevation and precipitation are presented in the y-axis on the left and erosion rate in the y-axis on the right.

6. DISCUSSION

6.1. Global variations in erosion of young orogens

A consistent correlation between the different topographic metrics, climate and long- term erosion rates is not evident in the results of this work. Thus, there is not a single topographic or climatic metric that can be used on a global scale to identify variations in long-term erosion rates in different orogens. 86 (30%) of the plots have a good R2 value (> 0.35) for the correlation regression line and 135 (47%) have an acceptable R2 value (> 0.2). Most of the plots with a good correlation have a positive slope for the trend line as only 11/86 have negative slopes. Overall, elevation has the worst correlation with long-term erosion rates, which is expected as the maximum erosion rate does not need to occur on the peak of a mountain where many erosion processes are inefficient, and 61 instead a nonlinear fit between erosion rates and elevation may be more appropriate. All the other variables have similar statistical correlation values as all have 15±2 /48 swaths with good statistical fit (R2 values > 0.35). Relief in a 2.5-km-diameter circle has the best correlation to long-term erosion rates as 16/48 have good R2 values and only one negative regression line slope. However, individual inspection is required because the good correlations are not evenly distributed throughout the orogens. Often there are swaths located next to each other in an orogen that have good R2 regression values for the same variable and that seem to form sections within the orogen. The sections can often be interpreted to have some kind of tectonic or geomorphic coupling (e.g., the western Alps and the western Mb). Lastly, note that a negative trend line slope with a good R2 value is also considered as good correlation in the analysis.

In the slope, 2.5-km relief and 5-km relief topographic metrics, it is notable that in all orogens the maximum steepness for the mean slope is ~30°. According to Kühni and Pfiffner (2001), 25° is a threshold value for hillslopes in the Swiss Alps, where hillslopes that exceed this critical slope angle become unstable and mass wasting processes such as bedrock landsliding start to reduce the slope angles (Gabet et al. 2004). In addition, because the mean critical slope angle seems to be same rate in all the orogens of this study, the local relief inside the 2.5 kilometer diameter circle seems to be limited by the maximum value of the mean slope. For example, the maximum of mean relief inside a 2.5 km diameter circle is between 1000 and 1500 meters in all the orogens. This could also be indicative of similar average bulk rock strength in the orogens, however locally the lithologies include a wide spectrum of rock types. In contrast to the smaller relief window, the maximum value of mean relief inside a 5- kilometer-diameter circle shows more variation (1500 to 3000 meters). Furthermore, the relief value increases in larger orogens. Thus, the distance and scale between valley bottoms and peaks would be affected by the size of the orogen.

The calculated normalized channel steepness index (ksn) values range from 0 to 4500, but values above 2000 are scattered from the rest of the data points. In the orogens I consider, swath profiles with concave topographic profiles have the largest ksn values, whereas the convex or straight topographic profiles do not have a single large peak in the ksn values. In addition, the average ksn values are smaller in swaths with convex or 62

straight profile. The ksn values also seem to be bigger in larger orogens such as the

Himalaya and the Central Andes where ksn > 1000. On the other hand, Taiwan has ksn values ~800, which is about the same or larger than the maximum ksn values in the Alps and the Southern Alps of New Zealand.

The ksn values are also not qualitatively comparable to long-term erosion rates across all the orogens. For example, a ksn value of 200 does not represent a particular erosion rate. Even on the scale of individual orogens that kind of generalization cannot be made though the ksn values often have a similar pattern as the erosion rates. In addition, the calculated normalized channel steepness (ksn) values correlate well with the relief and slope swaths. This is expected because the calculated ksn values, relief and slope angle are related. The best correlation is between ksn and relief in a 5-km-diameter circle, where the two metrics seem to covary. However, this covariance is not observed in a few cases; these occur as ksn gaps that have concave up shape in few swaths (Figure

6.1.1, 6.1.2). The reason for the ksn gaps is not clear, but they do not appear to result from data processing errors nor a scarcity of data points.

Figure 6.1.1. Relief in a 5-km diameter circle on the left y-axis with long-term erosion rates and normalized channel steepness index values on the right y-axis for the Sutlej swath 1. The relief and normalized channel steepness index seem covary nicely except for the location where the red ksn gap is indicated. 63

Figure 6.1.2. Relief in a 5-km diameter circle on the left y-axis with long-term erosion rates and normalized channel steepness index values on the right y-axis for the Alps swath 3. The relief and normalized channel steepness index seem covary nicely except for the location where the red ksn gap is indicated.

6.2. Swath profile analysis implications for tectonics

Though the swath profile analysis does not indicate a unique correlation between any of the topographic or climatic metrics and long-term thermochronometric erosion rates, in certain sections or parts of the orogens the metrics do seem to covary or have an obvious relationship. One interesting feature of the swath profile analysis is that it seems to provide an effective tool for identifying division of tectonic or tectonomorphic sections by interpreting topographic metrics, pattern of precipitation or/and variations in erosion rates. This kind of observation can be made both along the strike of an orogen by comparing different swath profiles and along the swath profile distance. For example, the western European Alps (swaths 1-6) could be distinguished from the eastern Alps (swaths 7-12) by the asymmetric topographic shape and more variable erosion rates, which can be related to the more recent tectonic events in the western European Alps.

Another important aspect is the detection of the main (or even minor) faults in an orogen. In the Southern Alps of New Zealand, the division between the southern, central and northern sections can be related to the Alpine fault and Marborough fault 64 zone. The southern section (NZ swaths 1-4) can be identified by the comparably low erosion rates though the precipitation is high. In the central section (NZ swaths 5-8) both high erosion rates and precipitation are focused on the western side of the orogen. In this section, the Alpine fault is also located at the western slopes of the mountain range. In the north (NZ swaths 9 and 10), the Alpine fault is splitting in to the Marlborough fault system, which is situated inland encircling the topographic peak. Rainfall has decreased dramatically and erosion rates have decreased from the highest values, but are still high and variable. Here, the peak of erosion rate is located close to the maximum elevation. Another region, where tectonic activity might play an interesting role is Taiwan. Taiwan has abundant precipitation across the island with even slightly higher rainfall rates on the western side of the orogen. Still, the maximum long-term erosion rates are experienced on the eastern side of the island, which also hosts many active faults and tectonically recently uplifted units (e.g. the Longitudinal Valley Fault separating the Eastern Coastal range and the Longitudinal Valley) (Huang et al. 1997).

In many of the studied orogens the mean topographic profile can be observed to vary between concave and straight/convex. The swaths in the northern Bolivian Andes and the western-central Himalayas can be considered as concave and the southern Bolivian Andes and the eastern Himalayas as straight/convex. The variation in topographic shape can also be linked to climatic or erosional factors. The swath profiles with concave- upward shape often have a peak in erosion rate either on the steep regions of the profile or at the crest of the mountain. Also, the rain shadow in these areas is typically strong. Precipitation has a peak on the steep flanks of the mountain, but it is often offset slightly from the maximum of the thermochronometric erosion rates. Combined, this suggests that high rainfall causes high erosion, which is more efficient than the tectonic uplift if rivers are the main erosive agent. On the other hand the swath profiles that resemble a straight or convex shape seem to have invariant and low erosion rates in comparison to what the maximum is at the orogen scale. In these regions, it appears that the low erosion rate causes minimal change in topography and thus the tectonic uplift dominates the morphology. 65

6.3. Recognizing glacial erosion

Another important question is whether the differences in topography and erosion rates from glaciated and fluvial landscapes can be clearly seen in topographic swath profile analysis. Based on this analysis, separating glacial erosion from river incision by interpreting topographic shapes in swath profiles is difficult. One problem can be that glaciers and rivers can occupy the same valleys sequentially (Montgomery 2002), potentially mixing their topographic signatures. In the orogens where extensive glaciation has been reported during the last 2 Ma (Himalayas, Southern Alps, Alps and Olympic Mountains), swath profiles show a pattern of maximum erosion rate where the present precipitation is highest or slightly upstream from that point, except for the central section of the Southern Alps. In the central section of the Southern Alps erosion rates are the highest of all the orogens and the peaks of erosion rates are located downslope or at the precipitation maximum. The rainfall magnitude is higher in the Southern Alps compared to the other orogens, so at least here the very high precipitation rate and potential for efficient fluvial erosion could be dominating the glacial erosion that is normally responsible for extremely high erosion rates. Though the efficiency of glacial versus fluvial erosion is dependent of many factors (e.g., cold-based or warm- based glaciers, tropical or desert rivers, etc.), Hallet et al. (1996) suggested that sediment yield is higher in glaciated than in unglaciated valleys. This suggests glaciers erode landscapes more efficiently than rivers. On the other hand, incision rates comparable to glaciated regions have been documented in rivers that are experiencing fast uplift (Burbank et al. 1996).

In Taiwan and the Central Andes, glacial erosion is considered insignificant since the Miocene. The erosion rates in the Central Andes and Taiwan are very different from each other as Taiwan has high and variable erosion rates whereas the Central Andes have low erosion rates with little variation. The long-term erosion rates in the non- glaciated orogens do not appear as distinctive pattern from the rest of the orogens. Both the non-glaciated orogens have many negative slopes for the regression lines but the negative slopes are not considered to be related to the fact that the orogens have been not been glaciated during the last 2 Ma. Neither the topographic shapes in the swath profiles of Taiwan and the Central Andes show a distinctive form for the mainly 66 fluvially eroded orogens. Differences between fluvially and glacially eroded valleys are not easy to be detected in the swaths oriented across the orogen.

Owen and Benn (2005) summarized modern and LGM ELAs from the Himalayas (Table 2.2.2), where local differences in ELAs can be linked to results from swath profile analysis. In the western section of Mb swaths, the maximum of erosion rate is located at relatively high altitude, where the maximum long-term erosion rate elevations in the swath profiles exceed 6000 meters. In this region (referred to as Langtang), the modern and LGM ELAs are just above 5300 meters elevation (Fort 1995, Owen and Benn 2005, Shiraiwa 1993, Shiraiwa and Watanabe 1991). In the eastern section (referred to as Kanchenjunga), modern and LGM ELAs are reported to be above 4200 meters (Asahi and Watanabe 2000, Owen and Benn 2005, Tsukamoto et al. 2002). In the Sutlej region, the altitude around maximum erosion rate region is high enough to be linked to erosion by glaciers. In the Garhwal Himalaya (close to the swaths of Sutlej), modern and LGM ELA is approximately at 4500 meters (Barnard et al. 2004, Owen and Benn 2005, Sharma and Owen 1996) and the maximum elevations in close proximity to the maximum erosion rates exceed 5000 meters (Table 2.2.2). These observations would support the idea that glaciation has a variable influence on the erosion distribution in the Himalaya during the Quaternary. As a side point, it is worthwhile to note that glaciers do not seem to have existed at the low elevations where thermochronometric erosion rates have been calculated for eastern Himalaya, which may provide an explanation for the missing peak in erosion rates.

7. CONCLUSIONS

Based on the swath profile analysis of topography, topographic metrics and climate data and long-term rates of erosion from six orogens, the following conclusions can be made: 1. The results support the notion by Telbisz et al. (2013) and Wang et al. (2009) that swath profile analysis is a valuable tool for recognizing differences in tectonic units or regional topography at the orogen scale. The location of active faults seems often to be the basis for regional divisions; either the fault positions can be used to understand the shape of the swath profiles if fault locations are known in 67

advance or the observations from swath profiles may be used for interpreting where the faults are hypothetically located.

2. Two general forms of topographic profiles were recognized in several orogens. In the concave upward profiles, erosion is dominating as long-term erosion rates and precipitation are both typically high (e.g., the central-western Himalaya and the northern Bolivian Andes). In contrast, convex or straight topographic profiles appear to be dominated by tectonic uplift, as erosion rates and rainfall are low (e.g., the eastern Himalaya and southern Bolivian Andes).

3. Differentiating glacial and fluvial signatures seems problematic because no clear difference is found in the swath profile variables for these two erosional effects. For example, U- and V-shaped valleys may be observed in some regions, but larger swath profiles do not seem to be affected by the valley forms.

4. No straightforward correlation between climatic or topographic metrics and long-term thermochronometric erosion rates was observed. Correlations between certain topographic and climatic variables can be strong within some orogens, but the lack of consistent pattern of correlation suggests topography is affected by several factors with comparable influence. The lack of a clear correlation between long-term erosion rates and climate or topography may be due to the difference in their respective timescales as climate can vary over shorter timescales than 10 -10 years. In that case, variations between fluvial and glacial erosion may have overprinted the erosional effects of one another. 68

8. REFERENCES

Adams, C.J., 1981. Uplift rates and thermal structure in the Alpine Fault Zone and Alpine Schists, Southern Alps, New Zealand. Geological Society, London, Special Publications 9, 211-222. Adams, J., 1980. Contemporary Uplift and Erosion of the Southern Alps, New Zealand. Geological Society of America Bulletin 91, 1-114. Allmendinger, R.W., Jordan, T.E., Kay, S.M.and Isacks, B.L., 1997. The evolution of the altiplano- puna plateau of the central andes. Annual Review of Earth and Planetary Sciences 25, 139-174. Asahi, K.and Watanabe, T., 2000. Past and present glacier fluctuations in Kanchenjunga Himal, Nepal. Journal of Nepal Geological Society 22, 481–490. Barnard, P.L., Owen, L.A.and Finkel, R.C., 2004. Style and timing of glacial and paraglacial sedimentation in a monsoon-influenced high Himalayan environment, the upper Bhagirathi Valley, Garhwal Himalaya. Sedimentary Geology 165, 199-221. Barnes, J.B.and Ehlers, T.A., 2009. End member models for Andean Plateau uplift. Earth- Science Reviews 97, 105-132. Barnes, J.B., Ehlers, T.A., Insel, N., McQuarrie, N.and Poulsen, C.J., 2012. Linking orography, climate, and exhumation across the central Andes. Geology 40, 1135-1138. Barry, R.G., 1994. Past and potential future changes in mountain environments: a review, in: Beniston, M. (Ed.), Mountain Environments in Changing Climates. Routledge Publishing Company, London and New York. Beaumont, C., Fullsack, P.and Hamilton, J., 1992. Erosional control of active compressional orogens, in: McClay, K.R. (Ed.), Thrust Tectonics. Springer Netherlands, pp. 1-18. Beniston, M., 2006. Mountain Weather and Climate: A General Overview and a Focus on Climatic Change in the Alps. Hydrobiologia 562, 3-16. Beniston, M.and Jungo, P., 2002. Shifts in the distributions of pressure, temperature and moisture and changes in the typical weather patterns in the Alpine region in response to the behavior of the North Atlantic Oscillation. Theor Appl Climatol 71, 29-42. Benn, D.I.and Lehmkuhl, F., 2000. Mass balance and equilibrium-line altitudes of glaciers in high-mountain environments. Quaternary International 65-66, 15-29. Bernet, M., Zattin, M., Garver, J.I., Brandon, M.T.and Vance, J.A., 2001. Steady-state exhumation of the European Alps. Geology 29, 35-38. Bookhagen, B., in review. High resolution spatiotemporal distribution of rainfall seasonality and extreme events based on a 12-year TRMM time series, in review. Bookhagen, B.and Strecker, M.R., 2010. Modern Andean rainfall variation during ENSO cycles and its impact on the Amazon Basin, in: Hoorn, C., Vonhof, H., Wesselingh, F. (Eds.), in Neogene history of Western Amazonia and its significance for modern diversity. Blackwell Publishing, Oxford. Brandon, M.T., 2004. The Cascadia subduction wedge: the role of accretion, uplift, and erosion., in: van der Pluijm, B.A., Marshak, S., Wiegman, L.A.W. (Eds.), Earth Structure: An Introduction to Structural Geology and Tectonics. WCB/McGraw Hill Press, New York, pp. 566-574. Brandon, M.T., Roden-Tice, M.K.and Garver, J.I., 1998. Late Cenozoic exhumation of the Cascadia accretionary wedge in the Olympic Mountains, northwest Washington State. Geological Society of America Bulletin 110, 985-1009. Bucher, S., Schmid, S.M., Bousquet, R.and Fügenschuh, B., 2003. Late-stage deformation in a collisional orogen (Western Alps): nappe refolding, back-thrusting or normal faulting? Terra Nova 15, 109-117. 69

Burbank, D.W., Leland, J., Fielding, E., Anderson, R.S., Brozovic, N., Reid, M.R.and Duncan, C., 1996. Bedrock incision, rock uplift and threshold hillslopes in the northwestern Himalayas. Nature 379, 505-510. Burg, J.-P., Proust, F., Tapponier, P.and Chen, G., 1983. Deformation phases and tectonic evolution of the Lhasa block (southern Tibet,China). Eclogae Geologicae Helvetiae 76, 643–665. Castillo, M., Muñoz-Salinas, E.and Ferrari, L., 2014. Response of a landscape to tectonics using channel steepness indices (ksn) and OSL: A case of study from the Jalisco Block, Western Mexico. Geomorphology 221, 204-214. Casty, C., Wanner, H., Luterbacher, J., Esper, J.and Böhm, R., 2005. Temperature and precipitation variability in the European Alps since 1500. International Journal of Climatology 25, 1855-1880. Charrier, R., Hérail, G., Pinto, L., García, M., Riquelme, R., Farías, M.and Muñoz, N., 2013. Cenozoic tectonic evolution in the Central Andes in northern Chile and west central Bolivia: implications for paleogeographic, magmatic and mountain building evolution. Int Journ Earth Sciences 102, 235-264. Chen, C.-S.and Chen, Y.-L., 2003. The Rainfall Characteristics of Taiwan. Monthly Weather Review 131, 1323-1341. Chinn, T.J., 1979. How wet is the wettest of the wet West Coast. New Zealand Alpine Journal 32, 85–87. Chinn, T.J.H., 1988. Glaciers of New Zealand, Satellite image atlas of glaciers of the world. U.S. Geological Survey professional paper, 1386. Cui, Z., Yang, C., Liu, G., Zhang, W., Wang, S.and Sung, Q., 2002. The Quaternary glaciation of Shesan Mountain in Taiwan and glacial classification in monsoon areas. Quaternary International 97-98, 147-153. Dadson, S.J., Hovius, N., Chen, H., Dade, W.B., Hsieh, M.-L., Willett, S.D., Hu, J.-C., Horng, M.-J., Chen, M.-C., Stark, C.P., Lague, D.and Lin, J.-C., 2003. Links between erosion, runoff variability and seismicity in the Taiwan orogen. Nature 426, 648-651. Daly, C., Halbleib, M., Smith, J.I., Gibson, W.P., Doggett, M.K., Taylor, G.H., Curtis, J.and Pasteris, P.P., 2008. Physiographically sensitive mapping of climatological temperature and precipitation across the conterminous United States. International Journal of Climatology 28, 2031-2064. de Silva, S.L., 1989. Altiplano-Puna volcanic complex of the central Andes. Geology 17, 1102- 1106. Dewey, J.F.and Bird, J.M., 1970. Mountain belts and the new global tectonics. Journal of Geophysical Research 75, 2625-2647. Dewey, J.F.and Lamb, S.H., 1992. Active tectonics of the Andes. Tectonophysics 205, 79-95. DiBiase, R.A., Whipple, K.X., Heimsath, A.M.and Ouimet, W.B., 2010. Landscape form and millennial erosion rates in the San Gabriel Mountains, CA. Earth and Planetary Science Letters 289, 134-144. England, P., 1981. Metamorphic pressure estimates and sediment volumes for the Alpine orogeny: an independent control on geobarometers? Earth and Planetary Science Letters 56, 387-397. Fielding, E., Isacks, B., Barazangi, M.and Duncan, C., 1994. How flat is Tibet? Geology 22, 163- 167. Fielding, E.J., 1996. Tibet uplift and erosion. Tectonophysics 260, 55-84. Fort, M., 1995. The Himalayan glaciation: myth and reality. Journal of Nepal Geological Society 11 (Special Issue), 257–272. Fort, M., 1996. Late Cenozoic environmental changes and uplift on the northern side of the central Himalaya: a reappraisal from field data Palaeogeography, Palaeoclimatology, Palaeoecology 120, 123-145. 70

Fox, M., Herman, F., Willett, S.D.and May, D.A., 2013. A linear inversion method to infer exhumation rates in space and time from thermochronometric data. Earth Surf. Dynam. Discuss. 1, 207-259. Fügenschuh, B., Loprieno, A., Ceriani, S.and Schmid, S.M., 1999. Structural analysis of the Subbriançonnais and Valais units in the area of Moûtiers (Savoy, Western Alps): paleogeographic and tectonic consequences. Int Journ Earth Sciences 88, 201-218. Gabet, E.J., Pratt-Sitaula, B.A.and Burbank, D.W., 2004. Climatic controls on hillslope angle and relief in the Himalayas. Geology 32, 629-632. Gaetani, M.and Garzanti, E., 1991. Multicyclic history of the northern India continental margin (northwestern Himalaya). AAPG Bulletin 75, 1427-1446. Galy, A.and France-Lanord, C., 2001. Higher erosion rates in the Himalaya: Geochemical constraints on riverine fluxes. Geology 29, 23-26. Gansser, A., 1964. Geology of the Himalayas. Interscience publishers, London. Grujic, D., Coutand, I., Bookhagen, B., Bonnet, S., Blythe, A.and Duncan, C., 2006. Climatic forcing of erosion, landscape, and tectonics in the Bhutan Himalayas. Geology 34, 801- 804. Hallet, B., Hunter, L.and Bogen, J., 1996. Rates of erosion and sediment evacuation by glaciers: A review of field data and their implications. Global and Planetary Change 12, 213-235. Hebenstreit, R.and Böse, M., 2003. Geomorphological evidence for a late pleistocene glaciation in the high mountains of Taiwan dated with age estimates by Optically Stimulated Luminescence (OSL). Zeitschrift fur Geomorphologie, Supplementband 130, 31-49. Hellwig, J.L., 2010. The interaction of climate, tectonics, and topography in the olympic mountains of washington state: The influence of erosion on tectonic steady-state and the synthesis of the alpine glacial history, Geology. University of Illinois at Urbana- Champaign. Henderson, R.D., 1993. Extreme storm rainfalls in the Southern Alps, New Zealand, in: Kundzewicz (Ed.), Extreme Hydrological Events: Precipitation, Floods and Droughts. IAHS, Proceedings of the Yokohama Symposium, pp. 113–120. Herman, F., Seward, D., Valla, P.G., Carter, A., Kohn, B., Willett, S.D.and Ehlers, T.A., 2013. Worldwide acceleration of mountain erosion under a cooling climate. Nature 504, 423- 426. Hinderer, M., 2001. Late Quaternary denudation of the Alps, valley and lake fillings and modern river loads. Geodinamica Acta 14, 231-263. Hodges, K.V., 2000. Tectonics of the Himalaya and southern Tibet from two perspectives. Geological Society of America Bulletin 112, 324-350. Hoke, G.D., Isacks, B.L., Jordan, T.E., Blanco, N., Tomlinson, A.J.and Ramezani, J., 2007. Geomorphic evidence for post-10 Ma uplift of the western flank of the central Andes 18°30ʹ–22°S. Tectonics 26, TC5021. Hoth, S., Adam, J., Kukowski, N.and Oncken, O., 2006. Influence of erosion on the kinematics of bivergent orogens: Results from scaled sandbox simulations. Geological Society of America Special Papers 398, 201-225. Howard, A.D., 1994. A detachment-limited model of drainage basin evolution. Water Resources Research 30, 2261-2285. Howard, A.D.and Kerby, G., 1983. Channel changes in badlands. Geological Society of America Bulletin 94, 739-752. Hovius, N., Stark, C.P.and Allen, P.A., 1997. Sediment flux from a mountain belt derived by landslide mapping. Geology 25, 231-234. Huang, C.-Y., Wu, W.-Y., Chang, C.-P., Tsao, S., Yuan, P.B., Lin, C.-W.and Xia, K.-Y., 1997. Tectonic evolution of accretionary prism in the arc-continent collision terrane of Taiwan. Tectonophysics 281, 31-51. 71

Huang, C.-Y., Yuan, P.B., Lin, C.-W., Wang, T.K.and Chang, C.-P., 2000. Geodynamic processes of Taiwan arc–continent collision and comparison with analogs in Timor, Papua New Guinea, Urals and Corsica. Tectonophysics 325, 1-21. Isotta, F.A., Frei, C., Weilguni, V., Perčec Tadić, M., Lassègues, P., Rudolf, B., Pavan, V., Cacciamani, C., Antolini, G., Ratto, S.M., Munari, M., Micheletti, S., Bonati, V., Lussana, C., Ronchi, C., Panettieri, E., Marigo, G.and Vertačnik, G., 2014. The climate of daily precipitation in the Alps: development and analysis of a high-resolution grid dataset from pan-Alpine rain-gauge data. International Journal of Climatology 34, 1657-1675. Jacobson, A.D., Blum, J.D., Chamberlain, C.P., Craw, D.and Koons, P.O., 2003. Climatic and tectonic controls on chemical weathering in the New Zealand Southern Alps. Geochimica et Cosmochimica Acta 67, 29-46. Jordan, E., 1998. Glaciers of South America: Glaciers of Bolivia, US Geological Survey Professional Paper 1386-I – Bolivia. US Geological Survey, Reston, VA. Keiler, M., Knight, J.and Harrison, S., 2010. Climate change and geomorphological hazards in the eastern European Alps. Kim, K.-H., Chiu, J.-M., Pujol, J., Chen, K.-C., Huang, B.-S., Yeh, Y.-H.and Shen, P., 2005. Three- dimensional VP and VS structural models associated with the active subduction and collision tectonics in the Taiwan region. Geophysical Journal International 162, 204- 220. Kirby, E.and Whipple, K.X., 2012. Expression of active tectonics in erosional landscapes. Journal of Structural Geology 44, 54-75. Klein, A.G., Seltzer, G.O.and Isacks, B.L., 1999. Modern and last local glacial maximum snowlines in the Central Andes of Peru, Bolivia, and Northern Chile. Quaternary Science Reviews 18, 63-84. Kley, J., 1996. Transition from basement-involved to thin-skinned thrusting in the Cordillera Oriental of southern Bolivia. Tectonics 15, 763-775. Koons, P.O., 1989 The topographic evolution of collisional mountain belts: A numerical look at the Southern Alps, New Zealand. American Journal of Science 289, 1041–1069. Koons, P.O.and Craw, D., 1991. Gold mineralization as a consequence of continental collision: an example from the Southern Alps, New Zealand. Earth and Planetary Science Letters 103, 1-9. Kühni, A.and Pfiffner, O.A., 2001. The relief of the Swiss Alps and adjacent areas and its relation to lithology and structure: topographic analysis from a 250-m DEM. Geomorphology 41, 285-307. Lamb, S.and Hoke, L., 1997. Origin of the high plateau in the central Andes, Bolivia, South America. Tectonics 16, 623-649. Little, T.A., Cox, S., Vry, J.K.and Batt, G., 2005. Variations in exhumation level and uplift rate along the obliqu-slip Alpine fault, central Southern Alps, New Zealand. Geological Society of America Bulletin 117, 707-723. Maloney, K.T., Clarke, G.L., Klepeis, K.A.and Quevedo, L., 2013. The Late Jurassic to present evolution of the Andean margin: Drivers and the geological record. Tectonics 32, 1049- 1065. McQuarrie, N., 2002a. Initial plate geometry, shortening variations, and evolution of the Bolivian orocline. Geology 30, 867-870. McQuarrie, N., 2002b. The kinematic history of the central Andean fold-thrust belt, Bolivia: Implications for building a high plateau. Geological Society of America Bulletin 114, 950-963. McQuarrie, N., Horton, B.K., Zandt, G., Beck, S.and DeCelles, P.G., 2005. Lithospheric evolution of the Andean fold–thrust belt, Bolivia, and the origin of the central Andean plateau. Tectonophysics 399, 15-37. Montgomery, D.R., 2002. Valley formation by fluvial and glacial erosion. Geology 30, 1047- 1050. 72

Montgomery, D.R., Balco, G.and Willett, S.D., 2001. Climate, tectonics, and the morphology of the Andes. Geology 29, 579-582. Montgomery, D.R.and Foufoula-Georgiou, E., 1993. Channel network source representation using digital elevation models. Water Resources Research 29, 3925-3934. Mosley, M.P.and Pearson, C.P., 1997. Introduction: hydrological extremes and climate in New Zealand, in: Mosley, M.P., Pearson, C.P. (Eds.), Floods and Droughts: the New Zealand Experience. New Zealand Hydrological Society, pp. 1–14. Muñoz, N.and Charrier, R., 1996. Uplift of the western border of the Altiplano on a west- vergent thrust system, Northern Chile. Journal of South American Earth Sciences 9, 171-181. Nelson, C.S., Cooke, P.J., Hendy, C.H.and Cuthbertson, A.M., 1993. Oceanographic and climatic changes over the past 160,000 years at Deep Sea Drilling Project Site 594 off southeastern New Zealand, southwest Pacific Ocean. Paleoceanography 8, 435-458. Norris, R.J., Koons, P.O.and Cooper, A.F., 1990. The obliquely convergent plate boundary in the South Island of New Zealand: Implications for ancient collision zones. Journal of Structural Geology 12, 715–725. Ono, Y., Aoki, T., Hasegawa, H.and Dali, L., 2005. Mountain glaciation in Japan and Taiwan at the global Last Glacial Maximum. Quaternary International 138–139, 79-92. Ono, Y.and Naruse, T., 1997. Snowline elevation and eolian dust flux in the Japanese Islands during Isotope Stages 2 and 4. Quaternary International 37, 45-54. Ono, Y., Shiraiwa, T.and Dali, L., 2003. Present and last-glacial Equilibrium Line Altitudes (ELAs) in the Japanese high mountains. Zeitschrift fur Geomorphologie, Supplementband 130, 217-236. Ouimet, W.B., Whipple, K.X.and Granger, D.E., 2009. Beyond threshold hillslopes: Channel adjustment to base-level fall in tectonically active mountain ranges. Geology 37, 579- 582. Owen, L.A.and Benn, D.I., 2005. Equilibrium-line altitudes of the Last Glacial Maximum for the Himalaya and Tibet: an assessment and evaluation of results. Quaternary International 138–139, 55-78. Owen, L.A., Derbyshire, E.and Fort, M., 1998. The Quaternary glacial history of the Himalaya. Journal of Quaternary Science 13, 91-120. Owen, L.A., Finkel, R.C.and Caffee, M.W., 2002. A note on the extent of glaciation throughout the Himalaya during the global Last Glacial Maximum. Quaternary Science Reviews 21, 147-157. Pazzaglia, F.J.and Brandon, M.T., 2001. A fluvial record of long-term steady-state uplift and erosion across the Cascadia forearc high, western Washington State. The American Journal of Science 301, 385-431. Porter, S.C., 1975. Equilibrium-line altitudes of late Quaternary glaciers in the Southern Alps, New Zealand. Quaternary Research 5, 27-47. Reiners, P.W.and Brandon, M.T., 2006. Using thermochronology to understand orogenic erosion. Annual Review of Earth and Planetary Sciences 34, 419-466. Reuter, H.I., Nelson, A.and Jarvis, A., 2007. An evaluation of void-filling interpolation methods for SRTM data. International Journal of Geographical Information Science 21, 983- 1008. Ring, U., Brandon, M.T., Willett, S.D.and Lister, G.S., 1999. Exhumation processes. Geological Society, London, Special Publications 154, 1-27. Roe, G.H., Stolar, D.B.and Willett, S.D., 2006. Response of a steady-state critical wedge orogen to changes in climate and tectonic forcing. Geological Society of America Special Papers 398, 227-239. Schmid, S.M., Fügenschuh, B., Kissling, E.K.and Schuster, R., 2004. Tectonic map and overall architecture of the Alpine orogen. Eclogae Geologicae Helvetiae 97, 93–117. 73

Schönswetter, P., Stehlik, I., Holderegger, R.and Tribsch, A., 2005. Molecular evidence for glacial refugia of mountain plants in the European Alps. Molecular Ecology 14, 3547- 3555. Searle, M.P., 1991. Geology and tectonics of the Karakoram Mountains. John Wiley and sons, Chichester, 358. Searle, M.P., 2007. Diagnostic features and processes in the construction and evolution of Oman-, Zagros-, Himalayan-, Karakoram-, and Tibetan-type orogenic belts. Geological Society of America Memoirs 200, 41-61. Sharma, M.C.and Owen, L.A., 1996. Quaternary glacial history of the NW Garhwal, central Himalayas, India. Quaternary Science, Reviews of Geophysics 15, 335–365. Shiraiwa, T., 1993. Glacial fluctuations and cryogenic environments in the Langtang Valley, Nepal Himalaya. Contributions from the Institute of Low Temperature Science A38, 1- 98. Shiraiwa, T.and Watanabe, T., 1991. Late Quaternary glacial fluctuations in the Langtang Valley, Nepal Himalaya, reconstructed by relative dating methods. . Arctic and Alpine Research 23, 404–416. Smith, D.W., Craw, D.and Koons, P.O., 1996. Tectonic hydrothermal gold mineralisation in the outboard zone of the Southern Alps, New Zealand. New Zealand Journal of Geology and Geophysics 39, 201-209. Smith, J.A., Mark, B.G.and Rodbell, D.T., 2008. The timing and magnitude of mountain glaciation in the tropical Andes. Journal of Quaternary Science 23, 609-634. Spicer, R.C., 1986. Glaciers in the Olympic Mountains, Washington: Present Distribution and Recent Variations., Geology. University of Washington, p. 158 p. Stampfli, G., 2000. Tethyan oceans, in: J.A., W., Bozkurt, E., Piper, J.D.A. (Eds.), Tectonics and Magmatism in Turkey and the surrounding area. Special Publications Geological Society, London, pp. 1–23. Stewart, R.J.and Brandon, M.T., 2004. Detrital-zircon fission-track ages for the “Hoh Formation”: Implications for late Cenozoic evolution of the Cascadia subduction wedge. Geological Society of America Bulletin 116, 60-75. Stüwe, K.and Schuster, R., 2010. Initiation of subduction in the Alps: Continent or ocean? Geology 38, 175-178. Suggate, R.P., 1978. The Geology of New Zealand, Wellington, New Zealand. Suggate, R.P., 1990. Late pliocene and quaternary glaciations of New Zealand. Quaternary Science Reviews 9, 175-197. Suggate, R.P.and Almond, P.C., 2005. The Last Glacial Maximum (LGM) in western South Island, New Zealand: implications for the global LGM and MIS 2. Quaternary Science Reviews 24, 1923-1940. Suppe, J., 1984. Kinematics of arc–continent collision, flipping of subduction, and back-arc spreading near Taiwan. Mem. Geol. Soc. China 6, 21–33. Tapponnier, P., Mercier, J.L., Proust, F., Andrieux, J., Armijo, R., Bassoullet, J.P., Brunel, M., Burg, J.P., Colchen, M., Dupre, B., Girardeau, J., Marcoux, J., Mascle, G., Matte, P., Nicolas, A., Tingdong, L., Xuchang, X., Chenfa, C., Paoyu, L., Guangcen, L., Naiwen, W., Guoming, C., Tonglin, H., Xibin, W., Wanming, D., Haixiang, Z., Huaibin, S., Yongong, C., Ji, Z.and Hongrong, Q., 1981. The Tibetan side of the India-Eurasia collision. Nature 294, 405-410. Tarboton, D.G., Bras, R.L.and Rodriguez-Iturbe, I., 1989. Scaling and elevation in river networks. Water Resources Research 25, 2037-2051. Telbisz, T., Kovács, G., Székely, B.and Szabó, J., 2013. Topographic swath profile analysis: a generalization and sensitivity evaluation of a digital terrain analysis tool. Zeitschrift für Geomorphologie 57, 485-513. Templetona, A.S., Chamberlaina, C.P., Koons, P.O.and Craw, D., 1998. Stable isotopic evidence for mixing between metamorphic fluids and surface-derived waters during recent 74

uplift of the Southern Alps, New Zealand. Earth and Planetary Science Letters 154, 73- 92. Thackray, G.D., 2001. Extensive Early and Middle Wisconsin Glaciation on the Western Olympic Peninsula, Washington, and the Variability of Pacific Moisture Delivery to the Northwestern United States. Quaternary Research 55, 257-270. Tsukamoto, S., Asahi, K., Watanabe, T.and Rink, W.J., 2002. Timing of past glaciations in Kanchenjunga Himal, Nepal by optically stimulated luminescence dating of tills. Quaternary International 97–98, 57-67. Walcott, R.I., 1998. Modes of oblique compression: Late Cenozoic tectonics of the south island of New Zealand. Reviews of Geophysics 36, 1-26. Wang, A., Wang, G., Zhang, K., Xiang, S., Li, D.and Liu, D., 2009. Late Neogene mountain building of eastern Kunlun orogen: Constrained by DEM analysis. J. Earth Sci. 20, 391- 400. Whipple, K.X., 2004. Bedrock rivers and the geomorphology of active orogens. Annual Review of Earth and Planetary Sciences 32, 151-185. Whittaker, A.C., 2012. How do landscapes record tectonics and climate? Lithosphere 4, 160- 164. Victor, P., Oncken, O.and Glodny, J., 2004. Uplift of the western Altiplano plateau: Evidence from the Precordillera between 20° and 21°S (northern Chile). Tectonics 23, TC4004. Willett, S., Beaumont, C.and Fullsack, P., 1993. Mechanical model for the tectonics of doubly vergent compressional orogens. Geology 21, 371-374. Willett, S.D., 1999. Orogeny and orography: The effects of erosion on the structure of mountain belts. Journal of Geophysical Research: Solid Earth 104, 28957-28981. Wittmann, H., von Blanckenburg, F., Guyot, J.-L., Maurice, L.and Kubik, P., 2011. Quantifying sediment discharge from the Bolivian Andes into the Beni foreland basin from cosmogenic 10Be-derived denudation rates. Brazilian Journal of Geology 41, 629-641. Wittmann, H., von Blanckenburg, F., Kruesmann, T., Norton, K.P.and Kubik, P.W., 2007. Relation between rock uplift and denudation from cosmogenic nuclides in river sediment in the Central Alps of Switzerland. Journal of Geophysical Research: Earth Surface 112, F04010. Wobus, C., Whipple, K.X., Kirby, E., Snyder, N., Johnson, J., Spyropolou, K., Crosby, B.and Sheehan, D., 2006. Tectonics from topography: Procedures, promise, and pitfalls. Geological Society of America Special Papers 398, 55-74. Wratt, D.S., Tait, A., Griffiths, G., Espie, P., Jessen, M., Keys, J., Ladd, M., Lew, D., Lowther, W., Mitchell, N., Morton, J., Reid, J., Reid, S., Richardson, A., Sansom, J.and Shankar, U., 2006. Climate for crops: integrating climate data with information about soils and crop requirements to reduce risks in agricultural decision-making. Meteorological Applications 13, 305-315. Wu, C.-C.and Kuo, Y.-H., 1999. Typhoons Affecting Taiwan: Current Understanding and Future Challenges. Bulletin of the American Meteorological Society 80, 67-80. Zeitler, P.K., Koons, P.O., Bishop, M.P., Chamberlain, C.P., Craw, D., Edwards, M.A., Hamidullah, S., Jan, M.Q., Khan, M.A., Khattak, M.U.K., Kidd, W.S.F., Mackie, R.L., Meltzer, A.S., Park, S.K., Pecher, A., Poage, M.A., Sarker, G., Schneider, D.A., Seeber, L.and Shroder, J.F., 2001. Crustal reworking at Nanga Parbat, Pakistan: Metamorphic consequences of thermal-mechanical coupling facilitated by erosion. Tectonics 20, 712-728. 75

9. APPENDIX A – Complete set of swath profile and correlation plots for all the examined orogens

This appendix includes a complete set of plots for the swath profile analysis and topographic/climatic data regression to the long-term erosion rate data for all six orogens in this study. A map of the swath profile locations is shown first, followed by the swath profile analysis of the topographic metrics and climate data, the ksn data and correlation plots. The orogens are in the same order as in the main text: Himalaya, Andes, Taiwan, Olympic Mountains, Southern Alps of New Zealand and the Alps. For the correlation plots, the regression line equations and the R2 values are also available. 76 77 78 79 80 81 82 83 84 85 86 87 88 89 90 91 92 93 94 95 96 97 98 99 100 101 102 103 104 105 106 107 108 109 110 111 112 113 114 115 116 117 118 119 120 121 122 123 124 125 126 127 128 129 130 131 132 133 134 135 136 137 138 139 140 141 142 143 144 145 146 147 148 149 150 151 152 153 154 155 156 157 158 159 160 161 162 163 164 165 166 167 168 169 170 171 172 173 174 175 176 177 178 179 180 181