点源时间序列数据缺失值的估值方法比较——以小流域气象和水文数据为例

对点源时间序列数据缺失值进行有效估值能提升其数据质量。为探究不同估值方法对点源时间序列数据缺失值的估值效果及其影响因素,以亚热带典型小流域长期定位观测的每日气象和水文数据（最高气温、最低气温、太阳辐射量、降雨量及地表径流量）为例,以均方根误差（RMSE）、绝对平均误差（MAE）和Pearson相关系数（r）为性能验证指标,比较了线性内插法（LIM）、K-最近邻插值法（KNNM）、样条插值法（SIM）、多项式插值法（PIM）和核密度估值法（KDEM）5种估值方法的估值性能差异及其主要影响因素。结果表明：（1）LIM、SIM和KDEM的估值性能总体上优于其它2种方法;（2）5种估值方法对气象数据（最高气温、最低气温和太阳辐射量）缺失值估值的RMSE为1.81～6.35,MAE为1.30～4.20,r为0.70～0.98（P＜0.05）,而对水文数据（降雨量和地表径流量）缺失值估值的RMSE为12.54～26.28,MAE为3.60～14.21,r为0.07～0.72。可见,各估值方法对气象数据的估值性能强于对水文数据;（3）上述数据集的变异系数（CV）与估值评估指标（RMSE、MAE及r）线性相关（P＜0.05）,是影响估值性能的重要因素。     

水稻格点作物模型在中国区域的不确定性评估

作物模型是评估气候变化对农业生产影响的主要手段之一,但中国对格点作物模型间的比较研究尚处于初始阶段。为全面评估不同作物模型在中国不同区域对水稻产量模拟的有效性,利用联合国粮农组织（FAO）和中国农业农村部种植业管理司（SYB）水稻年平均产量统计资料,对由2种气候资料（AgMERRA和WFDEI-GPCC）和 3种不同种植管理情景（Default、Fullharm和Harmnon情景）驱动的7种全球格点作物模型（CGMS-WOFOST、CLM-CROP、EPIC-BOKU、GEPIC、LPJML、PDSSAT和PEPIC模型）模拟的中国水稻产量进行了对比分析。结果表明：不同格点作物模型之间的模拟结果差异较大,在不同区域不同格点作物模型的模拟效果差异显著,不同格点作物模型对气候变化和种植管理情景的响应和敏感性不同,大部分模拟结果低估了水稻产量。使用不同水稻统计产量数据会对评估结果产生一定的影响。格点作物模型能够一定程度上模拟出水稻产量的年际变化和气候变化对产量的影响,但对于统计水稻产量上升的趋势较难模拟。通过综合分析产量在时间和空间上的波动情况,并利用2种评分方法对模拟表现打分,发现LPJML和PDSSAT在7种格点作物模型中模拟效果最好,同时也对不同气候数据和种植管理情景的变化最敏感,CLM-CROP的模拟效果最差。对不同种植管理情景,Default情景下的模拟效果显著高于Fullharm和Harmnon情景。多种格点作物模型集合平均（MME）可以降低单个格点作物模型模拟的误差,但需对MME中的集合模型进行挑选。     

基于MODIS的光化学植被指数反演水稻区域光能利用率

为了探究光化学反射植被指数（Photochemical Reflectance Index, PRI）在反演大面积水稻光能利用率（Light Use Efficiency, LUE）方面的潜力,利用安徽寿县国家气候观象台通量观测数据和同期MODIS卫星数据,分别构建不同参照波段下（488、551和667nm）的PRIs与站点像元LUE间的模型,并尝试利用最优模型反演区域尺度的水稻光能利用率。结果表明,在所测试的参照波段中,来自后向散射方向影像提取的PRIs与LUE的相关性更强,其中PRI₅₅₁表现最优;与MOD17算法相比,利用PRI₅₅₁-LUE模型反演的区域LUE值更符合实际,且空间分布差异明显,在多时相影像和完善数据插补方法的基础上,利用MODIS PRI在反演长时间、大面积植被的光能利用率或生产力方面具有可行性。     

日光温室内气温观测的适宜频次初探

根据河北省高邑县2008-2010年温室作物生长季的外界气温特征,将2008、2009和2010年度分别定为偏暖、偏冷和正常年型。以高邑县2008-2010年温室作物生长季采集周期为10min的小气候气温观测资料为基础,将一日内每10min的气温资料的算术平均值作为实际日平均气温,分别取每20min、30min、1h、2h、3h、4h的观测数据进行算术平均计算日平均气温,分析不同年型下计算结果与实际值之间的平均绝对误差MAE、标准差SD和积温相对误差R,并采用成对t检验进行显著性检验;以P＜0.05作为显著性检验指标,确定日光温室气温观测的适宜方法和频次。将MAE、SD和R作为评价计算方法优劣的主要因子,MAE、SD和R越小,表示计算效果越好。结果表明：（1）当观测间隔小于4h时,所计算的日平均气温与实际值无显著差异,MAE≤0.19℃,SD≤0.24℃,R≤1.1%,观测数据能满足温室环境日平均气温的计算,且观测间隔时间越短,观测频次越高,检验效果越好;观测间隔达到4h时,计算结果的准确性存在一定风险;（2）6个每日4时次的组合所计算的日平均气温与实际值存在显著差异,观测试验中不宜采用;（3）用日极值法计算的日平均气温较实际值显著偏高,不能代表实际气温,观测试验中亦不宜采用;（4）采用改进后的7：00、10：00、13：00、16：00、19：00和23：00气温组合计算的日平均气温与实际值无显著差异,MAE≤0.24℃,SD≤0.30℃,R≤1.4%,计算结果可以代表实际气温,且减少了夜间观测次数,观测时间更符合工作生活习惯,能够在实际生产中应用。总体来讲,每日观测应不少于6次,所计算的日平均气温才可能代表实际日平均气温。     

MyoG基因型对二花脸猪早期生长性状及肌肉组织学特性的影响

肌细胞生成素（myogenin,MyoG)是骨骼肌转录因子MyoG家族的一员，在肌细胞的形成过程中起着中心调控作用。利用21窝不同日龄的二花脸猪(Sus scrofa Erhua）共97头，根据日龄组分布分成1组（日龄阶段）：49d，2组：58－60d,3组：61－70d，4组：71－80d，5组：81－90 d，6组：228d,采用索氏脂肪抽提法和石蜡组织切片－图像分析法对其背景长肌及半腱肌的肌肉组织学特性进行了测定。根据MyoG基因PCR－RFLP多态性，对二花脸猪不同的MyoG基因型与相关性状即初生重、20日龄体重，断奶重及肌肉组织学特征的关系进行了分析。结果表明：（1）在二花脸猪不同的MyoG基因型间，初生重差异极显著（P＜0．01），但20日龄体重，断奶重无显著差异；（2）背最长肌的干肉样中的粗脂肪百分含量，肌纤维面积，肌纤维面密度、脂肪面密度在1－6日龄组内不同的MyoG基因型间差异均不显著，平均直径在1－5组增表现为AB＜BB，4组组内差异显著（P＜0．05），6组表现为AB＞BB；肌纤维密度在1－5组表现为AB＞BB，4组组内差异显著（P＜0．05），6组表现为AB＜BB；（3）半腱肌中除6组肌纤维密度表现为AB＜BB差异显著（P＜0．05），外其余性状表现为组内基因型间差异不显著。     

2001-2100年中国区域季节平均温度变化的时空格局

基于1km分辨率长时间序列温度数据集,采用距平法、Mann-Kendall趋势检验法和Sen’s斜率估计法，分析四季平均气温在历史时期（2001−2020年）与未来时期（2021−2100年）低强迫情景（SSP119）、中等强迫情景（SSP245）和高强迫情景（SSP585）下的变化幅度和变化趋势的时空格局，以期为气候变暖背景下制定详细的区域适应性策略提供依据。结果表明：（1）相比历史时期，未来时期在3个情景下的四季均温总体上升，且夏季增温区域面积最大，其中SSP119情景下增温1～2℃的区域占66.70%，SSP245和SSP585情景下增加2℃以上的区域分别占37.37%和99.06%；同时，3个SSP情景下的季节均温的整体变化幅度具有显著差异，SSP119情景下较缓和，SSP245情景次之，SSP585情景增温幅度最大。（2）在历史时期，相比其他季节，春季均温的显著上升速率最快（0.68±0.24℃∙10a⁻¹），且面积占比最大（14.44%），主要分布于华北、云贵川和江浙局部区域。（3）在未来时期，中国区域季节均温呈总体上升趋势，且具有显著的空间差异；其中，在SSP119情景下，春季和冬季均温显著上升的区域主要集中于中国南部和青藏高原局部区域，面积占比分别为29.03%和25.58%，在SSP245和SSP585情景下，中国所有区域的季节均温呈显著上升趋势；在SSP585情景下，北方的季节均温显著上升速率比南方快，全国区域在冬季的显著上升速率最快（0.66±0.09℃∙10a⁻¹）。     

基于被动微波遥感的中国干旱动态监测

目前用于中国干旱监测的遥感方法大多是可见光和热红外指数法,受云雨、植被和地形的影响较大,不能满足中国南方地区干旱监测的需求。该研究基于被动微波辐射传输方程,首先构建了基于AMSR-E（advanced microwave scanning radiometer-EOS）数据的地表温度反演模型,R2=0.79,RMSE（root mean square error）为2.54℃,实现了中国地表温度的被动微波遥感监测。然后,拟合了不同下垫面归一化植被指数（normalized difference vegetation index,NDVI）与微波极化差异指数（microwave polarization difference index,MPDI）的关系。在此基础上改进了植被供水指数（vegetation supply water index,VSWI）,构建了基于AMSR-E数据的被动微波遥感气象干旱指数,并以中国2009年的旱情为例进行实例验证。研究表明,该干旱指数与AMSR-E L3土壤湿度数据有着显著的负相关关系（R2=0.75）,且能基本表征2009年中国实际的气象干旱状况。     

TGIS及其在水土保持研究中的应用

介绍了动态地理信息系统（TGIS）的概念及研究进展，提出和界定了与TGIS相关的几个重要术语（包括比例尺、尺度、分辨率）；对TGIS发展概况，TGIS时空特征和TGIS数据构建方法，TGIS在水土保持研究中应用的几个主要方面——土地利用时空动态分析、试验观测数据表面模型建立与分析等进行了讨论。提出了TGIS研究中几个重要问题，包括：时间空间精度评估和其不确定性，元数据标准等。     

不同添加物对分离昆明系小鼠胚胎干细胞的影响研究

本试验以昆明系小鼠胚胎为材料,以丝裂霉素（10μg/mL）处理的MEF为饲养层,研究了在ESCs培养液中分别添加KSR、FBS、FBS+PD98059 (50 μmol/L)对昆明系小鼠胚胎贴壁、ICM集落形成及ESCs分离培养的影响。结果表明,在添加KSR的ESCs培养液中,胚胎贴壁率显著低于添加FBS的ESCs培养液（P＜0.05）,但ICM集落形成率和1代ESCs集落出现率差异不显著（P>0.05）,2~5代ESCs集落出现率显著高于添加FBS的ESCs培养液（P＜0.05）,2株ESCs被传到7代;在添加PD98059+FBS的ESCs培养液中,胚胎贴壁率、ICM集落形成率和1~5代ESCs集落出现率均显著低于添加KSR或FBS的ESCs培养液（P＜0.05）;用0.5 g/L胰酶+0.2g/L EDTA离散消化ICM细胞和ESCs并结合机械分割,1~5代ESCs集落出现率显著高于用2.5 g/L+0.2 g/L EDTA（P＜0.05）。结论：在ESCs培养液中添加KSR,较添加FBS或FBS+PD98059更适合用于分离培养昆明系小鼠ESCs,用0.5g/L胰酶+0.2g/L EDTA离散消化ICM细胞和ESCs并结合机械分割优于用2.5 g/L+0.2 g/L EDTA。     

华北平原干旱事件特征及农业用地暴露度演变分析

根据1961-2014年华北平原52个气象观测站月降水数据和区域气候模式COSMO-CLM（CCLM）输出的逐月降水预估数据,利用标准化降水指数,结合“强度-面积-持续时间”（Intensity-Area-Duration, IAD）方法,研究了华北平原过去（1961-2014年）和未来（2016-2050年）3种排放情景(RCP2.6、4.5、8.5）下,不同持续时间的区域最强干旱事件的强度-面积特征及其时空分布规律。同时,基于2000年的土地利用数据,分析了2016-2050年华北平原农业用地暴露度的演变。研究表明：（1）1961-2014年,华北平原干旱中心在空间上呈由南向北迁移的趋势。（2）相比基准期（1961-2005年）,过去45a未遇的干旱事件在2016-2050年RCP3种情景下均有可能发生;RCP2.6情景下发生频率最高。（3）2016-2050年,RCP2.6和RCP 4.5情景下,华北平原农业用地干旱暴露度（即暴露面积）呈增大趋势,RCP4.5情景下干旱暴露面积增加的速率更大,RCP8.5情景下则与之相反,呈减小趋势。3种情景下暴露度峰值分别出现在2040s后期,2040s前期及2020s中期。     

A comparison of multiple phenology data sources for estimating seasonal transitions in deciduous forest carbon exchange

There are currently numerous data sources available for estimating the timing of recurrent plant phenology transitions. We compared measurements from several phenology data sources to understand the relationship between phenology metrics derived from these data sources and the timing of seasonal transitions in net ecosystem exchange (NEE). We identified the timing of start, peak, end and the duration of the carbon uptake season, as well as the timing of the transitions from sink to source and source to sink using 11 years of NEE data from the University of Michigan Biological Station (UMBS). Using fitted logistic functions we identified proxy metrics for phenological transitions from the time series of Albedo, fraction of absorbed photosynthetically active radiation (fPAR), Plant Area Index (PAI), and MODIS normalized difference vegetation index (NDVI), enhanced vegetation index (EVI), and leaf area index (LAI) products of various spatial representations. We found that no single source of phenological data was able to accurately describe annual patterns of flux phenology. However, for each transition in NEE (e.g., start of season, transition to net sink), the metrics from one or more data sources were significantly (p < 0.05) correlated with the timing of these recurring events. A marginally significant trend toward a longer NEE carbon uptake period over 11 years was not detected by any of the metrics, primarily because none of the metrics were available for the full duration of the NEE data, and NEE did not show significant and consistent trends during the sub-sets of the time when proxy data were available. The results of our study highlight the relative strengths and weaknesses of each phenology data source for directly estimating seasonal transitions and interannual trends in carbon flux phenology of a deciduous forest.     

应用反射光谱法估测矿区附近农田土壤As, Cu污染

Concentrations of Iron (Fe), As, and Cu in soil samples from the fields near the Baoshan Mine in Hunan Province, China, were analyzed and soil spectral reflectance was measured with an ASD FieldSpec FR spectroradiometer (Analytical Spectral Devices, Inc., USA) under laboratory condition. Partial least square regression (PLSR) models were constructed for predicting soil metal concentrations. The data pre-processing methods, first and second derivatives (FD and SD), baseline correction (BC), standard normal variate (SNV), multiplicative scatter correction (MSC), and continuum removal (CR), were used for the spectral reflectance data pretreatments. Then, the prediction results were evaluated by relative root mean square error (RRMSE) and coefficients of determination (R²). According to the criteria of minimal RRMSE and maximal R², the PLSR models with the FD pretreatment (RRMSE = 0.24, R² = 0.61), SNV pretreatment (RRMSE = 0.08, R² = 0.78), and BC-pretreatment (RRMSE = 0.20, R² = 0.41) were considered as the final models for predicting As, Fe, and Cu, respectively. Wavebands at around 460, 1 400, 1 900, and 2 200 nm were selected as important spectral variables to construct final models. In conclusion, concentrations of heavy metals in contaminated soils could be indirectly assessed by soil spectra according to the correlation between the spectrally featureless components and Fe; therefore, spectral reflectance would be an alternative tool for monitoring soil heavy metals contamination.     

The REFLEX project: Comparing different algorithms and implementations for the inversion of a terrestrial ecosystem model against eddy covariance data

We describe a model-data fusion (MDF) inter-comparison project (REFLEX), which compared various algorithms for estimating carbon (C) model parameters consistent with both measured carbon fluxes and states and a simple C model. Participants were provided with the model and with both synthetic net ecosystem exchange (NEE) of CO₂ and leaf area index (LAI) data, generated from the model with added noise, and observed NEE and LAI data from two eddy covariance sites. Participants endeavoured to estimate model parameters and states consistent with the model for all cases over the two years for which data were provided, and generate predictions for one additional year without observations. Nine participants contributed results using Metropolis algorithms, Kalman filters and a genetic algorithm. For the synthetic data case, parameter estimates compared well with the true values. The results of the analyses indicated that parameters linked directly to gross primary production (GPP) and ecosystem respiration, such as those related to foliage allocation and turnover, or temperature sensitivity of heterotrophic respiration, were best constrained and characterised. Poorly estimated parameters were those related to the allocation to and turnover of fine root/wood pools. Estimates of confidence intervals varied among algorithms, but several algorithms successfully located the true values of annual fluxes from synthetic experiments within relatively narrow 90% confidence intervals, achieving >80% success rate and mean NEE confidence intervals <110 gC m⁻² year⁻¹ for the synthetic case. Annual C flux estimates generated by participants generally agreed with gap-filling approaches using half-hourly data. The estimation of ecosystem respiration and GPP through MDF agreed well with outputs from partitioning studies using half-hourly data. Confidence limits on annual NEE increased by an average of 88% in the prediction year compared to the previous year, when data were available. Confidence intervals on annual NEE increased by 30% when observed data were used instead of synthetic data, reflecting and quantifying the addition of model error. Finally, our analyses indicated that incorporating additional constraints, using data on C pools (wood, soil and fine roots) would help to reduce uncertainties for model parameters poorly served by eddy covariance data.     

基于不同PSO-ELM模型的碾压黏土抗剪强度预测方法研究

碾压黏土的抗剪强度直接影响碾压土石坝的质量和使用寿命。为得出碾压黏土抗剪强度的最优预测模型,通过粒子群算法优化极限学习机模型(PSO-ELM),分别以Sine函数、radbas函数和hardlim函数3种激活函数为基础,构建PSO-ELM_sin,PSO-ELM_rad和PSO-ELM_hard3种模型,并将模型结果与ELM模型、广义回归神经网络模型(GRNN)、随机森林模型(RF)和BP神经网络模型进行了对比。结果表明:在黏聚力和内摩擦角的拟合结果中,PSO-ELM_sin模型精度最高,其拟合方程斜率分别为1.005,1.032; 在月值模拟中,PSO-ELM_sin模型与实测值的拟合度最高,相对误差仅在6.0%～9.3%; PSO-ELM_sin模型在黏聚力模拟中RMSE,RRMSE,MAE,E_ns和R²分别为0.776 kPa,1.80%,0.641 kPa,0.993和0.997,该模型在内摩擦角模拟中RMSE,RRMSE,MAE,E_ns和R²分别为1.635°,6.98%,1.616°,0.983和0.998,模型精度均排名第一。因此,PSO-ELM_sin模型在所有模型中精度最高,可作为碾压黏土抗剪强度预测的标准模型使用。     

不同模型在渔业CPUE标准化中的比较分析

为了提高渔业数据单位捕捞努力量渔获量（catch per unite of effort,CPUE）标准化数据的质量和模型连续稳定预测能力,该文采用人工神经网络（artificial neural network,ANN）、回归树（regression trees,RT）、随机森林（random forest,RF）和支持向量机（support vector machine,SVM）等机器学习方法和传统的广义线性模型（generalized linear model,GLM）等方法,对2000－2013年大西洋大眼金枪鱼（Thunnus obesus）延绳钓CPUE数据进行标准化。采用平均绝对误差、平均均方误差、3种相关系数（Pearson's,Kendall's和Spearman's）和标准化均方误差等评价指标对不同模型标准化结果进行对比,寻找较优的标准化方法。研究结果表明,在验证数据集SVM方法得到的3种相关系数（0.596,0473和0.632）和RF（0.623,0.456,0.621）相似,高于RT(0.516,0.432和0.586)、ANN(0.428,0.249和0.365)和GLM(0.199,0.106和0.159)。SVM预测的均方误差（11.25）、平均绝对误差（2.107）和标准化均方误差（0.652）略低于RF（11.655,2.377和0.661）,明显低于RT（14.999,2.434和0.801）、ANN(16.692,2.883和0.823)和GLM（16.517,2.777和0.993）。各项指标揭示SVM方法要优于其他4种方法,RF次之,GLM计算结果在所有方法中最差,不适合渔业数据CPUE标准化。SVM和RF方法应该被优先考虑用于渔业数据CPUE标准化。研究结果为渔业资源管理和保护提供更好的支持。     

3种坡面流水深测量方法比较

[目的]为坡面流水深的快速、准确测量提供技术支撑。[方法]采用超声波系统测量变坡试验水槽不同坡度和流量条件下坡面流水深,同时用测针法和染色法对其进行平行测量,利用平均绝对误差(MAE),平均相对误差(MRE),相对均方差误差(RRMSE)和Nash—Sultcliffe系数(NSE)4个指标比较超声波法和染色法或测针法测量数据的接近程度。[结果]通过对3种方法测量结果的比较分析发现,不管以测针法还是以染色法的测量值为参照,超声波法测量的水深值与参照值更接近且相关程度更高。[结论]超声波法能快速有效地测量坡面流水深,用于侵蚀静床坡面流水动力学、土壤侵蚀机理等相关的研究工作。     

微波辅助提取葡萄皮中花色苷的动力学和热力学特性

为拓展微波辅助提取技术在食品工业领域的应用,研究了微波辅助技术对提取葡萄皮中花色苷的作用,并对其动力学和热力学进行了评价。结果表明：柠檬酸浓度、液料比、提取时间和微波功率可以显著影响总花色苷提取率。采用高浓度柠檬酸水溶液（1.00?mol/L）提取的总花色苷含量（172.99?mg/100?g）是低浓度柠檬酸水溶液（0.02?mol/L）提取的总花色苷含量（43.48?mg/100?g）的4倍。此外,在实验选取的动力学模型中,指数衰减模型可以更好的拟合提取动力学数据（R2＞0.9875）。微波辅助提取葡萄皮中花色苷的有效扩散系数（2.12～4.87×10-11?m2/s）随着微波功率的增加而增大。微波辅助提取过程的热力学参数表明微波提取方式有利于传质过程。所得研究结果将为微波辅助提取技术的工业化应用提供理论依据。     

CO2/heat fluxes in rice fields: Comparative assessment of flooded and non-flooded fields in the Philippines

The seasonal fluxes of heat, moisture and CO₂ were investigated under two different rice environments: flooded and aerobic soil conditions, using the eddy covariance technique during 2008 dry season. The fluxes were correlated with the microclimate prevalent in each location. This study was intended to monitor the environmental impact, in terms of C budget and heat exchange, of shifting from lowland rice production to aerobic rice cultivation as an alternative to maintain crop productivity under water scarcity.The aerobic rice fields had higher sensible heat flux (H) and lower latent heat flux (LE) compared to flooded fields. On seasonal average, aerobic rice fields had 48% more sensible heat flux while flooded rice fields had 20% more latent heat flux. Consequently, the aerobic rice fields had significantly higher Bowen ratio (0.25) than flooded fields (0.14), indicating that a larger proportion of the available net radiation was used for sensible heat transfer or for warming the surrounding air.The total C budget integrated over the cropping period showed that the net ecosystem exchange (NEE) in flooded rice fields was about three times higher than in aerobic fields while gross primary production (GPP) and ecosystem respiration (R_e) were 1.5 and 1.2 times higher, respectively. The high GPP of flooded rice ecosystem was evident because the photosynthetic capacity of lowland rice is naturally large. The R_e of flooded rice fields was also relatively high because it was enhanced by the high photosynthetic activities of lowland rice as manifested by larger above-ground plant biomass. The NEE, GPP, and R_e values for flooded rice fields were −258, 778, and 521 g C m⁻², respectively. For aerobic rice fields, values were −85, 515, and 430 g C m⁻² for NEE, GPP, and R_e, respectively. The ratio of R_e/GPP in flooded fields was 0.67 while it was 0.83 for aerobic rice fields.This short-term data showed significant differences in C budget and heat exchange between flooded and aerobic rice ecosystems. Further investigation is needed to clarify seasonal and inter-annual variations in microclimate, carbon and water budget of different rice production systems.