Kaustubh Thirumalai

and 2 more

Stable carbon (δ¹³C) and oxygen (δ¹⁸O) isotope measurements in lacustrine ostracodes are widely used to infer past climatic conditions. Previous work has used individual ostracode valves to resolve seasonal and subdecadal climate signals, yet environmental controls on geochemical variability within co-occurring specimens from modern samples are poorly constrained. Here we focus on individual ostracode valves in modern-aged Lake Turkana sediments, an alkaline desert lake in tropical East Africa. We present individual ostracode valve analyses (IOVA) of δ¹³C and δ¹⁸O measurements (n = 329) of extant species Sclerocypris clavularis from 17 sites spanning the entire lake (n-avg ~19 specimens per site). We demonstrate that the pooled statistics of individual valve measurements at each site overcome inter-specimen isotopic variance and are driven by hydrological variability in the lake. Mean IOVA-δ¹³C and -δ¹⁸O across the sites exhibit strong spatial trends with higher values at more southerly latitudes, modulated by distance from the inflow of the Omo River. Whereas the latitudinal δ¹³C gradient reflects low riverine δ¹³C and decreasing lacustrine productivity towards the southern part of the lake, the δ¹⁸O gradient is controlled by evaporation superimposed on the waning influence of low-δ¹⁸O Omo River waters, sourced from the Ethiopian highlands. We show that ostracode δ¹⁸Oproximal to Omo River inflow is deposited under near-equilibrium conditions and that inter-specimen δ¹⁸O variability across the basin is consistent with observed temperature and lake water δ¹⁸O variability. IOVA can provide skillful constraints on high-frequency paleoenvironmental signals and, in Omo-Turkana sediments, yield quantitative insights into East African paleohydrology.
Ecologists have proposed that montane grassland-shola (stunted evergreen forest) mosaics in the Western Ghats may represent alternative stable vegetation states. But paleoecology investigations seldom consider this framework, especially the role of short-term disturbances (fire, intense drought) other than long-term climatic changes, that can cause vegetation switches in landscapes with alternative vegetation states. The Sandynallah valley that hosts one of the oldest peat accumulations in the world at >50 kyr has been central to the reconstruction of paleovegetation in the montane Nilgiris, Western Ghats. Although the peat-forming vegetation here (dominated by sedges) is a unique vegetation state, its contribution to the paleovegetation signal has not been explicitly considered. We propose a conceptual framework of a tri-stability landscape with sedgeland on the valley floor, grassland on the hill slopes and shola vegetation in the boundary between sedgeland and grassland. While frost prevents shola saplings from establishing in grassland, waterlogging provides a barrier for their establishment in sedgeland, thus maintaining these distinct vegetation states under the same climate. We investigated the stable carbon isotope signatures of the cellulose fraction from two well-dated peat cores (Cores 1 and 2) collected from ~170m apart in the Sandynallah valley within the alternative stable states framework. We find that Core 1, which is closer to the boundary of valley and hill slope, shows dynamic switches between sedgeland and shola whereas Core 2, located in the centre of the valley floor, represents a stable sedgeland state. The vegetation switches and maintenance mechanisms at Core 1 is connected to a disturbance (fire) and to changing climate while Core 2 seems to be responding primarily to climatic changes. The simultaneously distinctive vegetation states in Cores 1 and 2 at such close proximity within the same valley is the first record of alternative stables states in the past in the Western Ghats.
The Sandynallah valley (Western Ghats, India) features one of the oldest peat accumulations in the world at >50 kyr and has been central to the reconstruction of late Quaternary paleoclimate using paleovegetation changes in the forest-grassland vegetation mosaic that coexist here. It is well-known that short-term disturbances (fire, frost, intense drought) can also cause vegetation switches when multiple stable states exist, but this framework has seldom been considered in paleoecology investigations. Using stable carbon isotope signatures (relative C3-C4 vegetation abundance) on the cellulose fraction from two well-dated peat cores ~170 m apart - Core 1 closer to the hillslope (32000 years old) and Core 2 from the centre of the valley floor (45,000 years old) - we looked at paleovegetation changes and the implications for paleoclimate reconstruction within the alternative stable states framework. Charcoal data from another sediment profile from the same valley was used to correlate with paleofires. We propose that the valley floor is bistable, switching between peat-forming vegetation ‘sedgeland’ and montane stunted evergreen forest ‘shola’, maintained by level of waterlogging. Core 1 shows shola-sedgeland dynamics with vegetation switching at c.22ka from shola (possibly due to fires) to a prolonged unstable state until 13 ka sustained by low waterlogging. Following a hiatus c.13-7 ka, sedgeland dominates, with a shift into shola at 3.75 ka driven by increasing aridity. Core 2 shows a relatively stable signature, enriched in C3-vegetation in the last glacial (45-20 ka) compared to the Holocene. Given temperature is the primary driver of abundance in C3-C4 mixed-grasslands, C4 dominance beginning c.18.5 ka followed by C4 enrichment is indicative of deglacial warming that continues into the Holocene except for a departure at ~10 ka. The record at Core 2 is indicative of changing climate while Core 1 shows disturbance-based vegetation dynamics. The simultaneously distinctive vegetation states in Cores 1 and 2 within the same valley is the first record of alternative stable states in the past in the montane tropics. Our results point to the need to account for short-term disturbances and site attributes before ascribing vegetation changes to changing climate in alternative stable states landscapes.