Reconstructing paleoclimates through 3D numerical modeling of tropical Andes paleoglacial flow.

The glaciology community aims to link climate with moraine positions; however, this is challenging because the glacial dynamics responsible for moraine deposition depend on both precipitation and temperature. These dynamics are further influenced by various physical parameters, such as valley hypsometry (Pratt-Sitaula et al., 2011). Consequently, many studies focus solely on dating morphologies to discuss relative chronologies, rather than attempting to reconstruct past climates from moraine extents. Thus, even though moraines are a key archive of continental paleoclimatology, they have not yet delivered their full potential in reconstructing past climates.To overcome these issues, we propose a novel methodology for interpreting past climate conditions using sets of moraines from different glaciers that (i) have close or similar ages and (ii) lie close enough to each other to assume they experienced the same paleoclimate. We rely on simulations performed with the numerical glacier model Elmer/Ice, which combines a 3D full Stoke-type glacier flow model with a distributed surface mass balance model, accurately capturing glacier dynamics and geometry sensitivity to valley hypsometry. Within a steady state setup, we simulate the extent of two neighboring paleo glaciers under different precipitation (P) and temperature (T) conditions, finding a P-T solution curve for each that explains the glaciers position/volume constrained by their moraines. The intersection of both curves, therefore, reveals the climatic condition that dominated the valley during the age of the moraines.We implemented our method in the altiplano valley of Zongo, Bolivia, taking advantage of the long-term glaciological monitoring of the Zongo glacier to calibrate our mass balance model. We studied late-glacial moraines from Telata and Charquini (5390 m asl), two neighboring paleo glaciers with almost the same geomorphological characteristics. We also considered the Zongo glacier itself (6088 m asl), but since it does not have moraines that match the ages of those from Telata and Charquini, we created a similar moraine formation for analysis purposes. Charquini and Telata showed close and almost parallel P-T solution curves, which were crossed by the steeper solution curve of Zongo, thus highlighting the dependence of our method on hypsometry contrasts between glacial valleys.The fact that these P-T curves have a single intersection suggests that, with a good elevation contrast, our methodology allows us to reconstruct past climatic conditions and decipher the joint contribution of precipitation and temperature from moraine positions only, which has never been achieved in previous studies based on ELA reconstruction or energy balance models (e.g., Autin et al., 2022; Rabatel et al., 2006) where strong assumptions about P or T are needed. We are currently working with two other Altiplano paleoglaciers (Aricoma and Llampu) that have a good hypsometry contrast and will allow us to compare the P-T results with the paleoclimatic conditions presented in the literature (e.g., Martin et al., 2018; Jomelli et al., 2014), enhancing our understanding of the regional signature of global climate changes.