6.3 Relationships to Climate Change and Tectonics
The progressive increase in the relative flux from the Himalaya since the Middle Miocene represents the progressive unroofing of these units. Structural reconstructions of the Western Himalaya predict that prior to 5.4 Ma the Greater and Lesser Himalaya were not exposed [Webb , 2013] implying that the Himalayan contribution was derived entirely from the Tethyan Himalaya during the Miocene. As we are not able to distinguish between Tethyan and Greater Himalaya derived sediment we focused on the first appearance of significant amounts of 1500–2300 Ma, Inner Lesser Himalayan detritus starting at 1.56 Ma. Previous studies considered these ranges to have been exposed somewhat before 1.6 Ma. Study of the Siwalik Group in the area of the Beas River Valley indicated an initial exposure of these units around 9 Ma and significant exposure by 6 Ma based on Nd isotope data [Najman et al. , 2009]. Our data support the findings of Clift et al . [2019b] that this exposure may only reflect the local situation in the paleo-Beas River area, but that widespread regional exposure of the Inner Lesser Himalayan units comes somewhat later. While Clift et al . [2019b]favored increased Inner Lesser Himalaya erosion starting at 1.9 Ma our new zircon data imply that 1.56 Ma is a more accurate age for this this transition.
Our result also contrasts with the suggestion by Myrow et al.[2015] that the Inner Lesser Himalaya were widely exposed and eroding by 16 Ma. Although we cannot exclude this from happening further east in the Ganges Basin our data do not support this over a wide area of the western Himalaya until much later.
The timing of Lesser Himalayan unroofing may reflect the development of the thrust duplex, which characterizes the structure of the Lesser Himalaya in this area [Huyghe et al. , 2001; Webb , 2013]. Integrated metamorphic and geochronologic data indicate rapid cooling of the Inner Lesser Himalaya before 6 Ma, following peak metamorphism around 10 Ma [Caddick et al. , 2007; Thiede et al. , 2009]}. We note that rapid cooling does not however require synchronous unroofing. The first major flux of Himalayan zircons to the submarine fan is dated at 7.99 to 7.78 Ma, although widespread Himalayan unroofing may not have started until 5.72 Ma, followed by Inner Lesser Himalayan unroofing starting around 1.56 Ma. This timing is younger than reconstructed by Colleps et al. [2018] who favor exposure of the Outer Lesser Himalaya after 16 Ma and of the Inner Lesser Himalaya after 11 Ma, although that study was again located in an area father east, within the wetter Ganges catchment, and need not apply to the drier Indus basin. A more erosive climate further east might favor earlier unroofing in that area. The erosion data support the concept of significant along strike diachroneity of unroofing.
Uplift of the Lesser Himalayan Duplex would have created a topographic barrier, susceptible to erosion as monsoon rains were focused along this topographic front. The increasing Himalayan character of the total zircon input comes at a time when the summer monsoon rains were generally weakening after ~8 Ma [Dettman et al. , 2001], or after 7.7 Ma based on new environmental data from Site U1456 [Clift et al. , 2019a](Fig. 13). Moisture delivery to this area from the winter westerlies has also been shown to have reduced around 7 Ma [Vögeli et al. , 2017]. In the recent geologic past, since the LGM, strong Himalayan rather than Karakoram erosion has occurred when the summer monsoon was strong, during interglacial times and not when it was weak during glacial times [Clift et al. , 2008a]. The increase in Himalayan erosion over longer periods of time, correlating with the weakening monsoon, is the opposite of this shorter-term trend. It is possible that solid Earth tectonic forces, rather than climate, have dominated the long-term evolution of erosion, although the temporal correlation of provenance and aridity is suggestive of a climatic control.
All of the samples show the presence of very young zircons (<25 Ma) that possibly correlate with bedrock dates from Nanga Parbat, although these are never very numerous. It is also possible that some of these young ages may in fact be derived from erosion of fast exhuming rocks in parts of the southern Karakoram [Wallis et al. , 2014]. However, even if that this material was derived from Nanga Parbat, the low abundance of such zircon grains in the Laxmi Basin sediments would suggest that this massif was not generating very high proportions of sediment in the trunk Indus river, unlike the situation in the eastern syntaxis [Garzanti et al. , 2004; Stewart et al. , 2008]. This is consistent with the U-Pb zircon ages in the modern Indus River downstream of Nanga Parbat [Alizai et al. , 2011] that show neither many <25 Ma zircons or older 1500–2300 Ma grains that would be associated with less deeply buried rocks but with the Lesser Himalayan affiliation typically made with Nanga Parbat [Whittington et al. , 1999].
We compare our detrital zircon budget with that of the Nd budget published by Clift et al. [2019b]. Translating zircon budgets into rock erosion budgets is not easy due to bedrock zircon fertility variations. However, whole-rock geochemical analysis of Alizai et al. [2012] suggested that on average the eastern, Himalaya-draining tributaries are around 2.2 times more fertile in zircon than the trunk Indus. If we simply use the source percentages from the zircon unmixing calculation described above and the average εNd values for these different units then it is possible to predict the average composition of the bulk sediment through time. We use an εNd value of -14.6 for the Greater and Tethyan Himalaya, -21.7 for the Lesser Himalaya, -9.3 for the Karakoram, -20 for Nanga Parbat and +5.1 for Kohistan and the Transhimalaya based on synthesis of the bedrock data, but especially the composition of river sediments that are derived from wide areas of these ranges [Clift et al. , 2002b]. Transhimalaya Nd data are from Rolland et al.[2002], Singh et al. [2002], and Khan et al.[1997]. Greater and Lesser Himalayan data are from Ahmad et al. [2000], Deniel et al. [1987], Inger et al.[1993] and Parrish and Hodges [1996]. Karakoram data are from Crawford and Searle [1992] and Schärer et al. [1990].
The results of this estimate are shown next to the smoothed long-term Nd isotope evolution from bulk sediment analysis [Clift et al. , 2019b] (Fig. 13). We account for the ±1 εNduncertainty value estimated from the Indus Quaternary [Jonell et al. , 2018]. We note that before 6 Ma the estimates overlap with the bulk sediment record that was derived from muddy lithologies, suggesting similar sources. After this time both the estimated and measured εNd values became more negative. However, the predicted Nd isotope compositions are always more negative than those measured from the bulk sediment and this implies an over estimation in the flux from isotopically negative sources, i.e. the Himalaya, using the zircon method. This is consistent with the geochemical data indicating that the Himalaya are more abundant in zircon than the Karakoram, but have similar concentrations in Nd [Alizai et al. , 2011]. As a result, our zircon budget (Fig. 13) represents an overestimate of the influence of the Himalaya compared the Karakoram through time in terms of total rock eroded. Nonetheless, the overall trends in the two data sets are consistent and the reconstruction of increasing Himalayan erosion since the 5.72 Ma may be considered robust.
7. Conclusions
Sandy and silty sediments recovered from the Laxmi Basin in the Eastern Arabian Sea provide a relatively continuous erosional record derived from the Indus River and spanning the last 15.5 m.y. In this study samples were taken from IODP Sites U1456 and U1457 for geochemical and geochronological analyses. Detrital zircon grains were dated by U-Pb methods to determine their provenance. The sediments themselves are defined as wackes and are relatively immature in composition, with bulk sediment characters similar to those found in the Quaternary delta of the Indus and in its submarine Canyon. They are readily distinguishable from sediments on the Western Indian Shelf, confirming their derivation from the Indus River and not the peninsula with one exception. The sediments are mostly of silty sand to silt size, with only a few being classified as fine sand. Although the sediments are relatively depleted in Ca, Na and P relative to the upper continental crust this reflects chemical weathering during transport and does not affect the provenance analysis conducted here.
Detrital zircon U-Pb ages fall into a number of categories which can be correlated with bedrock sources in the Himalaya. The ubiquitous presence of zircon grains younger than 200 Ma requires the sediments to be the erosional products of the Himalaya/Karakoram and not peninsular India. The progressive increase in zircon grains dating at 350–1250 Ma, as well as 1500–2300 Ma, indicates that the erosional flux from the Himalaya increased through the studied time interval. Almost all the samples contain grains that could be derived from the Karakoram or from Kohistan, and there is an appearance of very young zircon grains, younger than 25 Ma, that is especially marked since 3.17 Ma. Such young zircon grains may be from Nanga Parbat or parts of the eastern Karakoram. Statistical analysis shows that there are a number of groupings and an increase in Himalayan erosion through time. High flux from the Himalaya was noted at 7.99–7.78 Ma and starting between 7.0 and 5.87 Ma. Since 1.32 Ma the sediments are similar to the modern Indus River, but not like the glacial-era river, which has more similarities with the Miocene Laxmi Basin samples and with enhanced erosion in the Karakoram. Detrital zircon population unmixing techniques allow us to objectively confirm the progressive increase of Himalayan erosion relative to the Karakoram, and the sharp jump in erosion from the Inner Lesser Himalayas starting at 1.56 Ma. This is somewhat younger than the anticipated unroofing of these ranges derived from earlier foreland studies, although much of the earlier data comes from further east in the Ganges catchment. The shift to more Himalayan erosion through time occurs as the monsoon climate weakened, as well as when the Lesser Himalayan Duplex formed. This suggests that the changing patterns of erosion could be largely a function of solid Earth tectonic forces building topography, although the correlation of unroofing to the Late Miocene drying trend does raise the possible role for climate too, albeit in the opposite fashion to that seen since the LGM, when more Himalayan erosion correlates with strong summer monsoon rains.