Simulation of site-scale water fluxes in desert and natural oasis ecosystems of the arid region in Northwest China

The study of water fluxes is important to better understand hydrological cycles in arid regions. Data-driven machine learning models have been recently applied to water flux simulation. Previous studies have built site-scale simulation models of water fluxes for individual sites separately, requiring a large amount of data from each site and significant computation time. For arid areas, there is no consensus as to the optimal model and variable selection method to simulate water fluxes. Using data from seven flux observation sites in the arid region of Northwest China, this study compared the performance of random forest (RF), support vector machine (SVM), back propagation neural network (BPNN), and multiple linear regression (MLR) models in simulating water fluxes. Additionally, the study investigated inter-annual and seasonal variation in water fluxes and the dominant drivers of this variation at different sites. A universal simulation model for water flux was built using the RF approach and key variables as determined by MLR, incorporating data from all sites. Model performance of the SVM algorithm (R² = 0.25–0.90) was slightly worse than that of the RF algorithm (R² = 0.41–0.91); the BPNN algorithm performed poorly in most cases (R² = 0.15–0.88). Similarly, the MLR results were limited and unreliable (R² = 0.00–0.66). Using the universal RF model, annual water fluxes were found to be much higher than the precipitation received at each site, and natural oases showed higher fluxes than desert ecosystems. Water fluxes were highest during the growing season (May–September) and lowest during the non-growing season (October–April). Furthermore, the dominant drivers of water flux variation were various among different sites, but the normalized difference vegetation index (NDVI), soil moisture and soil temperature were important at most sites. This study provides useful insights for simulating water fluxes in desert and oasis ecosystems, understanding patterns of variation and the underlying mechanisms. Besides, these results can make a contribution as the decision-making basis to the water management in desert and oasis ecosystems.     

An analytical study on three‐dimensional versus two‐dimensional water table‐induced flow patterns in a Tóthian basin

Although it has been increasingly acknowledged that groundwater flow pattern is complicated in the three‐dimensional (3‐D) domain, two‐dimensional (2‐D) water table‐induced flow models are still widely used to delineate basin‐scale groundwater circulation. However, the validity of 2‐D cross‐sectional flow field induced by water table has been seldom examined. Here, we derive the analytical solution of 3‐D water table‐induced hydraulic head in a Tóthian basin and then examine the validity of 2‐D cross‐sectional models by comparing the flow fields of selected cross sections calculated by the 2‐D cross‐sectional model with those by the 3‐D model, which represents the “true” cases. For cross sections in the recharge or discharge area of the 3‐D basin, even if head difference is not significant, the 2‐D cross‐sectional models result in flow patterns absolutely different from the true ones. For the cross section following the principal direction of groundwater flow, although 2‐D cross‐sectional models would overestimate the penetrating depth of local flow systems and underestimate the recharge/discharge flux, the flow pattern from the cross‐sectional model is similar to the true one and could be close enough to the true one by adjusting the decay exponent and anisotropy ratio of permeability. Consequently, to determine whether a 2‐D cross‐sectional model is applicable, a comparison of hydraulic head difference between 2‐D and 3‐D solutions is not enough. Instead, the similarity of flow pattern should be considered to determine whether a cross‐sectional model is applicable. This study improves understanding of groundwater flow induced by more natural water table undulations in the 3‐D domain and the limitations of 2‐D models accounting for cross‐sectional water table undulation only.     

Shallow sediment deformation styles in north-western Campeche Knolls,Gulf of Mexico and their controls on the occurrence of hydrocarbon seepage

During German R/V Meteor M67/2 expedition to Campeche Knolls, southern Gulf of Mexico, a set of 2D high resolution seismic data was acquired to study the near-surface sediment structure and its relationship with hydrocarbon seepages in this salt province. The comprehensive survey covered 20 individual bathymetric highs or ridges and identified three principle structural types: Passive Type, Chaopopte Type, and Asymmetric Flap Type. The first type is the result of passive diapirism, whereas the latter two were initialized by a regional compressional event in the Miocene, but are later differently modified by salt tectonism. Chapopote Type structures appear as symmetrical domes, with uplifted coarse-grained Miocene sediments in their cores and rather thin syn-kinematic sediments covering the crests. Asymmetric Flap Type structures are also first folded as domes or ridges, but one flap later subsided together with salt evacuation, resulting in single uplifted monoclines. With the coarse-grained pre-kinematic sediments as reservoir units, both structural types can focus and accumulate hydrocarbons. The geometries of the structures suggest that hydrocarbons are accumulated in the center of the Chapopote Type structures and in the subsided flaps of the Asymmetric Flap Type structures. Hydrocarbon leakage from these thinly sealed reservoirs is regarded as the principle mechanism for the seepage in the study area, and accordingly the most seepage-prone positions are above these reservoirs. The seep locations suggested by analysis of sea-surface oil slick images of SAR satellite data are also examined in this study. These independently derived seep locations confirm the seepage-prone positions to be above the shallow buried reservoirs. This study suggest that the shallow sediment structures control the distribution of the hydrocarbon seeps of the north-western Campeche Knolls, although the hydrocarbons are sourced from the greater depth.     

Oxygen‐18 dynamics in precipitation and streamflow in a semi‐arid agricultural watershed,Eastern Washington,USA

Understanding flow pathways and mechanisms that generate streamflow is important to understanding agrochemical contamination in surface waters in agricultural watersheds. Two environmental tracers, δ¹⁸O and electrical conductivity (EC), were monitored in tile drainage (draining 12 ha) and stream water (draining nested catchments of 6‐5700 ha) from 2000 to 2008 in the semi‐arid agricultural Missouri Flat Creek (MFC) watershed, near Pullman Washington, USA. Tile drainage and streamflow generated in the watershed were found to have baseline δ¹⁸O value of ?14·7‰ (VSMOW) year round. Winter precipitation accounted for 67% of total annual precipitation and was found to dominate streamflow, tile drainage, and groundwater recharge. ‘Old’ and ‘new’ water partitioning in streamflow were not identifiable using δ¹⁸O, but seasonal shifts of nitrate‐corrected EC suggest that deep soil pathways primarily generated summer streamflow (mean EC 250 µS/cm) while shallow soil pathways dominated streamflow generation during winter (EC declining as low as 100 µS/cm). Using summer isotopic and EC excursions from tile drainage in larger catchment (4700‐5700 ha) stream waters, summer in‐stream evaporation fractions were estimated to be from 20% to 40%, with the greatest evaporation occurring from August to October. Seasonal watershed and environmental tracer dynamics in the MFC watershed appeared to be similar to those at larger watershed scales in the Palouse River basin. A 0·9‰ enrichment, in shallow groundwater drained to streams (tile drainage and soil seepage), of δ¹⁸O values from 2000 to 2008 may be evidence of altered precipitation conditions due to the Pacific Decadal Oscillation (PDO) in the Inland Northwest. Copyright © 2009 John Wiley & Sons, Ltd.     

Weichselian interstadials at Riipiharju,northern Sweden – interpretation of vegetation and climate from fossil and modern pollen records

Hättestrand, M. & Robertsson, A.‐M. 2010: Weichselian interstadials at Riipiharju, northern Sweden – interpretation of vegetation and climate from fossil and modern pollen records. Boreas, 10.1111/j.1502‐3885.2009.00129.x. ISSN 0300‐9483. The most complete records of Weichselian ice‐free conditions in northern Sweden have been retrieved from kettleholes in the Riipiharju esker. In an earlier study, the Riipiharju I core was described as containing two Weichselian interstadials and Riipiharju was chosen as type site for the second Weichselian interstadial in northern Sweden. Here, we present a palynological investigation of two new sediment cores (Riipiharju II and III) retrieved from Riipiharju. Together, the new cores comprise a late cold part of the first Weichselian interstadial recorded in northeastern Sweden (Tärendö I, earlier correlated with Peräpohjola in Finland) as well as a long sequence of the second Weichselian interstadial (Tärendö II, earlier named Tärendö). The results indicate that the climate during deposition of the Tärendö II sequence was more variable than earlier suggested. According to the present interpretation it was relatively warm in the early part of Tärendö II; thereafter a long cold phase persisted, and finally the climate was warmer again in the late part of Tärendö II. The warm phases are characterized by Betula‐dominant pollen assemblages, while the cold phase is characterized by high percentages of Artemisia and Gramineae pollen. Since there is still no firm chronology established of the interstadials in northeastern Sweden, two possible correlations are discussed; either Tärendö I and II are correlated with Brörup (MIS 5c) and Odderade (MIS 5a), or, perhaps more likely, they are correlated with Odderade and early Middle Weichselian (MIS 3) time.     

The data‐driven approach as an operational real‐time flood forecasting model

Accurate water level forecasts are essential for flood warning. This study adopts a data‐driven approach based on the adaptive network–based fuzzy inference system (ANFIS) to forecast the daily water levels of the Lower Mekong River at Pakse, Lao People's Democratic Republic. ANFIS is a hybrid system combining fuzzy inference system and artificial neural networks. Five ANFIS models were developed to provide water level forecasts from 1 to 5 days ahead, respectively. The results show that although ANFIS forecasts of water levels up to three lead days satisfied the benchmark, four‐ and five‐lead‐day forecasts were only slightly better in performance compared with the currently adopted operational model. This limitation is imposed by the auto‐ and cross‐correlations of the water level time series. Output updating procedures based on the autoregressive (AR) and recursive AR (RAR) models were used to enhance ANFIS model outputs. The RAR model performed better than the AR model. In addition, a partial recursive procedure that reduced the number of recursive steps when applying the AR or the RAR model for multi‐step‐ahead error prediction was superior to the fully recursive procedure. The RAR‐based partial recursive updating procedure significantly improved three‐, four‐ and five‐lead‐day forecasts. Our study further shows that for long lead times, ANFIS model errors are dominated by lag time errors. Although the ANFIS model with the RAR‐based partial recursive updating procedure provided the best results, this method was able to reduce the lag time errors significantly for the falling limbs only. Improvements for the rising limbs were modest. Copyright © 2011 John Wiley & Sons, Ltd.     

Dissolved nitrous oxide (N2O) dynamics in agricultural field drains and headwater streams in an intensive arable catchment

Indirect nitrous oxide (N₂O) emissions produced by nitrogen (N) leaching into surface water and groundwater bodies are poorly understood in comparison to direct N₂O emissions from soils. In this study, dissolved N₂O concentrations were measured weekly in both lowland headwater streams and subsurface agricultural field drain discharges over a 2‐year period (2013–2015) in an intensive arable catchment, Norfolk, UK. All field drain and stream water samples were found to have dissolved N₂O concentrations higher than the water–air equilibrium concentration, illustrating that all sites were acting as a net source of N₂O emissions to the atmosphere. Soil texture was found to significantly influence field drain N₂O dynamics, with mean concentrations from drains in clay loam soils (5.3 μg N L^?1) being greater than drains in sandy loam soils (4.0 μg N L^?1). Soil texture also impacted upon the relationships between field drain N₂O concentrations and other water quality parameters (pH, flow rate, and nitrate (NO₃) and nitrite (NO₂) concentrations), highlighting possible differences in N₂O production mechanisms in different soil types. Catchment antecedent moisture conditions influenced the storm event mobilisation of N₂O in both field drains and streams, with the greatest concentration increases recorded during precipitation events preceded by prolonged wet conditions. N₂O concentrations also varied seasonally, with the lowest mean concentrations typically occurring during the summer months (JJA). Nitrogen fertiliser application rates and different soil inversion regimes were found to have no effect on dissolved N₂O concentrations, whereas higher N₂O concentrations recorded in field drains under a winter cover crop compared to fallow fields revealed cover crops are an ineffective greenhouse gas emission mitigation strategy. Overall, this study highlights the complex interactions governing the dynamics of dissolved N₂O concentrations in field drains and headwater streams in a lowland intensive agricultural catchment.     

Tropical precipitation anomalies and d‐excess evolution during El Niño 2014‐16

The last 2014‐16 El Niño event was among the three strongest episodes on record. El Niño considerably changes annual and seasonal precipitation across the tropics. Here, we present a unique stable isotope data set of daily precipitation collected in Costa Rica prior to, during, and after El Niño 2014‐16, in combination with Lagrangian moisture source and precipitation anomaly diagnostics. δ²H composition ranged from ‐129.4 to +18.1 (‰) while δ¹⁸O ranged from ‐17.3 to +1.0 (‰). No significant difference was observed among δ¹⁸O (P=0.186) and δ²H (P=0.664) mean annual compositions. However, mean annual d‐excess showed a significant decreasing trend (from +13.3 to +8.7 ‰) (P<0.001) with values ranging from +26.6 to ‐13.9 ‰ prior to and during the El Niño evolution. The latter decrease in d‐excess can be partly explained by an enhanced moisture flux convergence across the southeastern Caribbean Sea coupled with moisture transport from northern South America by means of an increased Caribbean Low Level Jet regime. During 2014‐15, precipitation deficit across the Pacific domain averaged 46% resulting in a very severe drought; while a 94% precipitation surplus was observed in the Caribbean domain. Understanding these regional moisture transport mechanisms during a strong El Niño event may contribute to a) better understanding of precipitation anomalies in the tropics and b) re‐evaluate past stable isotope interpretations of ENSO events in paleoclimatic archives within the Central America region.     

Influence of heterogeneity on unsaturated hydraulic properties: (1) local heterogeneity and scale effect

Spatial heterogeneity is ubiquitous in nature, which may significantly affect the soil hydraulic property curves. The models of a closed‐form functional relationship of soil hydraulic property curves (e.g. VG model or exponential model) are valid at point or local scale based on a point‐scale hydrological process, but how do scale effects of heterogeneity have an influence on the parameters of these models when the models are used in a larger scale process? This paper uses a two‐dimensional variably saturated flow and solute transport finite element model (VSAFT2) to simulate variations of pressure and moisture content in the soil flume under a constant head boundary condition. By changing different numerical simulation block sizes, a quantitative evaluation of parameter variations in the VG model, resulting from the scale effects, is presented. Results show that the parameters of soil hydraulic properties are independent of scale in homogeneous media. Parameters of α and n in homogeneous media, which are estimated by using the unsaturated hydraulic conductivity curve (UHC) or the soil water retention curve (WRC), are identical. Variations of local heterogeneities strongly affect the soil hydraulic properties, and the scale affects the results of the parameter estimations when numerical experiments are conducted. Furthermore, the discrepancy of each curve becomes considerable when moisture content becomes closer to a dry situation. Parameters estimated by UHC are totally different from the ones estimated by WRC. Copyright © 2012 John Wiley & Sons, Ltd.     

Jet instability and the stabilizing effect of topography on jets in two-layer rotating systems

Abstract

Laboratory experiments concerning azimuthal jets in two-layer rotating systems in the absence and presence of bottom topography aligned along the jets have been conducted. The jets were forced by the selective withdrawal of fluid from the upper layer of a two-fluid system contained in a circular dishpan geometry. The principal parameters measured in the experiments were the jet Rossby number, Ro, and a stratification parameter F = r ₁/(λ₁λ₂)^1/2 where r ₁ is the radius of the circular disc used for the selective withdrawal (i.e., r ₁ is the approximate radius of curvature of the jet) and λ₁,λ₂ are the internal Rossby radii of deformation in the upper and lower fluids, respectively.

The no-topography experiments show that for a sufficiently small F, the particular value depending on Ro, the jet is stable for the duration of the experiment. For sufficiently large F, again as a function of Ro, the jet becomes unstable, exhibiting horizontal wave disturbances from modes three to seven. An Ro against F flow regime diagram is presented.

Experiments are then conducted in the presence of a bottom topography having constant cross-section and extending around a mid-radius of the dishpan. The axis of the topography is in the vicinity of the jet axis forced in the no-topography experiments and the crest of the topography is in the vicinity of the interface between the two fluids (i.e., the front associated with the jet). The experiments show that in all cases investigated the jet tends to be stabilized by the bottom topography. Experiments with the topography in place, but with the interface between the fluids being above the topography crest, are shown to be unstable but more irregular than their no-topography counterparts.

Various quantitative measurements of the jet are presented. It is shown, for example, that the jet Rossby number defined in terms of the fluid withdrawal rate from the tank. Q, can be well correlated with a dimensionless vorticity gradient, V_G , across the upper layer jet. This allows for an assessment of the stability characteristics of a jet based on a knowledge of V_G (which can be estimated given a jet profile) and F.