Near-term forecasts of stream temperature using deep learning and data assimilation in support of management decisions

Deep learning (DL) models are increasingly used to make accurate hindcasts of management-relevant variables, but they are less commonly used in forecasting applications. Data assimilation (DA) can be used for forecasts to leverage real-time observations, where the difference between model predictions and observations today is used to adjust the model to make better predictions tomorrow. In this use case, we developed a process-guided DL and DA approach to make 7-day probabilistic forecasts of daily maximum water temperature in the Delaware River Basin in support of water management decisions. Our modeling system produced forecasts of daily maximum water temperature with an average root mean squared error (RMSE) from 1.1 to 1.4°C for 1-day-ahead and 1.4 to 1.9°C for 7-day-ahead forecasts across all sites. The DA algorithm marginally improved forecast performance when compared with forecasts produced using the process-guided DL model alone (0%–14% lower RMSE with the DA algorithm). Across all sites and lead times, 65%–82% of observations were within 90% forecast confidence intervals, which allowed managers to anticipate probability of exceedances of ecologically relevant thresholds and aid in decisions about releasing reservoir water downstream. The flexibility of DL models shows promise for forecasting other important environmental variables and aid in decision-making.     

A multidecadal oscillation in precipitation and temperature series is pronounced in low flow series from Puget Sound streams

In Pacific Northwest streams, summer low flows limit water available to competing instream (salmon) and out-of-stream (human) uses, creating broad interest in how and why low flows are trending. Analyses that assumed linear (monotonic) change over the last ~60 years revealed declining low flow trends in minimally disturbed streams. Here, polynomials were used to model flow trends between 1929 and 2015. A multidecadal oscillation was observed in flows, which increased initially from the 1930s until the 1950s, declined until the 1990s, and then increased again. A similar oscillation was detected in precipitation series, and opposing oscillations in surface temperature, Pacific Decadal Oscillation, and Interdecadal Pacific Oscillation series. Multidecadal oscillations with similar periods to those described here are well known in climate indices. Fitted model terms were consistent with flow trends being influenced by at least two drivers, one oscillating and the other monotonic. Anthropogenic warming is a candidate driver for the monotonic decline, and variation in (internal) climatic circulation for the oscillating trend, but others were not ruled out. The recent upturn in streamflows suggests that anthropogenic warming has not been the dominant factor driving streamflow trends, at least until 2015. Climate projections based on simulations that omit drivers of multidecadal variation are likely to underestimate the range, and rate of change, of future climatic variation.     

A new approach to spatial source apportionment of haze pollution in large scale and its application in China

Long-lasting expansion of haze pollution in China has already presented a stern challenge to regional joint prevention and control. There is an urgent need to enlarge and reconstruct the coverage of joint prevention and control of air pollution in key area. Air quality models can identify and quantify the regional contribution of haze pollution and its key components with the help of numerical simulation, but it is difficult to be applied to larger spatial scale due to the complexity of model parameters. The time series analysis can recognize the existence of spatial interaction of haze pollution between cities, but it has not yet been used to further identify the spatial sources of haze pollution in large scale. Using econometric framework of time series analysis, this paper developed a new approach to perform spatial source apportionment. We applied this approach to calculate the contribution from spatial sources of haze pollution in China, using the monitoring data of particulate matter (PM_2.5) across 161 Chinese cities. This approach overcame the limitation of numerical simulation that the model complexity increases at excess with the expansion of sample range, and could effectively deal with severe large-scale haze episodes.     

屏蔽服冷却系统对人体表面温度影响的研究*

为科学有效地论证多种方式组合的冷却屏蔽服在不同环境条件下对人体表面温度控制的效果，需要对冷却系统及人体敏感部位发热量进行客观评估。通过对5名健康男性的高温测试，探究人体在不同环境温度下体表温度的变化，得出胸部、背部及额头为热量最高部位，并构建以“人体-降温屏蔽服-外界环境”为主体的冷却系统数值模型，对不同环境中的屏蔽服冷却效果展开研究，分析穿戴冷却屏蔽服时人体躯干部分的温度分布及影响。结果表明:在屏蔽服中靠近胸部、背部部位引入相变材料和风扇，均可帮助人体降低体温，提高舒适度。     

Thermal decomposition mechanism of 65% lysine sulfate powder and its thermal stability based on thermal analysis

Lysine is widely used in the fields of food, medicine and feed, which generally appears in the form of lysine sulfate or lysine hydrochloride dust because of the high instability of the free L-lysine. The L-lysine Sulfate is in high risk of decomposition, spontaneous ignition and even the dust explosion, because the control temperature in its production process is high up to 90 °C. Thus, the thermal behaviors and its thermal stability of 65% lysine sulfate are experimentally explored in Air and Nitrogen using the simultaneous TG-DSC measurements. Results show: (1) the decomposition of 65% lysine sulfate can be divided into three stages both in the atmospheres of air and nitrogen, and most of the weight loss occurred in the first two stages, which are related with the decarboxylation and deamination process. (2) The effects of atmosphere on the decomposition of 65% lysine sulfate mainly occur at the third stage. In this stage, the weight loss in nitrogen is only 14.2%, which is much lower than that in air (34.3%), which is related to the oxidative degradation at high temperature. Besides, the active energy is slightly increased in nitrogen compared to that in air. (3) The initial temperatures of the decomposition of the 65% lysine sulfate are 145 °C and 155 °C, for the air and nitrogen atmosphere, respectively, which are much lower than that (260 °C) of the pure lysine.     

Development of machine learning based prediction models for hazardous properties of chemical mixtures

Lower flammability limit (LFL), upper flammability limit (UFL), auto-ignition temperature (AIT) and flash point (FP) are crucial hazardous properties for fire and explosion hazards assessment and consequence analysis. In this study, a comprehensive prediction model set was constructed by using expanded chemical mixture databases of chemical mixture hazardous properties. Machine learning based gradient boosting quantitative structure-property relationship (GB-QSPR) method is implemented for the first time to improve the model performance and prediction accuracy. The result shows that all developed models have significantly higher accuracy than other regular QSPR models, with the 5-fold cross-validation RMSE of LFL, UFL, AIT, and FP models being 1.06, 1.14, 1.08, and 1.17, respectively. All developed QSPR models can be used to estimate reliable chemical mixture hazardous properties and provide useful guidance in chemical mixture hazard assessment and consequence analysis.     

肥煤自燃全过程热动力学特征参数研究

实验测定了林西矿肥煤样品30~900℃煤自燃全过程热动力学特征参数,得出:TG/DTG曲线显示煤样DTG初始临界温度45℃,干裂温度122℃,活性温度195℃,增速温度265℃,质量极大值温度342℃,着火温度465℃,最大热失重速率温度515℃和燃尽温度690℃;DSC曲线显示,煤样初始放热温度60℃、最大热释放速率温度511℃。结合TG-DTG-DSC曲线综合分析可知,煤温达到510℃左右时煤样反应最剧烈。由煤自燃标志气体测定实验系统得出:煤温130℃后CO,CO 2释放量迅速增加,210℃增加速度下降;CH 4,C 2 H 6含量变化具有规律性且两者变化相近;C 2 H 4出现温度为130℃;C 2 H 4/C 2 H 6比值在190~350℃有较强的规律性,呈上升趋势且上升速度较快;350℃之后,CH 4,C 2 H 6,C 2 H 4体积分数均开始急剧增大;C 2 H 4/CO与C 2 H 4/CO 2变化趋势大致相同,在130~350℃时缓慢增长,达到350℃后比值呈指数形式上升。经拟合曲线,得到活化能的3个突变点温度:70,180,220℃,其中180℃与交叉点温度相吻合。通过以上研究,得到了肥煤自燃全过程的热力学特征参数,为实际生产中防治煤自燃提供了理论依据。     

基于Landsat 8遥感数据的天津市地表温度反演

以Landsat 8遥感数据为数据源,进行天津市地表温度反演研究。首先采用单通道算法反演地表温度,并利用均值标准差法进行温度分级。然后建立不同温度等级面积比例的估算模型。再通过随机样点,从不同温度等级和土地覆盖类型2个角度,分别建立并比较不同类样点的地表温度与各指数的拟合模型。结果表明:次高温区域面积比例与人口密度、人均GDP都具有较高的决定系数;地表温度与NDVI、BAEM的二元线性回归决定系数高于地表温度与单一指数的决定系数;将样点分类后,低温点与MNDBI的决定系数高于其他温度等级样点,水域和植被样点与各种指数的决定系数高于其他地物类型样点。     

初始温度和压力对排污空间甲烷-空气混合物爆燃特性影响的模拟研究

市政排污空间作为城市公共基础设施的重要组成部分，易积聚可燃气体形成爆炸性环境。结合排污空间的特殊环境条件，采用Fluidyn-MP多物理场数值模拟软件，建立了20 L球形爆炸罐分析模型，通过改变初始温度和初始压力，对排污空间甲烷-空气混合物爆燃特性及其变化规律进行模拟研究。结果表明:初始温度升高导致甲烷-空气混合物最大爆炸压力降低，缩短了到达最大爆炸压力的时间；初始压力增加导致最大爆炸压力急剧升高，并延长了到达最大爆炸压力的时间；最大爆炸压力对初始压力的敏感程度远大于初始温度的影响。此外，随着初始温度和初始压力的升高，爆炸火焰平均传播速度增加，而火焰传播速度对初始温度较敏感。     

城市污泥-磷石膏陶粒的制备条件及性能分析

以城市污水处理厂剩余污泥和磷酸生产废渣磷石膏为原料制备多孔陶粒,考察不同配比、烧结温度及烧结时间对多孔陶粒堆积密度、吸水率及盐酸可溶率的影响。结果表明:在污泥与磷石膏混合比1∶4、烧结温度1 050℃、烧结时间15 min的条件下,可制得堆积密度为685.78 kg/m~3、吸水率27.34%、盐酸可溶率11.38%的陶粒,该陶粒内部含有丰富的膨胀气孔,可用作水处理滤料或建材骨料。