青海湖水量平衡及水位变化预测

青海湖是我国最大的内陆湖泊,流域面积29661km~(2),水面高程超过3000m,受人为活动影响相对较少,基本上还处于半自然状态。水量平衡计算结果表明,有观测资料的近30年来,青海湖处于负平蘅状态,水位下降了2.96m,平均每年下降10.2cm。如果未来湖区的气候大体保持过去的情况,水位将再下降5.8m,经过57年才能平衡。如果考虑“温室效应”所引起的西北地区未来气候变化,水位亦将下降,每年平均下降10.1cm。     

常规海面风预报质量检验的实现

常规海面风预报内容是“文字加字符”形式,实况数据是间隔均匀的数据序列,所以常规海面风预报无法像数值预报一样直接应用实况数据进行预报质量的检验评估。本文在详细调研全国海洋预报机构发布的海面风预报产品基础上,针对不同的常规预报内容,制定了一套客观、合理、标准化的处理方法,成功实现了对常规海面风预报的检验,并应用东海区实况数据对常规海面风预报进行了检验测试,通过对存在的问题进行研讨与论证,完善了检验评估规则,提升了其实用性、合理性和适用性。     

Verification of short-term runoff forecasts for a small Philippine basin (Marikina)

Storm runoff from the Marikina River Basin frequently causes flood events in the Philippine capital region Metro Manila. This paper presents and evaluates a system to predict short-term runoff from the upper part of that basin (380 km²). It was designed as a possible component of an operational warning system yet to be installed. For the purpose of forecast verification, hindcasts of streamflow were generated for a period of 15 months with a time-continuous, conceptual hydrological model. The latter was fed with real-time observations of rainfall. Both ground observations and weather radar data were tested as rainfall forcings. The radar-based precipitation estimates clearly outperformed the raingauge-based estimates in the hydrological verification. Nevertheless, the quality of the deterministic short-term runoff forecasts was found to be limited. For the radar-based predictions, the reduction of variance for lead times of 1, 2 and 3 hours was 0.61, 0.62 and 0.54, respectively, with reference to a “no-forecast” scenario, i.e. persistence. The probability of detection for major increases in streamflow was typically less than 0.5. Given the significance of flood events in the Marikina Basin, more effort needs to be put into the reduction of forecast errors and the quantification of remaining uncertainties.     

Evaluation of nonlinear models for precipitation forecasting

Prediction of factors affecting water resources systems is important for their design and operation. In hydrology, wavelet analysis (WA) is known as a new method for time series analysis. In this study, WA was combined with an artificial neural network (ANN) for prediction of precipitation at Varayeneh station, western Iran. The results obtained were compared with the adaptive neural fuzzy inference system (ANFIS) and ANN. Moreover, data on relative humidity and temperature were employed in addition to rainfall data to examine their influence on precipitation forecasting. Overall, this study concluded that the hybrid WANN model outperformed the other models in the estimation of maxima and minima, and is the best at forecasting precipitation. Furthermore, training and transfer functions are recommended for similar studies of precipitation forecasting.     

Longitudinal variations of phytoplankton compositions in lake-to-river systems

To test the hypothesis of longitudinal variations in phytoplankton compositions from a eutrophic lake to its river downstream and determine the length of the transition zone, we applied functional groups as well as taxonomical methods to this coupled aquatic system, which is composed of the Dianchi Lake upstream and the Tanglang River downstream, by sampling at 9 stations during Microcystis blooms in the Dianchi Lake in 2013. The longitudinal variations in phytoplankton compositions from lacustrine species to fluvial species were reflected by: (1) the shift from Microcystis to Chlorococcales green algae and centric diatoms; (2) the shift from the dominance of codon M to the coexistence of a variety of coda without one outstanding codon; and (3) except for codon M, the shift from lacustrine coda (H1, L_O, T) towards coda that are adapted to both lacustrine and fluvial circumstances (MP, X1, X2). The prominent difference of phytoplankton compositions between the Dianchi Lake and the lower reaches of the Tanglang River revealed that there was a transition zone in between. The upper and middle reaches of the Tanglang River with a length of approximately 26.4 km were considered the transition zone because: (1) the dominant lentic codon M in the Dianchi Lake disappeared at the lower reaches of the river; (2) the amount of codon P that is sensitive to stratification rose at the beginning of the river; and (3) the codon T, which is well adapted to the persistently mixed layer or epilimnia of lakes, lost a large number of biomass at the upper and middle reaches of the Tanglang River. In this study, we found that the eutrophic lake had a significant influence on the river downstream. In addition, we found that functional groups were sensitive to the changes of external aquatic conditions and helpful in determining the length of the transition zone.     

Macroinvertebrate community structure and feeding interactions along a pollution gradient in Gilgel Gibe watershed,Ethiopia: Implications for biomonitoring

Feeding interactions among functional feeding groups (FFGs) of macroinvertebrates are robust indicators of aquatic ecosystem interactions. They provide information regarding organic matter processing, habitat condition and trophic dynamics. In tropical rivers with pronounced wet and dry seasons, macroinvertebrate based ecological monitoring tools are explicitly focused on metrics and indices, while ignoring interactions of FFGs. Therefore, the objective of this study was to investigate the functional feeding type metrics, diversity indices and feeding interactions among FFGs of macroinvertebrates along the water pollution gradient in Gilgel Gibe watershed, Ethiopia. Water quality parameters and macroinvertebrate community attributes were assessed for samples collected from upstream sites (15 sites), urban-impacted stretches (12 sites) and wetland-affected river zones (7 sites) of the watershed during the rainy (July) and dry (February) seasons. To understand the effect of pollution on the feeding interactions, stable carbon and nitrogen isotopes were analyzed. Macroinvertebrate-based diversity indices and functional feeding type metric showed deterioration of ecological integrity at the urban-impacted sites and substantial recovery in the wetland-affected downstream sites. Omnivorous feeding behavior of macroinvertebrates was noted for the upstream sites, whereas clear trophic guilds of FFGs were suggested for the wetland-affected river zones by the stable isotope results. The results of pollution gradient analysis and feeding interactions among FFGs revealed that the urban-impacted sites showed weaker interactions when compared to upstream and wetland influenced sites. This affirms the potential importance of feeding interactions among FFGs of macroinvertebrates in water quality monitoring.     

Rapid adjustment of shoreline behavior to changing seasonality of storms: observations and modelling at an open‐coast beach

An 8‐year time series of weekly shoreline data collected at the Gold Coast, Australia, is used to examine the temporal evolution of a beach, focusing on the frequency response of the shoreline to time‐varying wave height and period. Intriguingly, during 2005 the movement of the shoreline at this site changed from a seasonally‐dominated mode (annual cycle) to a storm‐dominated (~monthly) mode. This unexpected observation provides the opportunity to explore the drivers of the observed shoreline response. Utilizing the calibration of an equilibrium shoreline model to explore the time‐scales of underlying beach behavior, the best‐fit frequency response (days^?1) is shown to be an order of magnitude higher post‐2004, suggesting that a relatively subtle change in wave forcing can drive a significant change in shoreline response. Analysis of available wave data reveals a statistically significant change in the seasonality of storms, from predominantly occurring at the start of the year pre‐2005 to being relatively consistent throughout the year after this time. The observed change from one mode of shoreline variability to another suggests that beaches can adapt relatively quickly to subtle changes in the intra‐annual distribution of wave energy. Copyright © 2016 John Wiley & Sons, Ltd.     

基于LBS和深度学习的旅游景区客流量的高时频预测

为实现精准的旅游景区客流量的高时频预测,本研究构建了一套基于LBS和深度学习模型的预测方法。此方法可通过对LBS数据的转换实现预测的空间范围与时频控制,并通过方法的核心模型——基于双向循环神经网络和GRU算法构建的深度双向GRU(DBi-GRU)模型完成预测。为检验方法的有效性,研究以深圳大梅沙海滨公园为例对方法进行实验测试。实验使用拟合曲线、误差指标及DM检验3种方法评估DBi-GRU模型的预测效果。此外,实验还设置了其他五种深度学习模型作为DBi-GRU的对照模型,测试基于不同深度学习算法的模型之间的预测水平差异。实验结果表明:(1)本研究提出的DBi-GRU模型在景区客流量高时频预测中具有理想的预测效果,在高峰时段的客流量预测方面也具有较高准确性,预测效果明显优于其他深度学习模型;(2)基于双向循环网络的模型的效果普遍优于基于常规循环网络的模型。尤其是基于双向LSTM算法的模型,虽然预测的准确度略逊色于DBi-GRU模型,但在模型性能上与其的差异并不显著;(3)在相同网络参数下,GRU算法较前人采用的LSTM和RNN算法有着更高的预测准确性。本研究为客流量预测领域的研究提供了一种...     

Location and probability of shoal margin collapses in a sandy estuary

Channel bank failure, and collapses of shoal margins and beaches due to flow slides, have been recorded in Dutch estuaries for the past 200 years but have hardly been recognized elsewhere. Current predictions lack forecasting capabilities, because they were validated and calibrated for historic data of cross‐sections in specific systems, allowing local hindcast rather than location and probability forecasting. The objectives of this study were to investigate where on shoal margins collapses typically occur and what shoal margin collapse geometries and volumes are, such that we can predict their occurrence. We identified shoal margin collapses, generally completely submerged, from bathymetry data by analyzing digital elevation models of difference of the Western Scheldt for the period 1959–2015. We used the bathymetry data to determine the conditions for occurrence, specifically to obtain slope height and angle, and applied these variables in a shoal margin collapse predictor. We found 299 collapses along 300 km of shoal margin boundaries over 56 years, meaning that more than five collapses occur on average per year. The average shoal margin collapse body is well approximated by a 1/3 ellipsoid shape, covers on average an area of 34 000 m² and has an average volume of 100 000 m³. Shoal margin collapses occur mainly at locations where shoals take up a proportionally larger area than average in the cross‐section of the entire estuary, and occur most frequently where lateral shoal margin displacement is low. A receiver operating characteristic curve shows that the forecasting method predicts the shoal margin collapse location well. We conclude that the locations of the shoal margin collapses are well predicted by the variation in conditions of the relative slope height and angle within the Western Scheldt, and likely locations are at laterally relatively stable shoal margins. This provides hypotheses aiding the recognition of these features in sandy estuaries worldwide. Copyright © 2018 John Wiley & Sons, Ltd.