Artificial Neural Networks in Proglacial Discharge Simulation: Application and Efficiency Analysis in Comparison to the Multivariate Regression; A Case Study of Waldemar River (Svalbard)

Artificial neural networks were applied to simulate runoff from the glacierized part of the Waldemar River catchment (Svalbard) based on hydrometeorological data collected in the summer seasons of 2010, 2011 and 2012. Continuous discharge monitoring was performed at about 1 km from the glacier snout, in the place where the river leaves the marginal zone. Averaged daily values of discharge and selected meteorological variables in a number of combinations were used to create several models based on the feed‐forward multilayer perceptron architecture. Due to specific conditions of melt water storing and releasing, two groups of models were established: the first is based on meteorological inputs only, while second includes the preceding day's mean discharge. Analysis of the multilayer perceptron simulation performance was done in comparison to the other black‐box model type, a multivariate regression method based on the following efficiency criteria: coefficient of determination (R²) and its adjusted form (adj. R²), weighted coefficient of determination (wR²), Nash–Sutcliffe coefficient of efficiency, mean absolute error, and error analysis. Moreover, the predictors' importance analysis for both multilayer perceptron and multivariate regression models was done. The performed study showed that the nonlinear estimation realized by the multilayer perceptron gives more accurate results than the multivariate regression approach in both groups of models.     

Precise Multibeam Acoustic Bathymetry

The maximum error in ocean depth measurement as specified by the International Hydrographic Organization is 1% for depth greater than 30m. Current acoustic multibeam bathymetric systems used for depth measurement are subject to errors from various sources which may significantly exceed this limit. The lack of sound speed profiles may be one significant source of error. Because of the limited ability of sound speed profile measurement, depth values are usually estimated using an assumed profile. If actual sound speed profiles are known, depth estimate errors can be corrected using ray-tracing methods. For depth measurements, the calculation of the location at which a sound pulse impinges on the sea bottom varies with the variation of the sound speed profile. We demonstrate that this location is almost unchanged for a family of sound speed profiles with the same surface value and the same area under them. Based on this observation, we can construct a simple constant-gradient equivalent sound speed profile to correct errors. Compared with ray-tracing methods, the equivalent sound speed profile method is more efficient. If a vertical depth is known (or independently measured), then depth correction for a multibeam system can be accomplished without knowledge of the actual sound speed profile. This leads to a new type of precise acoustic multibeam bathymetric system.     

The formation of fan-deltas in mountain lakes: Findings from Zelené Kežmarské Lake (Slovakia)

Fan-deltas are formed in mountain lakes, contributing to changes in shorelines' shapes and filling lake basins with sediments. Factors that condition sediment delivery to the lakes and the formation of fan-deltas are not fully understood. This study aims to identify processes forming fan-deltas in mountain lakes based on the sedimentary architecture of a fan-delta filling Zelené Kežmarské Lake (Slovakia). Our study is based on ground-penetrating radar and seismic refraction surveys conducted over the lake and debris-flow fans in its vicinity, and grain-size analyses of the surface deposits. The internal structure of the fan-delta comprises foreset deposits representing the fan-delta lobes. Mouth bars were identified in the near-shore zone. The fan-delta is built of sands and silty sands whereas an alluvial debris-flow dominated fan west of the lake contains gravels. A general trend of downslope fining of the fan surface sediments, disturbed by zones of coarse gravels, was identified in the surficial sediments forming the fan. The fan-delta was formed by depositional events in humid periods, alternating with dry periods featured with a small or no deposition. The lake-level steadily rose in the humid periods and remained stable in dry periods. The contrast between the coarse sediments forming the alluvial debris-flow dominated fan, and sands forming the fan-delta was caused by a selective deposition. Coarse gravels in the fan were mobilized in humid periods but did not reach the lake-shore. The reason was a basement rock ridge situated parallel to the lake shore. Sands were delivered to the fan-delta in humid periods owing to frequent high-energy flows. A deposition of silts took place in dry periods. The studied fan-delta might preserve the sedimentary record of three humid periods from the last 200 years. However, further studies are needed to establish the timing of the fan-delta formation.     

Variability of Sea-Ice Extent in Baffin Bay over the Last Millennium

Comparison of an ice core glaciochemical time-series developed from thePenny Ice Cap (PIC), Baffin Island and monthly sea-ice extent reveals astatisticallysignificant inverse relationship between changes in Baffin Bay spring sea-iceextent andPenny Ice Cap sea-salt concentrations for the period 1901–1990 AD.Empiricalorthogonal function analysis demonstrates the joint behavior between changesin PICsea-salt concentrations, sea-ice extent, and changes in North Atlanticatmosphericcirculation. Our results suggest that sea-salt concentrations in snowpreserved on thePIC reflect local to regional springtime sea-ice coverage. The PIC sea-saltrecord/sea-ice relationship is further supported by decadal and century scalecomparisonwith other paleoclimate records of eastern Arctic climate change over the last700 years. Our sea-salt record suggests that, while the turn of the century wascharacterized bygenerally milder sea-ice conditions in Baffin Bay, the last few decades ofsea-ice extentlie within Little Ice Age variability and correspond to instrumental recordsof lowertemperatures in the Eastern Canadian Arctic over the past three decades.     

A multiyear analysis of aerosol optical thickness over Europe and Central Poland using NAAPS model simulation

This study contains a comparative analysis of aerosol optical thickness (AOT) between numerical calculations obtained from the Navy Aerosol Analysis and Prediction System (NAAPS) model and direct observations from the AERONET robotic network and the Saharan Aerosol over WArsaw (SAWA) field campaign. AOT was calculated for 500 nm wavelength. The comparison shows underestimation of the total aerosol optical thickness simulated by NAAPS. The correlation coefficients between model and observation oscillates between 0.57 and 0.72. Results of seven-year (1998–2004) NAAPS simulation of aerosol components (sea salt, mineral dust, sulphate, and smoke) show large temporal and spatial variability of the aerosol optical thickness over Europe. The least polluted region is the Iberian Peninsula, while the highest aerosol burdens occurred in Central Europe, mostly due to anthropogenic sulphate particles. Finally, the analysis of mineral dust transport shows frequent episodes of Saharan dust inflow over Central Europe. There are about 20 days a year (4 days in May) when instantaneous AOT associated with mineral dust aerosol increases over 0.1.     

Slope processes in the High Arctic: a 30‐year‐long record from Spitsbergen

Ten monitoring sites were established in 1986 near the Scottbreen and Renardbreen glaciers (Bellsund, Spitsbergen) in order to determine the rates of gelifluction on raised marine terraces and on talus cones, accumulation on talus cones, and movement of the protalus rampart. The measurements were performed in 1987, 2007 and 2016. The results obtained revealed a very slow rate of changes on the surface of the analysed area. During 30 years, only one out of 84 points where the gelifluction rate was measured was displaced by 15 cm down‐slope. The maximal annual and average gelifluction rates were calculated at 0.5 cm and 0.18 cm a^?1, respectively. Several measurement points did not change their position over the observation period. As shown by the observations, the gelifluction rate largely does not depend on slope inclination. Studies of talus cones allowed the level of cone increment to be determined as 314.3 kg of scree per square metre. Analysis of factors influencing the rate of mass movement allowed for correlation of the small‐scale gelifluction movement at the analysed sites with the general trends of climate warming. In this part of the Arctic, these trends are manifested by soil desiccation. It appears that slope processes depend on very local topoclimatic factors. Analysis of the obtained data with regard to palaeogeographical and climatic investigations suggests that the ice segregation forms on raised marine terraces developed during the last Holocene cooling, i.e. the Little Ice Age.     

Free core nutation period inferred from the gravity measurements at Józefosław

The Free Core Nutation (FCN) is an important eigenmode which affects both Earth rotation and body tide. The FCN parameters, the resonance period and the quality factor are important for understanding the dynamics of the Earth at nearly diurnal periods. Those parameters are usually estimated either from the Very Long Baseline Interferometry (VLBI) observations of nutation, or from the tidal gravity measurements. In this paper we investigate the determination of the FCN parameters from gravity records covering a period of more than three years, collected with the use of a LaCoste&Romberg Earth Tide no. 26 gravimeter, located at Józefos?aw observatory near Warsaw. From the resonant enhancements of gravimetric factors and phases of diurnal tidal gravity waves, we could infer the FCN period to be equal to 430 sidereal days. This result is in very good agreement with previous gravimetric and VLBI nutation results, confirming the discrepancy in the dynamic flattening of the outer liquid core from its theoretical value based on the hydrostatic equilibrium assumption. The estimated FCN quality factor (Q ≈ 1300) is considerably smaller than the VLBI nutation result, which confirms that the local gravity measurements are more sensitive than VLBI global analyses to site-dependent phenomena (such as atmospheric and indirect ocean tidal effects). We also investigated the importance of gravimetric corrections in the FCN analysis, including numerical tests and simulations. This allowed us to estimate the uncertainty of FCN parameters due to improper or incomplete set of environmental corrections. We took also into account the impact of gravimetric factor errors and tidal wave selection on estimated FCN parameters. We demonstrated that despite relatively noisy measurements due to unfavorable gravimeter location, we were able to obtain very good results in case when proper correction and tidal wave selection were applied.     

Evolution of river planforms downstream of dams: Effect of dam construction or earlier human‐induced changes?

When studying the evolution of landscape, it is difficult to discriminate the influence of anthropogenic from natural causes, or recognise changes caused by different sources of human action. This is especially challenging when the influence of certain sources is overprinted. For instance, although dam closure is the most common method of altering river courses, dam construction is often preceded by hydro‐technical works such as channel straightening, embankment construction or sediment mining. Both dam construction and the hydro‐technical works that precede dam closure can result in changes in the balance between sediment supply and transport capacity, and often, changes in river planform. The main objective of this study was to verify whether the works preceding dam closure are an important driver of river planform changes on the lower Drava River (Hungary). The case study is based on geological and geophysical surveys, as well as the analysis of historical maps covering an anabranching, 23 km long valley section. We show that channel straightening conducted prior to dam closure resulted in a transition from a meandering to sinuous planform with channel bars. Dam construction itself then caused enhanced incision, exposure of bar surfaces, vegetation encroachment and the formation of an anabranching planform. Based on this study, we developed models of alluvial island and channel planform evolution downstream of dams. Dam construction enhances channel incision, narrowing, and the reduction of flow caused by earlier hydro‐technical works. Many rivers downstream of dams experience episodes of anabranching or wandering, with a multi‐thread pattern replacing sinuous, braided and meandering courses. When incision continues, river patterns evolve from anabranching to sinuous via the attachment of alluvial islands to floodplains. However, the timing and sequence of these changes depend on hydrological and sediment supply regimes, geomorphic settings and anthropogenic actions accompanying dam construction. Copyright © 2018 John Wiley & Sons, Ltd.     

Tectonic plates parameters estimated in International Terrestrial Reference Frame ITRF2008 based on DORIS stations

This paper concerns an analysis of the accuracy of the estimated parameters Ω (Φ, Λ, ω) which define the tectonic plate motions. The study is based on the velocities of station positions in the IERS (International Earth Rotation and Reference Systems Service) which has published new realization of the International Terrestrial Reference System—ITRF2008 for Doppler Orbitography by Radiopositioning Integrated on Satellite DORIS technique. Eurasian, African, Australian, North American, Australian, Pacific, Antarctic and South American plates were used in the analysis. The influence of the number and localization of stations on the plate surface on the estimation accuracy of the tectonic plate motion parameters were discussed. The results were compared with the APKIM 2005 IGN model and our earlier estimation for the SLR technique. In general, a remarkable concurrent agreement between the present and the APKIM 2005 solutions was found.