基于启发式Johnson算法优化BP神经网络的水产养殖产量预测模型

针对水产养殖产量预测难的现状,提出一种基于启发式Johnson算法优化的反向传播神经网络(BPNN)的产量预测模型。该模型在传统BP神经网络的基础上,针对网络训练时间长、易陷入局部最优的问题,通过启发式Johnson算法降低输入神经元维度,再结合试凑法确定神经网络隐层个数,构建启发式Johnson反向传播神经网络(HJA-BPNN)学习预测模型。实验结果表明,该模型在山东省对虾海水养殖产量预测中,预测的均方根误差小于传统BP神经网络和GM(1,1),且学习效率相比传统BP神经网络有所提升。研究表明,该学习预测模型在大量历史数据的模型构造上有更大的优势,能够缩短建模时间,同时获得良好的预测效果,为水产养殖产量预测提供了一种可行的新方法。     

Projections of temperature extremes based on preferred CMIP5 models:a case study in the Kaidu-Kongqi River basin in Northwest China

The extreme temperature has more outstanding impact on ecology and water resources in arid regions than the average temperature. Using the downscaled daily temperature data from 21 Coupled Model Inter-comparison Project(CMIP) models of NASA Earth Exchange Global Daily Downscaled Projections(NEX-GDDP) and the observation data, this paper analyzed the changes in temporal and spatiotemporal variation of temperature extremes, i.e., the maximum temperature(Tmax) and minimum temperature(Tmin), in the Kaidu-Kongqi River basin in Northwest China over the period 2020–2050 based on the evaluation of preferred Multi-Model Ensemble(MME). Results showed that the Partial Least Square ensemble mean participated by Preferred Models(PM-PLS) was better representing the temporal change and spatial distribution of temperature extremes during 1961–2005 and was chosen to project the future change. In 2020–2050, the increasing rate of Tmax(Tmin) under RCP(Representative Concentration Pathway) 8.5 will be 2.0(1.6) times that under RCP4.5, and that of Tmin will be larger than that of Tmax under each corresponding RCP. Tmin will keep contributing more to global warming than Tmax. The spatial distribution characteristics of Tmax and Tmin under the two RCPs will overall the same; but compared to the baseline period(1986–2005), the increments of Tmax and Tmin in plain area will be larger than those in mountainous area. With the emission concentration increased, however, the response of Tmax in mountainous area will be more sensitive than that in plain area, and that of Tmin will be equivalently sensitive in mountainous area and plain area. The impacts induced by Tmin will be universal and farreaching. Results of spatiotemporal variation of temperature extremes indicate that large increases in the magnitude of warming in the basin may occur in the future. The projections can provide the scientific basis for water and land plan management and disaster prevention and mitigation in the inland river basin.     

基于公共天气预报的三江平原ET0预报模型比较及敏感性分析

为了提出适合我国三江平原的高精度ET0预报方法,基于该区6个气象站点的天气预报数据和实测气象数据,以FAO56-Penman-Monteith(FAO56-PM)公式计算值为基准,比较Hargreaves-Samani(HS)、Thornthwaite(TH)和Blaney-Criddle(BC)3个ET₀预报模型的效果,对最优模型进行敏感性分析。结果表明:3个模型1~7 d预见期平均绝对误差均值分别为0.66、0.65、0.65 mm/d,均方根误差分别为0.93、0.96、0.95 mm/d,相关系数分别为0.857、0.828、0.840。1~5 d预见期最优预报模型为HS模型,6~7 d为TH模型。总体上预报精度由高到低为HS、TH、BC模型,建议采用HS模型在三江平原开展ET₀预报,HS模型预报对最高温预报的敏感性大于最低温。其预报值在夏季受温度预报误差影响最大,冬季最小,4季整体误差较小。研究可为灌溉预报提供较准确的数据基础。     

草地贪夜蛾缅甸虫源迁入中国的路径分析

草地贪夜蛾Spodoptera frugiperda(J.E. Smith)对非洲和南亚国家的入侵已对全球粮食安全造成重大影响,该虫2018年年底已在缅甸形成虫源基地,并零星进入中国云南西南部地区。本文利用历史数据分析了缅甸和华南地区春、夏两季(3-8月)925 hPa夜间平均风温场,并模拟预测了缅甸地区草地贪夜蛾在此期间进入中国的迁飞轨迹以及主要降落和波及的地区。结果表明:3-4月盛行的微弱西风不利于远距离迁移,但成虫的自主飞行可形成对云南和广西局部地区的近距离入侵;进入5月份后,随着西南夏季风的加强,云南和广西全境成为缅甸虫源的主要迁入地,并可能波及贵州、广东、海南和湖南等省。因此,4月份之前要重点监控云南和广西地区草地贪夜蛾的发生与为害,此后,应将监控区域扩大至中国中南部地区的各个省份。     

Improving cereal yield forecasts in Europe – The impact of weather extremes

The impact of extreme events (such as prolonged droughts, heat waves, cold shocks and frost) is poorly represented by most of the existing yield forecasting systems. Two new model-based approaches that account for the impact of extreme weather events on crop production are presented as a way to improve yield forecasts, both based on the Crop Growth Monitoring System (CGMS) of the European Commission. A first approach includes simple relations – consistent with the degree of complexity of the most generic crop simulators – to explicitly model the impact of these events on leaf development and yield formation. A second approach is a hybrid system which adds selected agro-climatic indicators (accounting for drought and cold/heat stress) to the previous one. The new proposed methods, together with the CGMS-standard approach and a system exclusively based on selected agro-climatic indicators, were evaluated in a comparative fashion for their forecasting reliability. The four systems were assessed for the main micro- and macro-thermal cereal crops grown in highly productive European countries. The workflow included the statistical post-processing of model outputs aggregated at national level with historical series (1995–2013) of official yields, followed by a cross-validation for forecasting events triggered at flowering, maturity and at an intermediate stage. With the system based on agro-climatic indicators, satisfactory performances were limited to microthermal crops grown in Mediterranean environments (i.e. crop production systems mainly driven by rainfall distribution). Compared to CGMS-standard system, the newly proposed approaches increased the forecasting reliability in 94% of the combinations crop × country × forecasting moment. In particular, the explicit simulation of the impact of extreme events explained a large part of the inter-annual variability (up to +44% for spring barley in Poland), while the addition of agro-climatic indicators to the workflow mostly added accuracy to an already satisfactory forecasting system.     

Can conservation tillage mitigate climate change impacts in Mediterranean cereal systems? A soil organic carbon assessment using long term experiments

Simulation models, informed and validated with datasets from long term experiments (LTEs), are considered useful tools to explore the effects of different management strategies on soil organic carbon (SOC) dynamics and evaluate suitable mitigative options for climate change. But, while there are several studies which assessed a better prediction of crop yields using an ensemble of models, no studies are currently available on the evaluation of a model ensemble on SOC stocks. In this study we assessed the advantages of using an ensemble of crop models (APSIM-NWheat, DSSAT, EPIC, SALUS), calibrated and validated with datasets from LTEs, to estimate SOC dynamics. Then we used the mean of the model ensemble to assess the impacts of climate change on SOC stocks under conventional (CT) and conservation tillage practices (NT: No Till; RT: Reduced Tillage). The assessment was completed for two long-term experiment sites (Agugliano – AN and Pisa – PI2 sites) in Italy under rainfed conditions. A durum wheat (Triticum turgidum subsp. durum (Desf.) Husn.) – maize (Zea mays L.) rotation system was evaluated under two different climate scenarios over the periods 1971–2000 (CP: Present Climate) and 2021–2050 (CF: Future Climate), generated by setting up a statistical model based on canonical correlation analysis. Our study showed a decrease of SOC stocks in both sites and tillage systems over CF when compared with CP. At the AN site, CT lost −7.3% and NT −7.9% of SOC stock (0–40 cm) under CF. At the PI2 site, CT lost −4.4% and RT −5.3% of SOC stocks (0–40 cm). Even if conservation tillage systems were more impacted under future scenarios, they were still able to store more SOC than CT, so that these practices can be considered viable options to mitigate climate change. Furthermore, at the AN site, under CF, NT demonstrated an annual increase of 0.4%, the target value suggested by the 4 per thousand initiative launched at the 21st meeting of the Conference of the Parties in Paris. However, RT at the PI2 needs to be coupled with other management strategies, as the introduction of cover crops, to achieve such target.     

基于生态足迹理论的新疆可持续发展研究

基于生态足迹和生态承载力理论,对2000-2014年新疆人均生态足迹、人均生态承载力、万元GDP生态足迹、生态协调系数的动态变化进行分析,运用灰色GM(1,1)预测模型预测新疆2020年、2025年人均生态足迹、人均生态承载力、万元GDP生态足迹。结果显示:2000-2014年新疆人均生态足迹呈逐年上升趋势,从2000年的1.699 3hm~2上升到2014年的4.100 5 hm~2,年均增幅为5.86%。新疆人均生态承载力基本保持在1.400 3~1.585 2 hm~2,且有逐年缓慢减小的趋势。人均生态足迹快速增长,人均生态承载力缓慢减小,造成生态赤字逐年扩大,人均赤字额由2000年的0.232 9 hm~2扩大到2014年的2.700 2 hm~2,赤字率由2000年的15.88%上升到2014年的192.82%。资源利用效益水平不断提高,生态协调度不断下降。预测结果表明:2020年新疆人均生态足迹将上升到6.496 4 hm~2,2025年将达到9.586 4 hm~2。2020年人均可利用生态承载力预计下降到1.294 7 hm~2,2025年为1.212 5 hm~2。万元GDP生态足迹2020年下降到0.649 9 hm~2,2025年预计为0.478 0 hm~2。基于此,提出降低化石能源消费比例,控制人口增长,提高土地产出效率是缓解新疆生态赤字的有效途径。     

基于极化分解和集成学习的PolSAR影像分类

为实现Pol SAR数据极化信息的充分利用,以进一步改善分类效果,该研究提出了一种基于极化分解和集成学习的Pol SAR影像分类方法。该方法首先利用多种极化分解方法从Pol SAR影像中提取极化参数;将提取的极化参数组合成一幅多通道影像;然后对多通道影像进行分割和特征提取,分别提取出各目标极化分解方法所对应的特征;并进行特征选择和分类,得到各目标极化分解方法的分类结果;最后利用集成学习技术对各分类结果进行集成。该研究以吉林省长春市部分区域为研究区,Radarsat2影像为数据源,将提出的方法应用于土地覆被分类中,取得了较好的分类效果,总体精度和Kappa系数分别达到了92.49%和0.90。此外,该研究还将提出方法与其他基于多种极化分解的分类方法进行比较,对比方法的总体精度和Kappa系数分别为90.74%和0.88,比提出方法分别低1.75%和0.02,对比结果进一步证明了提出方法的优越性。     

金融时间序列指标判别框架：以特质波动率为例

基于拐点集合判别的TBUD方法主要思路是分析拐点集合间的关系，并在高维空间进行划分，从而搭建判别模型，并将分析框架应用在特质波动率等若干指标上，利用实证数据得到结论。应用TBUD判别框架可以发现，特质波动率等指标无法对拐点集合进行清晰划分，因而并不具有预测能力。     

细分外推与模糊预测在近海水质非线性时序处理中的比较分析

近海水质非线性时间序列通常由于采集范围大、时间间隔长带有一定震荡性和模糊性,这使得对其进行分析与预测有一定的难度。本研究中以某近海水质指标磷酸盐(PO3-4-P)、硝酸盐(NO-3-N)、亚硝酸盐(NO-2-N)、铵盐(TNH+4-N)和硅酸盐(Si O2-3-Si)所形成的5种时间序列为例,采用逼近细分模式导出的细分外推法和多参考加权数据的模糊预测法对近海水质时序预测进行了比较分析,并通过图形与误差计算比较了两种方法的异同。结果表明:采用细分外推法预测序列在整体形状上能更好地逼近初始时序,而模糊预测法在整体逼近精度上占有优势。本研究中提出的预测比较方法可为同类问题的预测与模型选取提供参考依据。