Earth Planet. Sci. Lett.

658 (2025) 119305

Contents lists available at ScienceDirect

Earth and Planetary Science Letters

journal homepage: www.elsevier.com/locate/epsl

Untangling the interplay among tectonics, climate, and erosion in the

Himalayas using landscape evolution modeling
Yuqiang Li a , Xiaoping Yuan a,* , Charles M. Shobe b,c , Guillaume Dupont-Nivet d,e, Kai Cao a
a
State Key Laboratory of Geomicrobiology and Environmental Changes, Hubei Key Laboratory of Critical Zone Evolution, School of Earth Sciences, China University of
Geosciences, Wuhan, China
b
Department of Geology and Geography, West Virginia University, Morgantown, WV, USA.
c
U.S. Forest Service Rocky Mountain Research Station, Fort Collins, CO, USA.
d
Géosciences Rennes-UMR CNRS 6118, University of Rennes 1, Rennes, France.
e
Helmholtz Centre Potsdam, German Research Centre for Geosciences (GFZ), Potsdam, Germany.

A R T I C L E I N F O A B S T R A C T

Edited by: J.P. Avouac The interplay between tectonic rock uplift and climatically modulated erosion governs landscape evolution and
influences how mountain ranges affect climate, biogeochemical cycling, ecology, and biodiversity. The Hima-
Keywords: layas, Earth’s highest mountain range, have inspired a large body of work suggesting that Himalayan topography
Mountain growth is primarily governed by southward-propagating tectonic deformation. Here, we use a new coupled surface
Landscape evolution model
process and orographic precipitation model to test this hypothesis, and to assess the extent to which orographic
Himalayas
precipitation effects have modulated the influence of tectonics on Himalayan topography since the Neogene
Geomorphology
Uplift (circa 23 million years ago). The model is quantitatively constrained by observed topographic profiles, river
Erosion profiles, precipitation profiles, erosion rates, and thermochronologic ages from eight major rivers. Results
indicate that propagating rock uplift allows a maximum “no erosion” elevation of ~20 km, and largely governs
the formation of the present-day topography of the Himalayas, with a secondary role played by orographic-
rainfall-influenced fluvial processes as suggested by erosion/uplift ratios of 60− 70%. Modeled sediment fluxes
from the orogen are 30–40 × 106 m3/yr per 250-km width (i.e., approximately one drainage basin width). Our
methods enable the integration of diverse observations to reconstruct how tectonics and climate have interacted
to control the topographic evolution of mountain belts, and allow investigation into the long-term influence of
important geomorphic process parameters.

1. Introduction between tectonics and climate govern mountain belt topography.

It has been hypothesized that the large-scale topographic form of the
The long-term landscape evolution of mountain ranges affects Himalayas is largely controlled by southward-propagating deformation
climate, biogeochemical cycling, ecology, and biodiversity (Galy et al., (Eizenhöfer et al., 2019; Godard et al., 2009; Wolf et al., 2022).
2007). Since the foundational research on potential Late Cenozoic uplift Although tectonic processes that build Himalayan topography are
of mountain ranges by Molnar and England (1990), there has been a widely researched (Gao et al., 2016; Wolf et al., 2022), their interactions
long-term debate about the roles of tectonics and climate in mountain with the climatically-modulated fluvial erosion that usually lowers the
building. Some proposed that tectonics (e.g., mountain resistance and landscape remain elusive. Studies testing the relative importance of
friction on the main thrust) controls mountain height (Dielforder et al., tectonic and climatic factors in controlling Himalayan topography
2020; Lamb, 2006; Wang and He, 1999), while others suggested that (Adams et al., 2020; Burbank et al., 2003; Godard et al., 2014) have not
climate-modulated surface processes (e.g., fluvial and glacial erosion) reached a consistent conclusion, in part due to the inherent temporal
limit mountain height (Brozović et al., 1997; Egholm et al., 2009; and spatial limitations of geomorphological, thermochronological, and
Whipple et al., 1999). The Himalayas, the highest mountain belt in the other observational data. Burbank et al. (2003) suggested that the ge-
world, provide a natural laboratory for assessing how the interactions ometry of overthrusting dominantly influences exhumation across the

* Corresponding author.
E-mail address: [email protected] (X. Yuan).

https://doi.org/10.1016/j.epsl.2025.119305
Received 23 September 2024; Received in revised form 27 January 2025; Accepted 5 March 2025
Available online 17 March 2025
0012-821X/© 2025 Elsevier B.V. All rights are reserved, including those for text and data mining, AI training, and similar technologies.
Y. Li et al. Earth and Planetary Science Letters 658 (2025) 119305

Himalayas, but that surface processes are sufficient to balance the tec- Ma is poorly constrained with suggestions of lower elevations of 2− 3 km
tonics, by comparing apatite fission-track ages with meteorological and (Ding et al., 2017; Molnar and Stock, 2009), due to the earlier slow
geomorphological datasets. Godard et al. (2014) proposed that Hima- rising before widening outward (Wolf et al., 2022). We thus derive an
layan denudation is dominated by tectonic processes, and that precipi- initial, simple uplift function for the proto-Himalayas from 50 Ma to 23
tation plays a second-order role, by comparing 10Be denudation rates Ma, which ensures a successive transition from vertical uplift to
with modern precipitation and rock uplift rates. Adams et al. (2020) widening of topography at 23 Ma (see Appendix A.3).
found that climate, as represented by modern mean annual precipita- The Himalayan tectonic sequence of southward propagation over the
tion, strongly controls fluvial erosion in the Bhutan Himalaya. past 23 Ma is well established by thermo-kinematic modeling
Wolf et al. (2022) provided a consensus resolution to this debate by (McQuarrie and Ehlers, 2017; Wolf et al., 2022). This trend is recorded
showing that the relative influence of tectonics and surface processes in mountain ranges and foreland basins by stratigraphic, sedimento-
could be predicted by a dimensionless number (called the Beaumont logical, provenance, and thermochronological data (Li et al., 2015).
number and encompassing rheology and surface process parameters), Geophysical observations from GPS, leveling, InSAR, and microseis-
and indicated that the Himalaya–Tibet, as a non-steady-state micity data indicate that rock uplift reaches a maximum at the Hima-
strength-controlled orogen, is dominated by tectonics. This view is layan front, but decreases towards the northern plateau and the
consistent with the results of Godard et al. (2009), who interpreted that southern foreland (Feldl and Bilham, 2006). Southward propagation of
the coupling between deformation and erosion is the mechanism for rock uplift is also shown by spatial patterns in the discharge-adjusted
building plateau margins, and Eizenhöfer et al. (2019), who suggested river channel steepness index ksnQ , a proxy for rock uplift rate (Adams
that lateral advection over a fixed ramp sets patterns of the river et al., 2020), showing an asymmetric peak with rapidly
steepness (ksn ) in convergent orogens. southward-decreasing values downstream from the Himalayan front
Thus, tectonics and climate have varying degrees of control over towards the foreland (Fig. 1B).
mountain building, and Himalayan topography is largely dictated by Due to the rise of the Himalayas since the Paleogene (~50 Ma), warm
propagating deformation and resulting rock uplift over surface pro- moist air over the Indian Plate is insulated by these mountains and
cesses. However, to quantitatively test this hypothesis for the evolution produces a strong South Asian monsoon (Ding et al., 2017) (Fig. 1).
of the orogen has remained difficult. First, one would need to obtain Since the Neogene (~23 Ma), following the rapid rise of the Himalayan
quantitative evidence with which to separate the contributions of tec- topography, the monsoon has caused high rainfall through orographic
tonics and climatically-modulated erosion to modern topography. Sec- precipitation over the Himalayan front, forming trans-Himalayan rivers
ond, due to the much-debated climatic influence on exhumation and that drive significant fluvial erosion (Clift et al., 2008). Modern mete-
deformation in the Himalayas (Beaumont et al., 2001; Clift et al., 2008), orologic observations show a pronounced precipitation gradient,
it is necessary to resolve the potentially strong orographic precipitation ranging from 1 to 6 m/yr from south to north (Bookhagen and Burbank,
effects on erosion which were not considered in past studies. 2010).
Here we complement and expand on past studies by developing a
new method to quantitatively test the hypothesis that southward-
propagating deformation is the dominant control on the topographic 2.2. Coupled surface process and orographic precipitation model
form of the Himalayas. We present a new approach to inferring controls
on orogenic topography that relies on using geomorphologic and We model landscape change in the Himalayas by coupling the linear
geochronologic data to constrain the best-fit parameter values of a nu- feedback precipitation model LFPM (Hergarten and Robl, 2022) and the
merical model combining tectonic, landscape evolution, and orographic propagating uplift function of plateau growth (Yuan et al., 2021) into
precipitation processes. We first reconstruct the growth of the Hima- the landscape evolution model FastScape (Braun and Willett, 2013;
layas using numerical landscape evolution modeling, taking into ac- Yuan et al., 2019), thereby simulating the interactions among uplift,
count tectonic uplift and orographic precipitation (Braun and Willett, precipitation, and erosion which all vary in space and time during
2013; Hergarten and Robl, 2022; Yuan et al., 2019). To determine landscape evolution (Fig. 1D). FastScape solves for two-dimensional
best-fit models, we compare modeled river profiles, topographic pro- river erosion, sediment deposition, and hillslope diffusion (Appendix
files, precipitation profiles, erosion rates, and thermochronologic ages to A.1). River erosion rate, according to stream power law (SPL), is
observations from eight rivers in the Himalayas across ~1500 km. We expressed by a power-law function of the stream slope (Sn ), contributing
m
then use the best-fit models to assess the influence of drainage area (Am ) and upstream averaged precipitation (P ), and
southward-propagating deformation, fluvial erosion enhanced by controlled by the erosional efficiency coefficient (Kp ) representing
orographic precipitation, and other processes in controlling the topo- mainly lithology. The sediment deposition rate is obtained by an inte-
graphic form of the Himalayas. gration of upstream erosion rates, which is controlled by the dimen-
sionless deposition coefficient (G). Hillslope material flux is modeled
2. Materials and methods using a linear diffusion equation in which flux depends on topographic
slope and a hillslope diffusion coefficient (KD ).
2.1. Geological background, propagating uplift, and monsoon The propagating uplift function is a smoothed uplift wave evolving in
precipitation of the Himalayas space and time derived from the outward-growth model of assuming a
free “no erosion” topography at elevation h0 , a value greater than
The India-Eurasia collision led to crustal accretion and the propa- present-day observed elevation (Appendices A.3 and A.4). The slope of
gation of a fold-thrust belt responsible for the crustal shortening and the assumed “no erosion” topography and the derived rock uplift wave
rock uplift of the Himalayas (Li et al., 2015) (Figs. 1 and S1; Table S1), are controlled by another parameter: the characteristic width of prop-
well documented and dated by field investigations (DeCelles et al., 2001; agating uplift (W). Given the consistent displacement of the Himalayan
Yin, 2006) and validated by numerical experiments (Wolf et al., 2022). faults (Fig. 1C; Table S1) and stable India-Eurasia convergence since
Although the India-Eurasia collision started early in the Paleogene (~50 ~20 Ma (Molnar and Stock, 2009), we simplify our model by assuming,
Ma) (van Hinsbergen et al., 2012), thermochronologic ages (Table S1) through our propagating uplift function (Yuan et al., 2021), constant
and stratigraphic and provenance studies suggest that rapid rock uplift rates of propagation and uplift although some studies have reported
and exhumation in the Himalayas began in the early Neogene (~23 Ma) phased uplift patterns (McQuarrie and Ehlers, 2017; Wolf et al., 2022).
(Clift et al., 2008; Yin, 2006). This is supported by paleoelevation studies The simplified uplift function mimicking continuous advection gener-
in the Himalayas, indicating rapid rise from ~24− 21 Ma (Ding et al., ates a spatiotemporally smoothed uplift pattern caused by discontinuous
2017; Ibarra et al., 2023). The proto-Himalayan uplift history before 23 faults, such that the model smooths the distributions of rock uplift and

2
Y. Li et al. Earth and Planetary Science Letters 658 (2025) 119305

(caption on next page)

3
Y. Li et al. Earth and Planetary Science Letters 658 (2025) 119305

Fig. 1. Geological and climatic backgrounds in the Himalayas. (A) Location of Fig. 1B in red. (B) Topographic map. Over the Himalayas, the river networks are
colored by the precipitation-weighted variant of the normalized channel steepness index ksnQ (a proxy of uplift rate in the quasi-steady state) (Appendix B). Eight
major catchments are outlined with white (see Fig. 2 for their names). White and cyan arrows respectively show uplift propagation and moisture transport
(Appendices A.3 and A.4). The major geological structures include the Yarlung–Tsangpo suture zone (YTSZ), the Karakoram fault, the Main Frontal thrust belt (MFT),
the Main Boundary thrust belt (MBT), and the Main Central thrust belt (MCT). The black dashed rectangle outlines a zone showing a geological cross section in (C).
(C) The propagating uplift model and geological structures in the central Himalayas (near Kathmandu, Nepal). The major crust-scale faults, the Main Himalayan
Thrust (MHT), and Moho boundary are adapted from seismic reflections (Gao et al., 2016) and geological surveys (Murphy and Yin, 2003). For each of the faults, its
name acronym, initiation time (Ma), end time (Ma), slip and displacement (km) are labeled (see more details in Table S1). The geological strata beneath the
Himalayas are colored with different green colors from the Lhasa Terrane to the Quaternary basin (Taylor et al., 2021). The dark- to light- red blocks show the
southward propagation of the Himalayan topography generally following in sequence thrusting (Li et al., 2015). The brown dashed line shows the present-day
vertical velocity fields derived from observations and synthesis of GPS data (Feldl and Bilham, 2006). (D) The illustration of two-stage uplift and monsoon pre-
cipitation in the Himalayas. The time and space domains are estimated from the references and Google Earth (Appendices A.3 and A.4). The scenario at 0 Ma
corresponds to (B). The reference is set at the YTSZ in this study.

erosion. This propagating wave of vertical uplift represents the (Figs. 1B and 2A). We use the MATLAB-based software TopoToolbox to
large-scale variations of deformation over geological time, which is extract channel profiles as well as maximum and mean swath profiles
different from the material advection along a fixed crustal ramp along the eight river trunks, and transform their horizontal coordinate
assumed in previous models (Coutand et al., 2014; Eizenhöfer et al., from distance to upstream integral of drainage area χ (Appendix B)
2019). (Fig. 2). The precipitation-weighted variant of the normalized channel
The orographic precipitation model LFPM adjusts precipitation steepness index, ksnQ , is calculated for all the rivers, including the eight
dynamically in response to topographic change (Appendix A.4). trunks (Appendix B). We also compute broader swath profiles of
Incoming vapor and cloud water (of total influx Fin ), driven by wind, topography and precipitation, sampled from the area of 500-km length
undergo advection (controlled by the decay length Ll ), dispersion and catchment width (Fig. 3). In addition, we compiled 10Be-derived
(controlled by the dispersion length Ld ), and interaction (i.e., conden- erosion rates (Codilean et al., 2018) and thermochronological ages (van
sation and re-evaporation due to change of topography, controlled by der Beek and Schildgen, 2023) for the eight studied catchments (Figs. 4
the condensation length Lc ). Cloud water ultimately falls as orographic and 5), and verified that general patterns of most cooling ages and
precipitation (controlled by the fallout length Lf ). erosion rates respond to the orogen-building phase of interest in the
modeling rather than to modern active tectonics (see section 3.2).
2.3. Geomorphic and geochronologic analysis of observed and modeled We model each of the eight river systems separately, considering
river systems lateral variations of uplift, precipitation, and lithology across the
Himalayas. For each river system, we defined an initial rectangular
We chose the eight largest river systems (including trunks and domain of 730 × 250 km, and assumed an initial topography with
catchments) across the Himalayas based on the following criteria: (1) random ≤ 200-m elevation noise close to the present-day Indus-Ganga
flowing from the peaks or the interior plateau to the Indus-Ganga foreland (Fig. 1D). Our models run with initial margin uplift (~0.4 mm/
foreland, and (2) minimal fault-parallel-flowing sections of trunks yr) from 50-23 Ma and then propagating uplift (~2 mm/yr) since 23 Ma

Fig. 2. Comparison of modeled and observed river profiles. (A) Map of the eight of the main Himalayan rivers analyzed, colored by ksnQ . The headwaters of rivers 1,
4, and 6 connect with the plateau lakes as shown in Google Earth images, so their upstream parts are in the same elevations in χ − elevation plots (the dashed lines in
(B)). (B) Modeled profiles (thin lines) best fitting the observed profiles (thick lines) for each of the eight rivers (Appendix B). Right-bottom inset on each panel shows
modeled topography (the grayed base map) and rivers extracted from 10 headwaters (the lines colored by ksnQ ).

4
Y. Li et al. Earth and Planetary Science Letters 658 (2025) 119305

Fig. 3. Comparison of modeled and observed topography and precipitation. (A) Map of the Himalayan topography, overlapped with mean annual precipitation (P) of
1998 to 2007 (Bookhagen and Burbank, 2010). Eight rivers and corresponding catchments are outlines with yellow. The black boxes show the areas sampled to
calculate swath profiles in (B), and on which the following erosion rates (Fig. 4) and cooling ages (Fig. 5) are correspondingly projected. (B) Modeled and observed
topography and precipitation profiles for each of the eight catchments, according to the legend. YTSZ is the north boundary in (A) and (B).

Fig. 4. Comparison of modeled and observed erosion rates. (A) 10Be-derived tributary-catchment erosion rates in the Himalayas accessed from the Octopus archive
(Codilean et al., 2018). The map shows the (trunk) rivers 1–8 (gray thick lines) and corresponding catchments (black thin lines). (B) The stream networks colored by
the upstream area in the observed and modeled Himalayas where river 5 drains. (C) Modeled and observed short-term erosion rates for each of the eight catchments,
projected on the corresponding black box in Fig. 3A. YTSZ is the north boundary in (A) and (C).

(Appendix A.3). A time step length of 10,000 years is used. In each headwaters distributed uniformly across the 250-km width and extract
model, we omit belt curvature and assume a uniform horizontal short- channel elevation, maximum and mean elevation, and ksnQ values along
ening rate of 10 mm/yr since 23 Ma (Appendix A.2), because belt cur- trunks (Fig. 2B). Swath profiles for topography and precipitation are
vature exists at the scale of the entire orogen but is slight for a single sampled from the entire model domain (i.e., 500 × 250 km at 0 Ma)
250-km wide catchment (Fig. 1). (Fig. 3B). We extract modeled tributary-catchment erosion rates over the
At the end of each simulation corresponding to one of the eight river last time step (i.e., 10 kyr) (Fig. 4C). 10Be-derived erosion rates are
systems, we sample a set of drainage pathways originating from ten averaged over millennial time scales (Godard et al., 2014), and our

5
Y. Li et al. Earth and Planetary Science Letters 658 (2025) 119305

Fig. 5. Comparison of modeled and observed thermochronologic ages. (A) Compilation of thermochronologic ages in the eight Himalayan catchments (van der Beek
and Schildgen, 2023). Different thermochronometric systems including Apatite (U-Th)/He (AHe), Apatite fission track (AFT), Zircon (U-Th)/He (ZHe), and Zircon
fission track (ZFT) are represented by diamond, square, circle, and triangle, respectively, which are all colored by cooling ages. The eight rivers (gray) with cor-
responding catchments (black) are shown. (B) Modeled and observed thermochronologic ages for each of the eight catchments, projected on the corresponding black
box in Fig. 3A, according to the legend. YTSZ is the north boundary in (A) and (B).

6
Y. Li et al. Earth and Planetary Science Letters 658 (2025) 119305

modeled results are not obviously sensitive to the model time step, such 3. Results
that we can compare modeled versus measured erosion rates. Here, the
tributaries are defined to be the cells with drainage area less than 100 3.1. Model parameters
km2, similar to the observed tributary basins (Fig. 4B). The
low-temperature thermochronometric ages along the extracted trunks The optimal values of all the model parameters for eight river sys-
can be predicted from the modeled 50-Myr erosion history using a 1D tems are given in Table 1, showing that: (1) The maximum elevation of
heat diffusion and advection model (Braun et al., 2006) (Fig. 5B). The “no erosion” topography h0 is large in the central Himalayas but small in
observed near-surface geothermal gradients in the Himalayas is from the eastern Himalayas, implying more exhumation in the central
~50 to ~100 ◦ C/km (Table S2). Himalayas but less exhumation in the eastern Himalayas, consistent
with observed cooling ages (van der Beek and Schildgen, 2023); (2) The
2.4. Constraints on model parameters total moisture influx Fin is large in the eastern Himalayas but gradually
decreases westward, agreeing with modeled and observed moisture
In this study, we take into account the lateral variations of uplift (i.e., transport as well as monsoon precipitation trends over geological time
propagating distance, “no erosion” elevation h0 , and characteristic (Bookhagen and Burbank, 2010; Sarr et al., 2022); (3) The erosional
width W), precipitation (i.e., total incoming moisture influx Fin , efficiency Kp remains relatively consistent among the eight river sys-
condensation length Lc , fallout length Lf , and decay length Ll ), and rock tems, potentially reflecting similar material properties (e.g., lithology)
erosional efficiency Kp from the west to east Himalayas (Appendix A). from the eastern to western Himalayas.
We estimate their best-fit values by comparing the modeled and The erosion efficiency Kp and the uplift rate U (set by the parameter
observed river, topography, and precipitation profiles, because these h0 ) have competing effects on elevation. By varying the maximum “no
profiles have different sensitivities to each parameter as indicated by our erosion” elevation h0 across a range of 10–30 km and the erosion effi-
previous study (Li and Yuan, 2024). Quantitative comparisons between ciency Kp across a range of 1.1–5.4 × 10− 6 m− 0.35/yr0.55 while simul-
modeled and measured short-term erosion rates and long-term cooling taneously fitting the river profiles and comparing the erosion proxies (i.
ages give additional constraints on the model parameters. The rock e., relief, topography, erosion rate, and cooling age) for the river 5 as a
uplift and erosion that occurred in each region over the 50-Myr simu- test example, we find a unique best-fit value for h0 of approximately 20
lation time represent the relative contributions of tectonics and km (Fig. 6).
climatically-driven river erosion to Himalayan topography, and can be The presence or absence of modeled orographic rainfall dramatically
quantified by comparing the “no erosion” and “erosion” scenarios of the influences the estimates of the uplift and erosion amounts implied by the
optimal model. “no erosion” elevation h0 (Fig. 7). The optimal h0 is estimated to be 20
The main group of models uses the simplest settings of the linear SPL km for matching the modeled and observed results, using propagating
(i.e., m/n = 0.45/1) with sediment deposition (i.e., G = 1) (Table 1) to uplift and orographic precipitation in our modeling (Figs. 6 and 7A). The
simulate landscape evolution, driven by two-stage uplift and orographic use of uniform precipitation leads to underestimation of h0 by ~10 km,
precipitation (Fig. S3; Movie S1). In addition, we carry out a sensitivity influencing the resulting uplift and erosion amounts, and causing
analysis to study the influences of various conditions, including the modeled erosion rates to be too low and cooling ages to be too old
classical stream power model (SPL) (i.e., m/n = 1/2.2) with no sediment (Fig. 7B to D) when comparing to observations. The inclusion of
deposition (i.e., G = 0) (Fig. S4), different surface shortening rates orographic precipitation drives high values of precipitation at the
(Fig. S5), and superimposed uplift caused by a foreland fold-and-thrust mountain front of ~4 m/yr. Orographically enhanced precipitation can
belt (Fig. S6). largely unroof the topographic bulge at the Himalayan front, yielding
higher erosion rates and lower cooling ages that are better fits to the
observations than can be obtained in the uniform precipitation case
(Fig. 7A and D). This result highlights that incorporating orographic
precipitation is necessary for simulating landscape evolution of

Table 1
The parameters best fitting the multiple observations for each river system in the Himalayas.
Catchment Landscape evolution model Propagating uplift function Orographic precipitation model

m* n* G/̃
p† Kp Distance W h0 Fin Lc = Lf § Ll # Ld ** ϵ0 Pb
(m1-3m/yr1-m) (km) (km) (km) (km2/yr) (km) (km) (km) (%) (m/yr)
6
Range for main models†† 0.45 1 1 (1.0− 6.0) × 10− ksnQ §§ 10− 60 10− 30 0.1− 2.0 1− 100 100− 1,000 25 50 0.1
River 1 (i.e., Sutlej) 0.45 1 1 3.0 × 10− 6 180 28 15 0.60 20 500 25 50 0.1
River 2 (i.e., Ganges) 0.45 1 1 3.8 × 10− 6 150 40 20 0.70 25 500 25 50 0.1
River 3 (i.e., Gomti) 0.45 1 1 3.0 × 10− 6 120 24 20 0.78 16 500 25 50 0.1
River 4 (i.e., Karnali) 0.45 1 1 4.0 × 10− 6 170 40 20 0.80 25 450 25 50 0.1
River 5 (i.e., Trisuli) 0.45 1 1 3.3 × 10− 6 160 28 20 1.00 25 450 25 50 0.1
River 6 (i.e., Arun) 0.45 1 1 2.6 × 10− 6 200 14 15 0.80 25 350 25 50 0.1
River 7 (i.e., Puna Tsang) 0.45 1 1 3.6 × 10− 6 250 45 15 1.15 16 300 25 50 0.1
River 8 (i.e., Kuri) 0.45 1 1 3.4 × 10− 6 240 30 15 1.00 20 300 25 50 0.1
*
m/n in stream power law (SPL) is fixed to be 0.45 for the Himalayas (Adams et al., 2020; Olen et al., 2015; Wobus et al., 2006), and n is suggested to 1 in the Siwalik
Hills, Nepal (Wobus et al., 2006).
†
G/̃p controlling sediment deposition is usually close to 1 when considering precipitation, estimated from experimental and natural landscapes (Guerit et al., 2019).
§
Lc and Lf are usually assumed to be equal suggested by Hergarten and Robl (2022); Smith and Barstad (2004), which are usually 25 km in the modern Himalayas
(Hergarten and Robl, 2022).
#
Ll is about 500 km in the modern Himalayas (Hergarten and Robl, 2022).
**
Ld is usually 25 km in the modern Himalayas (Hergarten and Robl, 2022).
††
This table only list tested ranges of parameters for all main models in main text, not including the supplementary models.
§§
Propagating distance of uplift wave refers to the spatial distribution of ksnQ , detailed in Fig. 1B and Appendix B.

7
Y. Li et al. Earth and Planetary Science Letters 658 (2025) 119305

Fig. 6. The different “no erosion” height h0 and corresponding modeled results for river 5 as a test example. (A) Modeled and observed river and elevation profiles,
following Fig. 2. (B) Modeled and observed topography and precipitation profiles, following Fig. 3. (C) Modeled and observed tributary-averaged erosion rates,
following Fig. 4. (D) Modeled and observed thermochronologic ages, following Fig. 5. Arrows on the upper show the increase (+) and decrease (− ) of resulting relief
(A), topography (B), erosion rates (C), and cooling ages (D), when fitting the modeled river profiles to the observations in (A) using different “no erosion” height h0
and erosion efficiency Kp , with the other same parameters in Table 1. The height h0 = 20 km is most suitable for multiple comparisons and selected as the optimal
result for river 5 (thick boxes).

mountain belts and estimating tectonic and climatic effects on 3.2. Multiple comparisons of modeled and observed results
topography.
Because we assume uniform erosion efficiency within each of the The fit for the eight rivers indicates that the propagating uplift model
eight river systems in the Himalayas, our modeling plausibly captures reasonably replicates the observed river profiles which are convex in
the consistency of erodibility within the main parts of the Himalayas (i. their upstream reaches and concave in their downstream reaches, and
e., the Lesser and Greater Himalayas composed of metamorphic reproduces appropriately the maximum and mean elevation profiles
bedrock) but not the Indo-Gangetic foredeep (e.g., the Siwalik Hills along the rivers (Fig. 2). The upstream-most reaches of rivers 1, 4 and 6
which are much lower and are underlain by sediments rather than are not incorporated into the comparison because they extend into
bedrock). The Kp derived from our main modeling exercise, 2.6− 4.0 × plateau lakes at high elevations. The ksnQ trends are also similar between
10− 6, suggests hard to very hard rocks, compatible with granitic or high- the modeled and observed trunks. In addition, our modeling produces
grade metasedimentary rocks (Stock and Montgomery, 1999) and 4− 6 large rivers per 250 km of orogen width, comparable to one trunk
similar to the values fitted by Adams et al. (2020) (~2.5 × 10− 6 for n = per 50 km in the Himalayas.
1). Kp ranges from 10− 4− 10− 5 in the Siwalik Hills (Kirby and Whipple, Combining propagating uplift and orographic precipitation re-
2001) to 10− 6 in the Lesser and Greater Himalayas (Adams et al., 2020) produces topography broadly consistent with the Himalayas consisting
for m/n = 0.45 and n = 1, suggesting that our modeling yields estimates of a gentle plateau in the north and steep mountains in the south (Fig. 3).
of Kp consistent with past work on the Lesser and Greater Himalayas. Modeled relief is higher at peaks but lower towards the plateau interior
Although our propagating uplift model is excessively smoothed relative and the foreland, basically consistent with the observed topographic
to the real phased uplift pattern caused by a series of faults, the Kp swath profiles. We also find modeled precipitation patterns, in which
derived from our supplementary modeling with classical SPL (i.e., m /n mean precipitation rates are ~1.0 m/yr in the southern low elevations,
= 1/2.2 and G = 0) is about 1.4 × 10− 9 (Fig. S4), comparable to 2.2 × abruptly increase to ~4 m/yr in the northern high elevations, and then
10− 9 for n = 2.2 suggested by Adams et al. (2020). Previous studies decrease northward to ~0.2 m/yr, that are consistent with the broad
suggested that there is no substantial difference in the erodibility of patterns of Himalayan precipitation. An abrupt precipitation peak
different Himalayan rocks or tectonostratigraphic units (Adams et al., located at around 100 km in River 8 is caused by the Shillong plateau
2020; Godard et al., 2014), indicating that our assumption of uniform Kp growing in the front of the eastern Bhutan Himalayas, outside our study
is reasonable at these scales. area (Fig. 3).
Modeled tributary erosion rates are greatest at the leading edge of
the propagating plateau with lower erosion rates in the upstream and

8
Y. Li et al. Earth and Planetary Science Letters 658 (2025) 119305

Fig. 7. The modeled results with (A) and without (B to D) orographic precipitation, fitting observed river profiles and comparing other datasets, for river 5 as a test
example. Precipitation condition is shown in the second column. h0 is 20 km and Kp is 3.3 × 10− 6 m− 0.35/yr0.55 in (A); h0 is 20 km and Kp is 3.0 × 10− 6 m− 0.35/yr0.55
in (B); h0 is 15 km and Kp is 2.2 × 10− 6 m− 0.35/yr0.55 in (C); h0 is 10 km and Kp is 1.4 × 10− 6 m− 0.35/yr0.55 in (D). The other parameters, except the precipitation
parameters, are the same in the four models (Table 1). The first column: river and elevation profiles, following Fig. 2; the second column: topography swath profiles,
following Fig. 3; the third column: tributary-averaged erosion rates, following Fig. 4; the fourth column: thermochronologic ages, following Fig. 5.

downstream areas, consistent with observed 10Be-derived tributary- sharply to the south. Our modeled Himalayan front (i.e., the Greater
catchment erosion rates (Fig. 4). Except for a small number of high Himalaya) has a maximum exhumation of ~15 km along the rivers (i.e.,
outliers possibly caused by glaciation in catchments 1 and 7 (Adams 0 Ma in Fig. S3). More exhumation would occur at the modeled Hima-
et al., 2016; Scherler et al., 2014) and by active faults at the north end of layan front (i.e., the Lesser and Greater Himalayas) if we took the
catchment 7 (Le Roux-Mallouf et al., 2015), we find that all erosion rates foreland thrusts into account (Fig. S6). Given the geothermal gradient of
concentrate in the 0–2.0 mm/yr range and differ among the eight 50− 100 ◦ C/km (Table S2), ~15 km exhumation may cause the exposure
catchments. This is similar to our modeled results (Fig. 4), showing that of high-grade metamorphic rocks (>550 ◦ C) even though we ignore the
both observed and modeled erosion rates have different maxima effect of high pressure in the collision, approaching the metamorphic
concentrating at different places, e.g., (1) the central Himalayas (rivers 5 series (300− 700 ◦ C) along the rivers observed in the Himalayas
and 6) have experienced more rapid erosion, consistent with exhuma- (Beaumont et al., 2001). It should be noted that some of the thermo-
tion rates (van der Beek and Schildgen, 2023) and local relief (Adams chronometric data come from summits, which usually imply larger ages
et al., 2016); (2) the maximum erosion rates in the eastern Himalayas and thus cannot represent maximum exhumation, e.g., the AFT cooling
(rivers 7 and 8) occur further south, consistent with the spatial pattern of ages of 10− 20 Ma located at around 350 km in catchment 6 (Fig. 5B). In
ksnQ (Fig. 1B). addition, our idealization of a propagating uplift wave ignores some
To validate our best-fit models over million-year timescales, we also lateral structural variations specific to some Himalayan catchments. The
compare the modeled and observed ages for different thermochrono- AHe and AFT cooling ages in catchment 4 exceed modeled ages between
metric systems (Fig. 5). The general pattern of cooling ages still reflects 300 and 350 km, because the Almora-Dadeldhura klippe located there
the million-year exhumation time, although there are frequent earth- has formed since 12− 14 Ma and uplifted slower than the surrounding
quakes occurring at local faults. Focusing on the maximum erosion area (Sherpa et al., 2022; van der Beek et al., 2016). The AFT and ZHe
amounts appearing along trunk valleys, we mainly compare our pre- cooling ages in catchment 8 present are greater (i.e., 5− 10 Ma) than
dicted ages along trunk valleys with the lower bound of the observed modeled between 250 and 300 km because the Shillong Plateau in the
ages. Modeled and observed ages are roughly comparable in pattern and front of catchment 8 has accommodated a part of India-Eurasia
magnitude; minimum ages of ~1 Ma are found at the leading edge of the convergence since 9− 14 Ma, reducing uplift and erosion in the eastern
propagating plateau, and ages increase gently to the north and increase Bhutan Himalaya (Coutand et al., 2014).

9
Y. Li et al. Earth and Planetary Science Letters 658 (2025) 119305

Our models reproduce the morphologic features of most of the numerical modeling (Feldl and Bilham, 2006). Also, the resulting
Himalayas, including the Lesser, Greater and Tethyan Himalayas, using erosion rates are broadly consistent with 10Be-derived erosion rates in
the simplified tectonic conditions we impose (Figs. 2 to 5). The only pattern and magnitude (Fig. 4).
notable misfit corresponds to the foot of the Himalayas, which exhibits Our coupled model of tectonics, landscape evolution, and orographic
local rock uplift near the faults MFT and MBT (Fig. 1A). Seismic to- precipitation, constrained by diverse geomorphic and geologic data,
mography and geodynamic modeling suggest a secondary foreland fold- provides a novel line of evidence supporting the hypothesis that a dy-
and-thrust belt in front of the strongly-deformed hinterland in the namic uplift propagating outward may be primary reason for Himalayan
Himalayas (Gao et al., 2016; Wolf et al., 2021). To test the influence of topographic form. Results from data-driven landscape evolution
this localized rock uplift, we test an additional model with 0.28 mm/yr modeling that incorporates orographic precipitation are therefore
of rock uplift between the Lesser Himalaya Duplex (LHD) and the MFT consistent with geological hypotheses for Li et al. (2015), and numerical
(i.e., from 250 to 350 km along our orogen-perpendicular transect) since modeling of Wolf et al. (2022), the southward propagation of defor-
10 Ma, in addition to the main propagating uplift (Fig. S6). This sup- mation and topography in the Himalayas. Our results also suggest that
plementary model produces the missing topography, precipitation, and orographic precipitation strongly influences unroofing. Although our
cooling ages at the foot of the Himalayas. Note that two separate pre- study treats Himalayan tectonics as a boundary condition, previous
cipitation peaks (e.g., in Rivers 1− 6, Fig. 3) are caused by simultaneous work has suggested plausible mechanisms of propagation in the Hima-
activation of one old fault and another renewed fault (e.g., MBT in the layas that would yield a rock uplift pattern consistent with the one we
foreland) in the Himalayas (Yuan et al., 2024), indicating complexity used.
beyond the simplified foreland uplift (Fig. S6).
Our sensitivity analysis indicates that the uncertainties on the SPL 4.2. Quantifying relative contributions of tectonics and climate to the
slope exponent n (Fig. S4), sediment deposition coefficient G (Fig. S4), Himalayan landscape
surface shortening rate (Fig. S5), and the superimposed uplift caused by
a foreland fold-and-thrust belt (Fig. S6) have no significant influences on The hypothetical “no erosion” topography at maximum elevation h0 ,
uplift and erosion patterns. For the nonlinear and detachment-limited under only uplift without erosion and deposition (Appendix A.3), re-
SPL (i.e., m/n = 1/2.2 and G = 0), there are some trade-offs between veals the tectonic contribution to mountain building. Our tests with
n and Kp values (Goren et al., 2014), and between G and Kp values (Yuan different pairs of h0 and Kp , fitting river profiles and comparing erosion
et al., 2019, 2021). We compare the modeled and observed erosion proxies (i.e., relief, maximum topography, erosion rates, cooling ages),
rates, as well as other data, so there will be one good case for each set of indicate that landscape evolution modeling can uniquely determine the
m/n and G values to fit the observed rates. Our modeling obtains similar total uplift reflected by h0 (Fig. 6). Our best-fit h0 is about 20 km for the
results between the case with m/n = 0.45/1 and G = 1 and the case with central Himalaya (catchment 5, for example), which agrees with ~20-
m/n = 1/2.2 and G = 0 (Fig. S4). Due to the increase in the number of km “no erosion” elevation and ~15-km erosion values derived from
model grid cells when increasing the shortening rate from 10 to 20 or 30 balanced cross sections (DeCelles et al., 2001) and thermal-kinematic
mm/yr, the parameters associated with model length (i.e., W, Fin , Ll , Lc , modeling (Bollinger et al., 2006). Thus, our modeling approach in-
and Lf ) correspondingly increase. The modeled erosional efficiency Kp is corporates diverse geomorphic and geochronologic data to quantify
a little larger with higher shortening rates. However, the horizontal tectonic contributions to mountain building that are consistent with
shortening rate does not influence the modeled “no erosion” height and those derived from traditional methods.
the ratio of the eroded to uplifted volumes, which can reflect the relative One key benefit of our modeling approach is the ability to quantify
contributions of tectonics and climate, because the modeling is con- the relative contributions of tectonic uplift and climatically-modulated
strained by observed erosion rates (Fig. S5). The foreland erosion to landscape evolution by comparing a hypothetical “no
fold-and-thrust belt contributes additional sediment flux out of the erosion” scenario with the best-fit model. We reran our best-fit simula-
Himalayas since ~10 Ma, but this influence is secondary at the scale of tion with zero fluvial erosion efficiency, and compared the topographic
the entire Himalayas (Fig. S6). difference between the cases without erosion and with erosion using the
optimal values of Kp (Figs. 8A and S3); results suggest that the ratio of
4. Discussion erosion to uplift amounts falls between 60− 70% (Fig. 8B). This pro-
portion of amount lost to erosion relative to the hypothetical no-erosion
4.1. Contributions of propagating rock uplift to Himalayan topography case is that predicted by minimizing misfit between the modeled and
observed river profiles, topographic profiles, precipitation profiles,
Our best-fit models with propagating rock uplift since the Miocene erosion rates, and thermochronologic ages (Figs. 2 to 7).
(23 Ma), following a slow uplift during the Paleogene (from ~50 Ma to We can obtain time-evolving erosion/uplift history in the Himalayas
23 Ma), generate modeled topography consistent with modern obser- by tracking the sediment flux out of the model domain (of the 250-km
vations at the orogen scale. Models can replicate the geomorphological model width) from west to east and from Paleogene to Quaternary
record characterized by elevated and gradually declining ksnQ values, (Fig. 8C). The evolution of erosion/uplift since Miocene (~23 Ma) shows
tapered topography, and high-relief advection-parallel interfluves an increase in the ratio of erosion/uplift amounts through time, because
(Fig. 2). This supports previous hypotheses that the topographic signa- the uplifted rock volume increases rapidly but the remaining rock vol-
ture of the tectonic process may be preserved at the locus of active rock ume increases slowly. Sediment flux first increases rapidly and then
uplift in convergent orogens like the Himalayas (Eizenhöfer et al., slowly through time, similar to observations of the other drainages
2019). Modeled river profiles are consistent with large Himalayan rivers originating from the Himalaya-Tibet (Clift, 2006), indicating a gradually
that display convex-up profiles in their upstream reaches and stabilizing sediment flux caused by the outward propagation of tectonic
concave-up profiles in their downstream reaches (Fig. 2) (Adams et al., deformation and orographic precipitation. The sediment flux from the
2016), providing a mechanistic link between the region’s propagating model domain of 250 × 500 km (i.e., approximately one Himalayan
rock uplift patterns and channels that deviate from the canonical catchment) is estimated to be 30–40 × 106 m3/yr (Fig. 8C), comparable
expectation of generally concave-up rivers (Whipple and Tucker, 1999). with measurements from Himalayan catchments (Dingle et al., 2016;
Our propagating uplift function with the optimal parameter values Lupker et al., 2012). Thus, the modeled total flux from the entire Hi-
(Table 1) results in a smoothed wave-form uplift rate with a maximum of malayan front is about 240–320 × 106 m3/yr if we simply assume that
~2 mm/yr (Fig. S3; Movie S1), which is close to the uplift patterns the size of the Himalayan Range is ~2000 km (i.e., eight times the model
derived from InSAR, leveling, geomorphologic features, and previous width 250 km), much more than that of all SE and E Asia not including
the Himalayas and YTSZ (ca. 60–100 × 106 m3/yr (Clift, 2006)). Our

10
Y. Li et al. Earth and Planetary Science Letters 658 (2025) 119305

Fig. 8. Amounts of uplift and erosion. (A) Illustration of the maximum “no erosion” height h0 , erosion/uplift, and sediment flux in the contexts of propagating uplift
and orographic precipitation. (B) The ratio of erosion to uplift amounts for eight river systems, derived from the modeled topographies with (i.e., the best-fit values of
Kp ) and without (i.e., Kp = 0.00 × 10− 6 m− 0.35/yr0.55) erosion. The upper (right) y-axis shows the modeled “no erosion” height (h0 ) for each river system from west to
east Himalayas and the “no erosion” surface elevations in the Nepal Himalaya modeled by Bollinger et al. (2006). (C) The average elevations of “no erosion” to-
pographies and eroded topographies, and sediment flux, for eight river systems through time. Our model domain is 250 km wide and 500 km long, including half
mountain and half plain, similar to the observed catchment size (Dingle et al., 2016; Lupker et al., 2012) (Fig. 1A).

Fig. 9. Relationships between erosion/uplift ratio and erosion efficiency Kp for eight Himalayan river systems. The circles, joined by a curve, represent the modeled
results with variable Kp values, and the star represents the result with the optimal Kp value, for each river. The circles and stars thickly outlined are the examples
shown in subplots (I)− (V). In each subplot (I)− (V), the surface colored with elevation is the resulting topography with the Kp value in the main figure of Erosion/
Uplift− Kp relationships, the surface colored cyan represents the topography with no erosion (i.e., Kp = 0.00 × 10− 6 m− 0.35/yr0.55), and the blue curve shows the
modeled river profile with the corresponding Kp value.

11
Y. Li et al. Earth and Planetary Science Letters 658 (2025) 119305

modeling approach allows reconstruction of the sediment flux out of process-based support for existing theoretical hypotheses.
Himalayas through time, which has important implications for the
amount of CO2 consumed by silicate weathering and thus climate 4.4. Model limitations
change over geological timescales (Galy et al., 2007).
The supplementary models in which we varied the SPL parameters The model reproduces river profiles, topographic profiles, precipi-
(Fig. S4), the surface shortening rate (Fig. S5), and the superimposed tation profiles, erosion rates, and cooling ages along the eight main
uplift caused by a foreland fold-and-thrust belt (Fig. S6), show results rivers across the Himalayas (Figs. 2 to 5) despite incorporating only
similar to those from our main model, i.e., “no erosion” height h0 (20 tectonic uplift, fluvial processes, and hillslope diffusion. Neither glacial
km), ratio of erosion to uplift amounts (~70%), and sediment flux 30–40 nor rockslide erosion is included in our model because they are limited
× 106 m3/yr. We conclude that, due to the dynamics of propagating to the high peaks (Lavé et al., 2023). 10Be concentrations measured in
uplift and orographic precipitation, the Himalayas should be uplifted to the downstream plain and fan indicate that the erosion rates averaged
~20-km elevation, and be eroded by 60− 70%, which produces a sedi- over each of the upstream Himalayan catchments range from 0.5 to 2.4
ment flux of 30–40 × 106 m3/yr per 250-km width, and that these mm/yr (average ~1.0 mm/yr) (Lupker et al., 2012), similar to our
general results are reasonably insensitive to many details of model setup modeled erosion rates of 0− 2 mm/yr (Fig. 4). Our best-fit model likely
and parameterization. subsumes the erosive effects of glaciation and rockslides into a slightly
higher erosional efficiency than would be the case for a landscape
4.3. Relationship between erosion/uplift and erosion efficiency in the evolving exclusively under fluvial processes. If we were to take glacia-
Himalayas tion and rockslides into account explicitly in the model, it might lead to
slightly greater “no erosion” height h0 , ratio of erosion/uplift amounts,
We study the impact of fluvial erosion efficiency on erosion/uplift and sediment flux.
assuming a constant uplift function in the sensitivity analysis (Fig. 9). A tectonic simplification was made in modeling the north of the
The erosion efficiency Kp encapsulates the resistance of the channel bed South Tibetan detachment (STD), where its activation in the Miocene
to fluvial erosion, is influenced by lithology and other factors (Adams remains poorly constrained (Fig. 1C). However, the total displacement
et al., 2020), and is highly variable in nature (Stock and Montgomery, of the STD (20–65 km) is much less than those of the other thrust faults
1999), ranging from 1 × 10− 7 to 1 × 10− 2 m0.2/yr (assuming m/n=0.4). such as the GCT, THFT, MCT, MBT, and MFT (>100 km) (Fig. 1C;
Our modeling results indicate that changes in erosion efficiency can Table S1). The STD is also much shallower than the MCT and other
significantly alter the relative contributions of tectonics and climate thrust faults (Fig. 1C) (Gao et al., 2016). In addition, the entire Hima-
(Fig. 9). For the Himalayas, erosion/uplift ratios are sensitive to erosion layas should be contractional, although the STD is an extensional
efficiency when the erosion efficiency is between 1 × 10− 7 and 1 × 10− 5 structure (Gao et al., 2016; Yin, 2006). Thus, simplification of the STD in
(m1-3m/yr1-m, m=0.45), but become constant for other values of erosion our uplift function likely did not have a first-order influence on our
efficiency. outcomes.
This idea can be generalized from the Himalayas to orogens globally The organization of river networks is influenced by the other controls
because the erosion efficiency can control the non-dimensional Beau- not included in our simplified models, especially in river reaches flowing
mont number, Bm, determining types of growing orogens (Wolf et al., parallel to, rather than perpendicular to, the strike of the Himalayan
2022; Yuan et al., 2024). We therefore relate the change of material range (e.g., rivers 1, 4, and 6 in Fig. 2A). Olen et al. (2015) argued that
amounts on Earth’s surface to the types of orogens determined by the upper reach of river 6 (i.e., Arun) and its slope-break knickzones
theoretical Bm. For example, high mountains (e.g., the Tian Shan and correspond to the STD, activated at 23− 9 Ma (Fig. 1C). Han et al. (2024)
Altai Mountains) or wide plateaus (e.g., the Tibetan Plateau and Central proposed that, for the Arun River, the lateral-flowing reach in the
Andes), having Bm greater than 0.5 (Wolf et al., 2022), can form with Tethyan Himalaya was captured by the downstream reaches at ~89
low erosion efficiency and small erosion/uplift amounts (Fig. 9). In thousand years ago. However, Adams et al. (2016) suggested that duplex
contrast, under high to very high erosion efficiencies, small and narrow deformation, rather than river capture, produced the similar knickpoints
mountain belts (e.g., Pyrenees, Taiwan, and Southern Alps of New along river profiles in the Bhutan Himalaya, and blockage of renewed
Zealand), having Bm less than 0.4 (Wolf et al., 2022), will form with faults can also cause lateral flow of rivers (Wolf et al., 2022; Yuan et al.,
large erosion/uplift amounts (Fig. 9). This connection between the rock 2024). Despite simplification of the reorganization of Himalayan
erosion efficiency, the relative proportions of uplift and erosion, and the drainages, e.g., caused by the STD and capture events, our results sug-
shape of mountain ranges expands insight into mountain belt evolution. gest that propagating uplift and orographic precipitation can explain the
Under the India-Eurasia collision and the South Asian monsoon (Figs. 1 major features of trans-Himalayan river profiles (Fig. 2).
and S1), the Himalayas are propagating a high-elevation plateau with
erosion/uplift ratios of ~70%. When the erosion efficiency is less than 5. Conclusions
3–4 times current value (i.e., Kp ≤ 1 × 10− 5 m1-3m/yr1-m, m=0.45), the
Himalayas still have a wide plateau (Fig. 9). With much greater values of This study uses coupled tectonic, landscape evolution, and
erosion efficiency, the Himalayas would be a narrow mountain belts and orographic precipitation modeling, constrained by diverse topographic
the ratios of erosion/uplift would approach the limit (>90%), similar to and geochronologic datasets, to test the hypothesis that propagating
the conditions observed in Taiwan and Southern Alps of New Zealand. rock uplift represents the primary control on Himalayan topography.
Our findings quantitatively show how the form and evolution of the Within the limitations inherent to our modeling assumptions, our
Himalayas reflect the interplay between the primary influence of method of constraining a landscape evolution model with readily
propagating rock uplift and the secondary influence of orographic pre- available datasets allows for rough inference of time-varying quantities
cipitation. This result provides new evidence for the hypothesis that the such as paleo-altitude and rock uplift and erosion amounts, which are
Himalayas is a strength-limited orogen (Wolf et al., 2022). Though we difficult to accurately determine using traditional approaches. Our re-
do not incorporate a geodynamic model to assess orogenic strength sults provide a new line of evidence supporting the idea that propagating
controls, our propagating rock uplift condition reflects the crust reach- rock uplift sets the broad topographic template of the Himalayas; we
ing its strength limit because it sets the maximum uplift amount (h0 ) and also suggest a secondary role for orographic precipitation effects.
therefore governs the extent of orogenic growth (Wolf et al., 2022). Our Propagating uplift could drive ~20-km elevations in the Himalayas, but
landscape evolution modeling study leverages diverse topographic and orographically-influenced precipitation eroded 60− 70% of the uplifted
geochronologic data to relate the amount of material at Earth’s surface material, corresponding to a present-day sediment flux of 30–40 × 106
to the relative influence of tectonics and climate, providing m3/yr per 250-km width (i.e., approximately one Himalayan

12
Y. Li et al. Earth and Planetary Science Letters 658 (2025) 119305

catchment). Our study provides an inversion method to estimate topo- Funding acquisition, Formal analysis, Conceptualization. Charles M.
graphic evolution caused by tectonics and climate, providing a process- Shobe: Writing – review & editing, Validation, Supervision, Investiga-
based way to assess how mountainous topography reflects tectonic tion, Formal analysis. Guillaume Dupont-Nivet: Writing – review &
dominance and climatic modulation. Studies using approaches similar to editing, Supervision, Funding acquisition, Formal analysis, Conceptu-
ours could shed new light on the growth of other mountain belts on alization. Kai Cao: Writing – review & editing, Supervision, Investiga-
Earth, and help quantitatively establish links among tectonics, climate, tion, Formal analysis.
mountain belt topography, and biogeochemical cycling.
Declaration of competing interest
Funding sources
The authors declare that they have no known competing financial
X.Y. acknowledges funding from National Natural Science Founda- interests or personal relationships that could have appeared to influence
tion of China (no. 42272261). G.D.-N. acknowledges funding from ANR the work reported in this paper.
grant TIBETOP.
Acknowledgments
CRediT authorship contribution statement
We thank two anonymous reviewers for their very constructive
Yuqiang Li: Writing – review & editing, Writing – original draft, reviewing and Editor Jean-Philippe Avouac for editorial handling. The
Visualization, Validation, Methodology, Investigation, Formal analysis, findings and conclusions in this publication are those of the authors and
Data curation, Conceptualization. Xiaoping Yuan: Writing – review & should not be construed to represent any official USDA or U.S. Gov-
editing, Supervision, Software, Project administration, Methodology, ernment determination or policy.

Appendix A. Explanation of coupled surface process and orographic precipitation model

A.1. Landscape evolution model

We model the geomorphic evolution of the Himalayas using the FastScape landscape evolution model which solves the equation governing the
long-term tectonic uplift, fluvial erosion and sediment deposition (Braun and Willett, 2013; Yuan et al., 2019) as
∫( )
∂h m G ∂h
= U − Kf ̃
p Am S n + U− dA, (A.1)
∂t pA
̃ ∂t
A

where ∂h/∂t is the change rate of topography (m/yr), U is the uplift rate (m/yr), Kf is the fluvial erosion efficiency related to lithology, vegetation, and
referential precipitation rate P0 = 1 m/yr (m1-2m/yr), ̃ p is the ratio of upstream averaged precipitation P to the referential precipitation P0 (dimen-
sionless), A is the upstream drainage area (m2), S is the slope in the steepest-descent drainage direction (dimensionless), m and n are the stream power
law (SPL) exponents, and G is a dimensionless lumped deposition coefficient encapsulating sediment concentration ratio in transport and settling
velocity of sediment (Yuan et al., 2019). It should be noted that, due to the referential precipitation P0 = 1 m/yr,
m m m
p = Kp P0 m ̃
Kf ̃ p = Kp P , (A.2)

where Kp is a new erosional efficiency coefficient that is independent of precipitation, but still encapsulates a number of factors including lithology,
vegetation, and other factors (m1-3m/yr1-m) (Adams et al., 2020). Here, the upstream averaged precipitation P for each cell i is calculated by
/
∑ ∑
P= Pj ΔxΔy ΔxΔy, (A.3)
j=ups(i) j=ups(i)

where Pj is the precipitation rate in each upstream cell j.

Fluvial erosion leads to the formation of hillslopes along river channels. Hillslope processes are commonly represented by a linear diffusion term
(Culling, 1960):
∂h
= KD ∇2 h, (A.4)
∂t

where KD is the hillslope diffusion coefficient (m2/yr).

Concavity m/n in stream power law (SPL) is suggested to be 0.45 for the Himalayas (Fig. S2) (Adams et al., 2020; Olen et al., 2015; Wobus et al.,
2006). Wobus et al. (2006) suggested SPL slope exponent n = 1 in the Siwalik Hills, Nepal, according to compiled data of many basins subject to a
known but variable forcing. Also, deposition coefficient G/̃ p is widely close to one when n = 1, estimated from experimental and natural landscapes
when considering precipitation (Guerit et al., 2019). They, therefore, are used in a set of main models for eight river systems in the Himalayas
(Table 1). In addition, detachment-limited SPL (i.e., G = 0) suggests n = ~2.2 fitting the channel steepness ksnQ and 10Be-derived erosion rates in the
Himalayas (Adams et al., 2020). Thus, we supply a test for the classical stream power law (i.e., m/n = 1/2.2 and G/̃ p = 0). It should be noted that a
larger n corresponding to a smaller G agrees with the classicl stream law (Guerit et al., 2019), so it is simplified but reasonable to test with m/n = 1/2.2
and G/̃ p = 0 (Fig. S4). Finally, the erosion efficiency Kp can uniquely be estimated by simultaneously comparing both river profiles and maximum
topographic profiles (Fig. 6; Table 1) (Li and Yuan, 2024). Cosmogenic and thermochronologic data can give more similar constraints on the modeled
Kp .

13
Y. Li et al. Earth and Planetary Science Letters 658 (2025) 119305

Hillslope diffusion coefficient KD , ranging from 1 × 10− 1 to 1 × 10− 4 m2/yr around the world (Richardson et al., 2019), has no substantial in-
fluences on modeled results according to Li and Yuan (2024). Thus, KD is set to be 1 × 10− 2 m2/yr throughout this study for simplicity.
A.2. Boundary conditions

The south boundary of the model domain is set as the base level, i.e., an open boundary where water and sediment flux can leave the system
(Fig. 1D). The other three boundaries are all closed where no flux is allowed to leave. Further, model domain, for each catchment, is determined from
Google Earth. At present day (i.e., 0 Ma), the N-S length should be 500 km, representing the distance from the Yarlung–Tsangpo suture zone (YTSZ) to
a part of the Indus-Ganga plain, and the W-E width is about 250 km, based on the observed catchment width in the Himalayas on average (Fig. 1, B and
D). The upstream parts of the Himalayan rivers are limited by the northern Yarlung–Tsangpo catchments, so the comparisons in topography, river
length, and drainage divide location on the upstream will be slightly influenced.
We simplify the geological background, assuming a uniform surface shortening along the Himalayan arc because collision occurs anywhere
(Fig. 1). Current Global Positioning System (GPS) data shows that, relative to a stable Eurasia, the southern foot of the Himalayas moves northward at
~33− 36 mm/yr, whereas the northern margin, i.e., the YTSZ, moves northwards at ~22− 25 mm/yr (Liang et al., 2013). In geological history, the
average long-term Indo-Himalayan shortening rates determined by various methods are 14±4 mm/yr (Powers et al., 1998). Balanced cross-sections
and river terraces indicate that arc-perpendicular shortening velocity along the Himalayas is about 10 mm/yr on average from south to north since the
Holocene (Paul Burgess et al., 2012). Meanwhile, the plate reconstructions have suggested that India’s convergence with Eurasia slows and stabilizes
since ~20 Ma (Molnar and Stock, 2009; van Hinsbergen et al., 2012), so it can be assumed that the development of the Himalayas was the same at
steady state with the constant crustal shortening. Thus, we set the N-S shortening rate perpendicular to the Himalayan arc to 10 mm/yr on average for
each river system since 23 Ma, i.e., a shortening distance of 230 km. In addition, the shortening of topography before 23 Ma is not taken into account in
this study, because the mountain building and exhumation are relatively weak (Clift et al., 2008; Ding et al., 2022) before “hard” collision (Molnar and
Stock, 2009; van Hinsbergen et al., 2012). Thus, the initial N-S length of our model is 730 = 500 + 230 km (Fig. 1D). In addition, considering un-
certainty on shortening rate which may be larger in the Himalayas, we supply and compare different models with shortening rate of 10 mm/yr (i.e.,
our main model), 20 mm/yr, and 30 mm/yr (Fig. S5).

A.3. Propagating uplift function

Inspired by the topographic features around the Himalaya-Tibet-Hengduan region (Clark and Royden, 2000; Yuan et al., 2021), we assume a
simplified curved topography only with uplift, termed as the hypothetical “no erosion” topography. According to the outward-growth of the
Himalayas described in the main text, this imaged “no erosion” topography also propagates southward in the Himalayas during the fast rise stage (i.e.,
since the Miocene). Here, the “no erosion” topography should be a final result of all the underground processes, such as faulting, folding, channel
flowing, and isostatic rebound, opposite to climatic effects. Thus, the propagating uplift rate derived below is a lumped bedrock uplift rate resulting
from multiple tectonic activities.
The elevations hf of the “no erosion” topography at distances x south of the “no erosion” topography margin, x0 , are well represented by a logistic
function as (Yuan et al., 2021):
h0
hf (x) = , (A.5)
1 + e(x− x0 )/W

where h0 is the maximum elevation of the hypothetical “no erosion” topography, and W is the characteristic width (of the propagating uplift wave
described below). The “no erosion” topography resulting from the outward-growth of the Himalayas also agrees with the high-elevation, low-relief
landscape via duplex deformation in the Bhutan Himalaya (Adams et al., 2016) and the distributed basins in the Himalayan hinterland since the
Miocene (Ibarra et al., 2023). Assuming the topography margin propagating southward in time t at velocity v0 , the propagating distance x0 is
x0 = v0 t. (A.6)
Combining the Eqs. (A.5) and (A.6), the uplift rate U can be obtained from the derivative of the “no erosion” topographic profile with respect to
time, given that the morphology of this profile is dominated by tectonic uplift without climatic effects (i.e., fluvial erosion and deposition):
h0 v0 e(x− x0 )/W
U= 2
. (A.7)
W[1 + e(x− x0 )/W ]
From the high precipitation-weighted variant of the normalized channel steepness index ksnQ (a proxy for uplift rate) section (Fig. 1B), we can
approximate the present propagating margin and the corresponding distance of uplift wave during the fast rise stage (i.e., since 23 Ma) (Table 1). For
river 5 with its catchment as an example, the propagating distance of uplift wave is ~160 km at present day (Fig. 1B), but the historical propagating
distance should be (500 + 230) × (160/500) = 233.6 km in 23 Myr taking the surface shortening into account. Thus, the propagating velocity of uplift
wave v0 is 233.6 km/23 Myr ≈ 10 mm/yr. Finally, we estimate the optimal values of h0 and W in Eq. (A.7) by comparing modeled and observed river
profiles and swath elevation profiles along the rivers (Fig. 6; Table 1) (Li and Yuan, 2024). Cosmogenic and thermochronologic data can give more
similar constraints on the modeled h0 .
According to the plate reconstruction (van Hinsbergen et al., 2012), palaeo-altimetric investigation (Ding et al., 2022), and tectonic-surface process
modeling (Wolf et al., 2022), the rise of the Himalayan topography involves basically two stages, i.e., one fast propagating uplift began since Neogene
(i.e., from 23 Ma to 0 Ma, as the above), and another relatively slow uplift occurred in Paleogene (i.e., from ~50 Ma to 23 Ma) (Fig. 1D). We do not
take into account the detailed thrusting processes occurred around the northern boundary YTSZ (including GT, GCT, and more faults) before Neogene
(23 Ma) (Murphy and Yin, 2003), and simplify the proto-Himalayan uplift with limited suggestions of lower elevations of 2− 3 km (Ding et al., 2022;
Molnar and Stock, 2009). Not taking into account climatic effects, the “no erosion” topography at 23 Ma can be calculated based on Eq. (A.5) with an
initial propagating distance x0 = 0 m. Thus, the uplift rate in the slow rise stage (i.e., from ~50 Ma to 23 Ma) is simply derived from the “no erosion”
topography divided by the period 27 Myr. Given the suggested low palaeo-topography (Ding et al., 2017; Molnar and Stock, 2009) and modest erosion
(Clift, 2006; Clift et al., 2008) before 23 Ma, the simplifying assumption of slow uplift driven by thrusting in the YTSZ between 50 and 23 Ma should
have only a minor influence on the total uplift and erosion amounts accumulated from 50 to 0 Ma.

14
Y. Li et al. Earth and Planetary Science Letters 658 (2025) 119305

Although the uplift functions for before and after 23 Ma are simplified, our model coupling uplift and precipitation yields the palaeo-elevations
close to the measured palaeo-elevations in the Qiabulin area and Liuqu Basin near the YTSZ (Ding et al., 2022; Ibarra et al., 2023), i.e., ~2 km before
23 Ma and 5− 6 km after 23 Ma (Fig. S3).

A.4. Orographic precipitation model

The main South Asian monsoon transports a large amount of moisture to the Himalayas, in the forms of vapor and cloud water at a given wind
velocity. Each component undergoes advection, dispersion, and interaction (i.e., condensation and re-evaporation due to change of topography)
processes, keeping in respective moisture balance. The cloud water will finally transfer to be orographic precipitation. In addition, a fraction of the
precipitation will evaporate immediately. The above processes for orographic precipitation are described in the linear feedback precipitation model
LFPM (Hergarten and Robl, 2022), as the following:
∂Fv ∂2 Fv Fv − βFc Fc
− + Ld 2 − + ϵ = 0, (A.8)
∂x ∂y Lc Lf

∂Fc ∂2 Fc Fv − βFc Fc
− + Ld 2 + − = 0, (A.9)
∂x ∂y Lc Lf

H
ϵ = ϵ0 e (A.10)
−
H0 ,

where the subscript v/c means the relation for vapor (v) or cloud water (c), F is the advective flux per unit width (m2/yr), Ld is the dispersion length
(m), Lc is the length scale of condensation (m), Lf is the length scale of fallout (m), and ϵ and ϵ0 are evapotranspiration rates at any altitude H and a
reference altitude H0 respectively (dimensionless). Note that H also represents the topographic elevation h in Eq. (A.1), but in the unit of grid spacing
(i.e., 1 km here), and thus H0 is usually 1 unit grid spacing for convenience. At the same time, the nondimensional coefficient β that determines the
dynamic equilibrium in source interaction is
H
β = β0 e (A.11)
−
H0 ,

( )( )
Lc Ll
β0 = 1− − 1 , (A.12)
Ll Lf

where Ll is the length scale of the decay (m). Finally, the effective precipitation that contributes to runoff is
Fc
P = (1 − ϵ) + Pb , (A.13)
Lf

where Pb is a background precipitation rate including other secondary rainfall contributions by other sources (e.g., local lakes), in addition to the
South Asian monsoon moisture (m/yr).
Note that (1) Lc and Lf are usually assumed to be equal suggested by Smith and Barstad (2004) and Hergarten and Robl (2022); (2) the dispersion
length Ld is set to be 25 km in all the modeling based on the modern Himalayas (Hergarten and Robl, 2022), and its change does not obviously in-
fluence the spatially-continuous mountain belt like the Himalayas (Hergarten and Robl, 2022; Li and Yuan, 2024); (3) the reference evapotranspi-
ration rate ϵ0 is set to be 50% in all the modeling, which can produce the total evapotranspiration amount close to the observation (Li and Yuan, 2024),
because our evapotranspiration mechanism is excessively simplified and evapotranspiration amount is much less than the effective precipitation rate
in the Himalayas (Bookhagen and Burbank, 2010; Li and Yuan, 2024); (4) the background precipitation rate Pb is set to be 0.1 m/yr in all the modeling,
which is close to observed precipitation rate in the rear of Himalayas and also can produce suitable modeling river profile for the Himalayas (Li and
Yuan, 2024); (5) The remaining precipitation parameters (i.e., Fin , Lc = Lf , and Ll ) can uniquely be estimated by comparing the modeled and observed
features of precipitation profiles (Table 1) (Li and Yuan, 2024).
The monsoon precipitation is further related to the atmosphere dynamic, especially the moisture and wind direction for our model. The South
Asian monsoon significantly intensified during the Middle Miocene (17− 12 Ma), while keep advancing northwards from the tropic since the middle
Eocene (50–45 Ma) (Sarr et al., 2022; Zuo et al., 2024). Around the Himalayas, the vertical integrated moisture transport is estimated to be 100− 200
kg/m2/s at the late Miocene, while the difference of vertical integrated moisture transport between early and late Miocene is about 50− 100 kg/m2/s
due to the uplift of the East African and Middle Eastern topography (Sarr et al., 2022). In this study, we preliminarily assume that the total incoming
influx of monsoon moisture Fin (km2/yr), including vapor and cloud water, is temporally constant in each modeling. The more comprehensive model
coupling our landscape evolution and ocean-atmosphere dynamics is needed to improve our inversion results.

Appendix B. River morphology analysis

The digital elevation model (DEM), accessed from the National Aeronautics and Space Administration (NASA) Shuttle Radar Topography Mission
(SRTM) 3 arc (~90 m) second product, is resampled to 200-m resolution. The mean annual precipitation (P) of 1998 to 2007 is calculated from the
Tropical Rainfall Measurement Mission (TRMM) product (Bookhagen and Burbank, 2010).
In this study, we extracted the river networks with drainage area exceeding 10 km2 in the Himalayas using the MATLAB-based software Top-
oToolbox (Schwanghart and Scherler, 2014). Next, we compute the precipitation-weighted variant of the normalized channel steepness index (i.e.,
ksnQ ) for the Himalayan rivers (Fig. 1B), according to (Adams et al., 2020):
ksnQ = SQm/n , (B.1)

15
Y. Li et al. Earth and Planetary Science Letters 658 (2025) 119305

where Q is discharge (= AP). In the quasi-equilibrium landscapes with respect to current uplift and climate conditions, here the Himalayas (Adams
et al., 2020), ksnQ can reflect the current uplift information with small uncertainties caused by nonuniform erosional efficiency (i.e., Kp ).
To compare the modeled and observed longitudinal profiles, including the river channels and relief, we also apply the χ -transform to the trunks in
observations and modeling (Fig. 2), transforming the horizontal coordinate of longitudinal profile from distance to upstream integral of drainage area
χ , according to (Perron and Royden, 2013):
∫x ( )m/n
A0
χ (x) = dx, (B.2)
A(x)
xb

from the base level xb to the channel reach x, where A0 (= 1 m2) is a reference drainage area. Using TopoToolbox (Schwanghart and Scherler, 2014),
we also extract the river profiles, and mean and maximum elevation profiles within 50 km-radius swaths along rivers, based on the observed trunk
catchment sizes, for the eight main trans-Himalayan trunks from the DEM. We slightly smooth all the profiles (Schwanghart and Scherler, 2014) to
remove unimportant noise.
Because (1) the drainage area is difficult to be accurately found for the reaches in the plain, and (2) the change of profile located in the plain is too
small relative to that in the mountain, we focus on comparing the χ -elevation profiles located in mountains. Here we rewrite the Eq. (B.2) as
∫x ( )m/n ∫ ( )m/n ∫ ( )m/n
A0 A0 A0
dx = dx + dx, (B.3)
A(x) A(x) A(x)
xb plain mountain

where the first term on the right-hand side of Eq. (B.3) represents the integration from the base level xb to the separation place of plain and mountain,
whereas the second term is from the separation place to a point x.

Supplementary materials

Supplementary material associated with this article can be found, in the online version, at doi:10.1016/j.epsl.2025.119305.

Data availability Clark, M.K., Royden, L.H., 2000. Topographic ooze: Building the eastern margin of Tibet
by lower crustal flow. Geology 28 (8), 703–706. https://doi.org/10.1130/0091-
7613(2000)28<703:TOBTEM>2.0.CO;2.
All data are cited in the article and supplementary material. Nu- Clift, P.D., 2006. Controls on the erosion of Cenozoic Asia and the flux of clastic sediment
merical models are computed with published methods and codes. The to the ocean. Earth Planet. Sci. Lett. 241 (3), 571–580. https://doi.org/10.1016/j.
epsl.2005.11.028.
software to model landscape evolution is available on https://fastscape.
Clift, P.D., Hodges, K.V., Heslop, D., Hannigan, R., Van Long, H., Calves, G., 2008.
org/fastscapelib-fortran/ (Braun and Willett, 2013; Yuan et al., 2019). Correlation of Himalayan exhumation rates and Asian monsoon intensity. Nat.
The software to model orographic precipitation is available on Geosci. 1 (12), 875–880. https://doi.org/10.1038/ngeo351.
Codilean, A.T., Munack, H., Cohen, T.J., Saktura, W.M., Gray, A., Mudd, S.M., 2018.
http://jura.geologie.uni-freiburg.de/openlem/lfpm.php (Hergarten and
OCTOPUS: an open cosmogenic isotope and luminescence database. Earth Syst. Sci.
Robl, 2022). Data 10 (4), 2123–2139. https://doi.org/10.5194/essd-10-2123-2018.
Coutand, I., Whipp Jr, D.M., Grujic, D., Bernet, M., Fellin, M.G., Bookhagen, B.,
References Landry, K.R., Ghalley, S.K., Duncan, C., 2014. Geometry and kinematics of the Main
Himalayan Thrust and Neogene crustal exhumation in the Bhutanese Himalaya
derived from inversion of multithermochronologic data. J. Geophys. Res. Solid Earth
Adams, B.A., Whipple, K.X., Forte, A.M., Heimsath, A.M., Hodges, K.V., 2020. Climate 119 (2), 1446–1481. https://doi.org/10.1002/2013JB010891.
controls on erosion in tectonically active landscapes. Sci. Adv. 6 (42), eaaz3166. Culling, W.E.H., 1960. Analytical Theory of Erosion. J. Geol. 68 (3), 336–344. https://
https://doi.org/10.1126/sciadv.aaz3166. doi.org/10.1086/626663.
Adams, B.A., Whipple, K.X., Hodges, K.V., Heimsath, A.M., 2016. In situ development of DeCelles, P.G., Robinson, D.M., Quade, J., Ojha, T.P., Garzione, C.N., Copeland, P.,
high-elevation, low-relief landscapes via duplex deformation in the Eastern Upreti, B.N., 2001. Stratigraphy, structure, and tectonic evolution of the Himalayan
Himalayan hinterland. Bhutan. J. Geophys. Res. Earth Surf. 121 (2), 294–319. fold-thrust belt in western Nepal. Tectonics 20 (4), 487–509. https://doi.org/
https://doi.org/10.1002/2015JF003508. 10.1029/2000TC001226.
Beaumont, C., Jamieson, R.A., Nguyen, M.H., Lee, B., 2001. Himalayan tectonics Dielforder, A., Hetzel, R., Oncken, O., 2020. Megathrust shear force controls mountain
explained by extrusion of a low-viscosity crustal channel coupled to focused surface height at convergent plate margins. Nature 582 (7811), 225–229. https://doi.org/
denudation. Nature 414 (6865), 738–742. https://doi.org/10.1038/414738a. 10.1038/s41586-020-2340-7.
Bollinger, L., Henry, P., Avouac, J.P., 2006. Mountain building in the Nepal Himalaya: Ding, L., Kapp, P., Cai, F.L., Garzione, C.N., Xiong, Z.Y., Wang, H.Q., Wang, C., 2022.
Thermal and kinematic model. Earth Planet. Sci. Lett. 244 (1), 58–71. https://doi. Timing and mechanisms of Tibetan Plateau uplift. Nat. Rev. Earth Environ. 3 (10),
org/10.1016/j.epsl.2006.01.045. 652–667. https://doi.org/10.1038/s43017-022-00318-4.
Bookhagen, B., Burbank, D.W., 2010. Toward a complete Himalayan hydrological Ding, L., Spicer, R.A., Yang, J., Xu, Q., Cai, F., Li, S., Lai, Q., Wang, H., Spicer, T.E.V.,
budget: Spatiotemporal distribution of snowmelt and rainfall and their impact on Yue, Y., Shukla, A., Srivastava, G., Khan, M.A., Bera, S., Mehrotra, R., 2017.
river discharge. J. Geophys. Res. Earth Surf. 115 (F3). https://doi.org/10.1029/ Quantifying the rise of the Himalaya orogen and implications for the South Asian
2009JF001426. monsoon. Geology 45 (3), 215–218. https://doi.org/10.1130/G38583.1.
Braun, J., van der Beek, P., Batt, G., 2006. Quantitative Thermochronology: Numerical Dingle, E.H., Sinclair, H.D., Attal, M., Milodowski, D.T., Singh, V., 2016. Subsidence
Methods for the Interpretation of Thermochronological Data. Cambridge University control on river morphology and grain size in the Ganga Plain. Am. J. Sci. 316 (8),
Press, Cambridge. 778. https://doi.org/10.2475/08.2016.03.
Braun, J., Willett, S.D., 2013. A very efficient O(n), implicit and parallel method to solve Egholm, D.L., Nielsen, S.B., Pedersen, V.K., Lesemann, J.E., 2009. Glacial effects limiting
the stream power equation governing fluvial incision and landscape evolution. mountain height. Nature 460 (7257), 884–887. https://doi.org/10.1038/
Geomorphology 180-181, 170–179. https://doi.org/10.1016/j. nature08263.
geomorph.2012.10.008. Eizenhöfer, P.R., McQuarrie, N., Shelef, E., Ehlers, T.A., 2019. Landscape Response to
Brozović, N., Burbank, D.W., Meigs, A.J., 1997. Climatic Limits on Landscape Lateral Advection in Convergent Orogens Over Geologic Time Scales. J. Geophys.
Development in the Northwestern Himalaya. Science 276 (5312), 571–574. https:// Res. Earth Surf. 124 (8), 2056–2078. https://doi.org/10.1029/2019JF005100.
doi.org/10.1126/science.276.5312.571. Feldl, N., Bilham, R., 2006. Great Himalayan earthquakes and the Tibetan plateau.
Burbank, D.W., Blythe, A.E., Putkonen, J., Pratt-Sitaula, B., Gabet, E., Oskin, M., Nature 444 (7116), 165–170. https://doi.org/10.1038/nature05199.
Barros, A., Ojha, T.P., 2003. Decoupling of erosion and precipitation in the Galy, V., France-Lanord, C., Beyssac, O., Faure, P., Kudrass, H., Palhol, F., 2007. Efficient
Himalayas. Nature 426 (6967), 652–655. https://doi.org/10.1038/nature02187. organic carbon burial in the Bengal fan sustained by the Himalayan erosional
system. Nature 450 (7168), 407–410. https://doi.org/10.1038/nature06273.

16
Y. Li et al. Earth and Planetary Science Letters 658 (2025) 119305

Gao, R., Lu, Z., Klemperer, S.L., Wang, H., Dong, S., Li, W., Li, H., 2016. Crustal-scale Powers, P.M., Lillie, R.J., Yeats, R.S., 1998. Structure and shortening of the Kangra and
duplexing beneath the Yarlung Zangbo suture in the western Himalaya. Nat. Geosci. Dehra Dun reentrants, Sub-Himalaya, India. Geol. Soc. Am. Bull. 110 (8),
9 (7), 555–560. https://doi.org/10.1038/ngeo2730. 1010–1027. https://doi.org/10.1130/0016-7606(1998)110<1010:SASOTK>2.3.
Godard, V., Bourlès, D.L., Spinabella, F., Burbank, D.W., Bookhagen, B., Fisher, G.B., CO;2.
Moulin, A., Léanni, L., 2014. Dominance of tectonics over climate in Himalayan Richardson, P.W., Perron, J.T., Schurr, N.D., 2019. Influences of climate and life on
denudation. Geology 42 (3), 243–246. https://doi.org/10.1130/G35342.1. hillslope sediment transport. Geology 47 (5), 423–426. https://doi.org/10.1130/
Godard, V., Cattin, R., Lavé, J., 2009. Erosional control on the dynamics of low- G45305.1.
convergence rate continental plateau margins. Geophys. J. Int. 179 (2), 763–777. Sarr, A.-C., Donnadieu, Y., Bolton, C.T., Ladant, J.-B., Licht, A., Fluteau, F., Laugié, M.,
https://doi.org/10.1111/j.1365-246X.2009.04324.x. Tardif, D., Dupont-Nivet, G., 2022. Neogene South Asian monsoon rainfall and wind
Goren, L., Fox, M., Willett, S.D., 2014. Tectonics from fluvial topography using formal histories diverged due to topographic effects. Nat. Geosci. 15 (4), 314–319. https://
linear inversion: Theory and applications to the Inyo Mountains. California. J. doi.org/10.1038/s41561-022-00919-0.
Geophys. Res. Earth Surf. 119 (8), 1651–1681. https://doi.org/10.1002/ Scherler, D., Bookhagen, B., Strecker, M.R., 2014. Tectonic control on 10Be-derived
2014JF003079. erosion rates in the Garhwal Himalaya. India. J. Geophys. Res. Earth Surf. 119 (2),
Guerit, L., Yuan, X.-P., Carretier, S., Bonnet, S., Rohais, S., Braun, J., Rouby, D., 2019. 83–105. https://doi.org/10.1002/2013JF002955.
Fluvial landscape evolution controlled by the sediment deposition coefficient: Schwanghart, W., Scherler, D., 2014. Short Communication: TopoToolbox 2 – MATLAB-
Estimation from experimental and natural landscapes. Geology 47 (9), 853–856. based software for topographic analysis and modeling in Earth surface sciences.
https://doi.org/10.1130/G46356.1. Earth Surf. Dynam. 2 (1), 1–7. https://doi.org/10.5194/esurf-2-1-2014.
Han, X., Dai, J.-G., Smith, A.G.G., Xu, S.-Y., Liu, B.-R., Wang, C.-S., Fox, M., 2024. Recent Sherpa, T.Z.L., DeCelles, P.G., Carrapa, B., Schoenbohm, L.M., Wolpert, J., 2022.
uplift of Chomolungma enhanced by river drainage piracy. Nat. Geosci. https://doi. Bhumichula plateau: A remnant high-elevation low-relief surface in the Himalayan
org/10.1038/s41561-024-01535-w. thrust belt of western Nepal. Geol. Soc. Am. Bull. 135 (7-8), 2121–2140. https://doi.
Hergarten, S., Robl, J., 2022. The linear feedback precipitation model (LFPM 1.0) – a org/10.1130/B36481.1.
simple and efficient model for orographic precipitation in the context of landform Smith, R.B., Barstad, I., 2004. A Linear Theory of Orographic Precipitation. J. Atmos. Sci.
evolution modeling. Geosci. Model Dev. 15 (5), 2063–2084. https://doi.org/ 61 (12), 1377–1391. https://doi.org/10.1175/1520-0469(2004)061<1377:
10.5194/gmd-15-2063-2022. ALTOOP>2.0.CO;2.
Ibarra, D.E., Dai, J., Gao, Y., Lang, X., Duan, P., Gao, Z., Chen, J., Methner, K., Sha, L., Stock, J.D., Montgomery, D.R., 1999. Geologic constraints on bedrock river incision
Tong, H., Han, X., Zhu, D., Li, Y., Tang, J., Cheng, H., Chamberlain, C.P., Wang, C., using the stream power law. J. Geophys. Res. Solid Earth 104 (B3), 4983–4993.
2023. High-elevation Tibetan Plateau before India–Eurasia collision recorded by https://doi.org/10.1029/98JB02139.
triple oxygen isotopes. Nat. Geosci. https://doi.org/10.1038/s41561-023-01243-x. Taylor, M., Forte, A., Laskowski, A., Ding, L., 2021. Active uplift of southern Tibet
Kirby, E., Whipple, K., 2001. Quantifying differential rock-uplift rates via stream profile revealed. GSA Today 31 (8), 4–10. https://doi.org/10.1130/GSATG487A.1.
analysis. Geology 29 (5), 415–418. https://doi.org/10.1130/0091-7613(2001) van der Beek, P., Litty, C., Baudin, M., Mercier, J., Robert, X., Hardwick, E., 2016.
029<0415:QDRURV>2.0.CO;2. Contrasting tectonically driven exhumation and incision patterns, western versus
Lamb, S., 2006. Shear stresses on megathrusts: Implications for mountain building central Nepal Himalaya. Geology 44 (4), 327–330. https://doi.org/10.1130/
behind subduction zones. J. Geophys. Res. Solid Earth 111 (B7). https://doi.org/ G37579.1.
10.1029/2005JB003916. van der Beek, P., Schildgen, T.F., 2023. Short communication: age2exhume – a MATLAB/
Lavé, J., Guérin, C., Valla, P.G., Guillou, V., Rigaudier, T., Benedetti, L., France- Python script to calculate steady-state vertical exhumation rates from
Lanord, C., Gajurel, A.P., Morin, G., Dumoulin, J.P., Moreau, C., Galy, V., 2023. thermochronometric ages and application to the Himalaya. Geochronology 5 (1),
Medieval demise of a Himalayan giant summit induced by mega-landslide. Nature 35–49. https://doi.org/10.5194/gchron-5-35-2023.
619 (7968), 94–101. https://doi.org/10.1038/s41586-023-06040-5. van Hinsbergen, D.J.J., Lippert Peter, C., Dupont-Nivet, G., McQuarrie, N., Doubrovine
Le Roux-Mallouf, R., Godard, V., Cattin, R., Ferry, M., Gyeltshen, J., Ritz, J.-F., Pavel, V., Spakman, W., Torsvik Trond, H., 2012. Greater India Basin hypothesis and
Drupka, D., Guillou, V., Arnold, M., Aumaître, G., Bourlès, D.L., Keddadouche, K., a two-stage Cenozoic collision between India and Asia. Proc. Natl. Acad. Sci. U.S.A.
2015. Evidence for a wide and gently dipping Main Himalayan Thrust in western 109 (20), 7659–7664. https://doi.org/10.1073/pnas.1117262109.
Bhutan. Geophys. Res. Lett. 42 (9), 3257–3265. https://doi.org/10.1002/ Wang, K., He, J., 1999. Mechanics of low-stress forearcs: Nankai and Cascadia.
2015GL063767. J. Geophys. Res. Solid Earth 104 (B7), 15191–15205. https://doi.org/10.1029/
Li, Y., Wang, C., Dai, J., Xu, G., Hou, Y., Li, X., 2015. Propagation of the deformation and 1999JB900103.
growth of the Tibetan–Himalayan orogen: A review. Earth Sci. Rev. 143, 36–61. Whipple, K.X., Kirby, E., Brocklehurst, S.H., 1999. Geomorphic limits to climate-induced
https://doi.org/10.1016/j.earscirev.2015.01.001. increases in topographic relief. Nature 401 (6748), 39–43. https://doi.org/10.1038/
Li, Y., Yuan, X., 2024. Coupled Surface Process and Orographic Precipitation Model for 43375.
the Landscape Evolution of the Himalayas. J. Earth Sci. 35 (3), 1063–1068. https:// Whipple, K.X., Tucker, G.E., 1999. Dynamics of the stream-power river incision model:
doi.org/10.1007/s12583-024-2012-x. Implications for height limits of mountain ranges, landscape response timescales,
Liang, S., Gan, W., Shen, C., Xiao, G., Liu, J., Chen, W., Ding, X., Zhou, D., 2013. Three- and research needs. J. Geophys. Res. Solid Earth 104 (B8), 17661–17674. https://
dimensional velocity field of present-day crustal motion of the Tibetan Plateau doi.org/10.1029/1999JB900120.
derived from GPS measurements. J. Geophys. Res. Solid Earth 118 (10), 5722–5732. Wobus, C., Whipple, K.X., Kirby, E., Snyder, N., Johnson, J., Spyropolou, K., Crosby, B.,
https://doi.org/10.1002/2013JB010503. Sheehan, D., 2006. Tectonics from topography: Procedures, promise, and pitfalls. In:
Lupker, M., Blard, P.-H., Lavé, J., France-Lanord, C., Leanni, L., Puchol, N., Charreau, J., Willett, S.D., Hovius, N., Brandon, M.T., Fisher, D.M. (Eds.), Tectonics, Climate, and
Bourlès, D., 2012. 10Be-derived Himalayan denudation rates and sediment budgets Landscape Evolution Vol. 398. Geological Society of America, pp. 55–74.
in the Ganga basin. Earth Planet. Sci. Lett. 333-334, 146–156. https://doi.org/ Wolf, S.G., Huismans, R.S., Braun, J., Yuan, X., 2022. Topography of mountain belts
10.1016/j.epsl.2012.04.020. controlled by rheology and surface processes. Nature. https://doi.org/10.1038/
McQuarrie, N., Ehlers, T.A., 2017. Techniques for understanding fold-and-thrust belt s41586-022-04700-6.
kinematics and thermal evolution. In: Law, R.D., Thigpen, J.R., Merschat, A.J., Wolf, S.G., Huismans, R.S., Muñoz, J.-A., Curry, M.E., van der Beek, P., 2021. Growth of
Stowell, H.H. (Eds.), Linkages and Feedbacks in Orogenic Systems Vol. 213. Collisional Orogens From Small and Cold to Large and Hot—Inferences From
Geological Society of America. Geodynamic Models. J. Geophys. Res. Solid Earth 126 (2), e2020JB021168. https://
Molnar, P., England, P., 1990. Late Cenozoic uplift of mountain ranges and global doi.org/10.1029/2020JB021168.
climate change: chicken or egg? Nature 346 (6279), 29–34. https://doi.org/ Yin, A., 2006. Cenozoic tectonic evolution of the Himalayan orogen as constrained by
10.1038/346029a0. along-strike variation of structural geometry, exhumation history, and foreland
Molnar, P., Stock, J.M., 2009. Slowing of India’s convergence with Eurasia since 20 Ma sedimentation. Earth Sci. Rev. 76 (1), 1–131. https://doi.org/10.1016/j.
and its implications for Tibetan mantle dynamics. Tectonics 28 (3). https://doi.org/ earscirev.2005.05.004.
10.1029/2008TC002271. Yuan, X., Li, Y., Brune, S., Li, K., Pons, M., Wolf, S.G., 2024. Coordination between
Murphy, M.A., Yin, A., 2003. Structural evolution and sequence of thrusting in the deformation, precipitation, and erosion during orogenic growth. Nat. Commun. 15
Tethyan fold-thrust belt and Indus-Yalu suture zone, southwest Tibet. Geol. Soc. Am. (1), 10362. https://doi.org/10.1038/s41467-024-54690-4.
Bull. 115 (1), 21–34. https://doi.org/10.1130/0016-7606(2003)115<0021: Yuan, X.P., Braun, J., Guerit, L., Rouby, D., Cordonnier, G., 2019. A New Efficient
SEASOT>2.0.CO;2. Method to Solve the Stream Power Law Model Taking Into Account Sediment
Olen, S.M., Bookhagen, B., Hoffmann, B., Sachse, D., Adhikari, D.P., Strecker, M.R., Deposition. J. Geophys. Res. Earth Surf. 124 (6), 1346–1365. https://doi.org/
2015. Understanding erosion rates in the Himalayan orogen: A case study from the 10.1029/2018JF004867.
Arun Valley. J. Geophys. Res. Earth Surf. 120 (10), 2080–2102. https://doi.org/ Yuan, X.P., Huppert, K.L., Braun, J., Shen, X., Liu-Zeng, J., Guerit, L., Wolf, S.G.,
10.1002/2014JF003410. Zhang, J.F., Jolivet, M., 2021. Propagating uplift controls on high-elevation, low-
Paul Burgess, W., Yin, A., Dubey, C.S., Shen, Z.-K., Kelty, T.K., 2012. Holocene relief landscape formation in the southeast Tibetan Plateau. Geology 50 (1), 60–65.
shortening across the Main Frontal Thrust zone in the eastern Himalaya. Earth https://doi.org/10.1130/G49022.1.
Planet. Sci. Lett. 357-358, 152–167. https://doi.org/10.1016/j.epsl.2012.09.040. Zuo, M., Sun, Y., Zhao, Y., Ramstein, G., Ding, L., Zhou, T., 2024. South Asian summer
Perron, J.T., Royden, L., 2013. An integral approach to bedrock river profile analysis. monsoon enhanced by the uplift of the Iranian Plateau in Middle Miocene. Clim. Past
Earth Surf. Process. Landf. 38 (6), 570–576. https://doi.org/10.1002/esp.3302. 20 (8), 1817–1836. https://doi.org/10.5194/cp-20-1817-2024.

17