Ukinrek Maars, Alaska, II. Deposits and formation of the 1977 craters

The initial phase of the eruption forming Ukinrek Maars during March and April 1977 were explosions from the site of West Maar. These were mainly phreatomagmatic and initially transitional to strombolian. Activity at West Maar ceased after three days upon the initiation of the East Maar. The crater quickly grew by strong phreatomagmatic explosions. During the first phases of phreatomagmatic activity at East Maar, large exotic blocks derived from a subsurface till were ejected. Ballistic studies indicate muzzle velocities for these blocks of 80–90 m s⁻¹.Phreatomagmatic explosions ejected both juvenile and non-juvenile material which formed a low rim of ejecta (< 26 mhigh) around the crater and a localized, coarse, wellsorted (σ_φ = 1−1.5) juvenile and lithic fall deposit. Other fine ash beds, interstratified with the coarse beds, are more poorly sorted (σ_φ = 2−3) and are interpreted as fallout of wet, cohesive ash from probably milder phases of activity in the crater. Minor base surge activity damaged trees and deposited fine ash, including layers plastered on vertical surfaces. Viscous basalt lava appeared in the center of the East Maar crater almost immediately and a lava dome gradually grew in the crater despite phreatomagmatic eruptions adjacent to it.The development of these maars appears to be mainly controlled by gradual collapse of crater and conduit walls, and blasting-out of the slumped debris by phreatomagmatic explosions when rising magma contacted groundwater beneath the regional water table and a local perched aquifer.Ballistic analysis on the ejected blocks indicates a maximum muzzle velocity of 100–150 m s^-1, values similar to those obtained from other ballistic studies on maar ejecta.     

Luminescence dating of old (>70 ka) Chinese loess: A comparison of single-aliquot OSL and IRSL techniques

The applicability of two different approaches in the luminescence dating of old (>70 ka) Chinese loess is investigated. Both SAR-OSL ages obtained on 63–90 μm quartz grains and SAR-IRSL ages obtained on 4–11 μm polymineral grains, for samples collected from two sites in the Chinese Loess Plateau (Luochuan and Dongchuan) are presented. The characteristics of the luminescence signals stimulated by blue and infrared light are investigated in terms of dose response and dose recovery, and as a function of age. Additionally, anomalous fading measurements from the 410 nm IRSL emission in polymineral fine-grains are reported. An average value of g_2days amounting to 3% per decade was measured and seems to be independent of site location and age. For the samples from Luochuan, independent age control (pedostratigraphy and palaeomagnetism) is available. At both sites, the SAR-OSL ages are always lower than the SAR-IRSL ages after they have been corrected for anomalous fading. It seems that the quartz-based SAR-OSL ages are accurate for the younger ages, but that they underestimate the true age of deposition for loess that was deposited about 60–70 ka ago. The fading-corrected SAR-IRSL ages are in better agreement with the pedostratigraphic age control (75 and 130 ka) and allow dating beyond the quartz OSL range. Based on our results, we suggest that conventional SAR-OSL and SAR-IRSL protocols at these sites should be restricted to samples of ages not exceeding 40–50  and 100–120 ka, respectively.     

K-feldspar IRSL dating of a Pleistocene river terrace staircase sequence of the Lower Tejo River (Portugal,western Iberia)

We present the results of K-feldspar IRSL dating of the four lower terraces (T3–T6) of the Portuguese Tejo River, in the Arripiado-Chamusca area. Terrace correlation was based upon: a) analysis of aerial photographs, geomorphological mapping and field topographic survey; b) sedimentology of the deposits; and c) luminescence dating. Sediment sampled for luminescence dating gave unusually high dose rates, of between 3.4 and 6.2 Gy/ka and, as a result, quartz OSL was often found to be in saturation. We therefore used the IRSL signal from K-feldspar as the principal luminescence technique. The K-feldspar age results support sometimes complex geomorphic correlations, as fluvial terraces have been vertically displaced by faults (known from previous studies). Integration of these new ages with those obtained previously in the more upstream reaches of the Tejo River in Portugal indicates that the corrected K-feldspar IRSL ages are stratigraphically and geomorphologically consistent over a distance of 120 km along the Tejo valley. However, we are sceptical of the accuracy of the K-feldspar ages of samples from the T3 and T4 terraces (with uncorrected D_e values >500 Gy). In these cases the Dose Rate Correction (DRC) model puts the natural signals close to luminescence saturation, giving a minimum corrected D_e of about 1000 Gy, and thus minimum terrace ages; this may even be true for those doses >200 Gy. Luminescence dating results suggest that: T3 is older than 300 ka, probably ca. 420–360 ka (～Marine Isotope Stage [MIS]11); T4 is ca. 340–150 ka (～MIS9-6); T5 is 136–75 ka (～MIS5); T6 is 60–30 ka (MIS3); an aeolian sand unit that blankets T6 and some of the older terraces is 30–≥12 ka. Collectively, the luminescence ages seem to indicate that regional river downcutting events may be coincident with periods of low sea level (associated, respectively, with the MIS10, MIS6, MIS4 and MIS2).     

A hybrid optimization approach for efficient calibration of computationally intensive hydrological models

ABSTRACT

The calibration of hydrological models is formulated as a blackbox optimization problem where the only information available is the objective function value. Distributed hydrological models are generally computationally intensive, and their calibration may require several hours or days which can be an issue for many operational contexts. Different optimization algorithms have been developed over the years and exhibit different strengths when applied to the calibration of computationally intensive hydrological models. This paper shows how the dynamically dimensioned search (DDS) and the mesh adaptive direct search (MADS) algorithms can be combined to significantly reduce the computational time of calibrating distributed hydrological models while ensuring robustness and stability regarding the final objective function values. Five transitional features are described to adequately merge both algorithms. The hybrid approach is applied to the distributed and computationally intensive HYDROTEL model on three different river basins located in Québec (Canada).