基于格网的GCM数据修订分析未来海南岛农业水热资源的变化特征

以海南岛为研究区域，选用5个大气环流模式（GCMs）1970−1999年的逐日输出数据和同期地面气象观测数据，使用空间插值降尺度到0.5°×0.5°格网。以格网单元为基础，应用系统误差修订（修正值法或比值法）和多模式集合平均方法（贝叶斯模型平均法BMA或等权重平均法EW），训练与验证GCMs输出值并进行综合修订。在此基础上，分析RCP2.6、RCP4.5和RCP8.5情景下，未来海南岛近期（2020−2059年）和远期（2060−2099年）农业水热资源，包括年平均气温、1月平均气温、≥10℃积温、≥20℃积温、年降水量、1月降水量和≥20℃界限温度生长期间降水量的变化特征。结果表明：GCMs输出值的系统误差和BMA权重系数在格网间存在较大的空间差异，且GCMs输出值低估逐日最高气温约3.55℃，高估逐日最低气温约1.19℃，逐日降水量仅为观测值的54.35%。基于格网的综合修订，可有效降低GCMs输出值在空间上的不确定性，BMA与EW的修订结果相似，均优于单一GCM模式。通过格网BMA综合修订后，最高气温、最低气温和降水量在验证期的相关系数r分别约提升0.10、0.07和0.06；均方根误差RMSE分别约降低2.38℃、1.01℃和1.01mm；较单一GCM相对观测值的偏差平均约减少3.25℃、1.13℃和25.67mm。未来海南岛农业热量资源在空间上主要表现为从中部向外围逐渐升高，高温主要分布在南部至西部沿海地区，年平均气温的增幅全岛较为接近，1月平均气温、≥10℃积温和≥20℃积温的增幅分别表现为由东向西、由北向南和由中部向外围递减。在时间上，RCP8.5情景下所有农业热量资源均为极显著增加且增温最快，RCP4.5情景为先增加后平缓，RCP2.6情景较为平缓，远期无显著增温。未来海南岛降水资源在空间上转为由东向西逐步递减的格局，南部和北部沿海地区降水变率增加，西部和中部降水变率减少，在时间上无显著变化趋势。随着未来海南岛气候变暖和降水格局的改变，农作物适宜种植面积扩大，会对农业生产带来巨大挑战，应提前布局，做好趋利避害。     

Considering sink strength to model crop production under elevated atmospheric CO2

Climatic changes and elevated atmospheric CO₂ concentrations will affect crop growth and production in the near future. Rising CO₂ concentration is a novel environmental aspect that should be considered when projections for future agricultural productivity are made. In addition to a reducing effect on stomatal conductance and crop transpiration, elevated CO₂ concentration can stimulate crop production. The magnitude of this stimulatory effect (‘CO₂ fertilization’) is subject of discussion. In this study, different calculation procedures of the generic crop model AquaCrop based on a foregoing theoretical framework and a meta-analysis of field responses, respectively, were evaluated against experimental data of free air CO₂ enrichment (FACE) environments. A flexible response of the water productivity parameter of the model to CO₂ concentration was introduced as the best option to consider crop sink strength and responsiveness to CO₂. By varying the response factor, differences in crop sink capacity and trends in breeding and management, which alter crop responsiveness, can be addressed. Projections of maize (Zea mays L.) and potato (Solanum tuberosum L.) production reflecting the differences in responsiveness were simulated for future time horizons when elevated CO₂ concentrations and climatic changes are expected. Variation in future yield potential associated with sink strength could be as high as 27% of the total production. Thus, taking into account crop sink strength and variation in responsiveness is equally relevant to considering climatic changes and elevated CO₂ concentration when assessing future crop production. Indicative values representing the crop responsiveness to elevated CO₂ concentration were proposed for all crops currently available in the database of AquaCrop as a first step in reducing part of the uncertainty involved in modeling future agricultural production.     

基于高光谱与集成学习的单粒玉米种子水分检测模型

为建立单粒玉米种子水分含量的高精度检测模型,制备了80份不同水分含量的玉米种子样本。针对玉米种胚朝上和种胚朝下分别进行高光谱反射图像采集,每份样本取样100粒,波长范围为968.05～2 575.05 nm。采用PCA快速提取单粒种子光谱,经多元散射校正预处理后,分别采用随机森林(RF)和AdaBoost算法建立单粒种子水分检测模型,并集成两种算法特征提出基于加权策略的改进RF用于单粒种子水分含量建模。利用单粒玉米种子胚朝上的光谱信息建立的改进RF模型训练集相关系数R为0.969,训练集均方根误差(RMSEC)为0.094%,测试集R为0.881,测试集均方根误差(RMSEP)为0.404%;利用单粒玉米种子胚朝下的光谱信息建立的改进RF模型训练集R为0.966,RMSEC为0.100%,测试集R为0.793,RMSEP为0.544%。实验结果表明：改进RF的泛化能力和预测精度明显优于RF和AdaBoost算法;种胚朝上的单粒玉米种子水分含量检测模型优于种胚朝下的模型。高光谱检测技术结合集成学习算法建立的玉米种子水分检测模型预测精度高,稳健性好。     

基于遥感数据同化的土壤含盐量估算方法

为探究同化遥感数据对监测区域尺度土壤含盐量时空信息的适用性,以河套灌区沙壕渠灌域为研究区,以高分一号卫星影像为数据源,通过灰度关联法筛选光谱指数,采用岭回归法构建不同深度的土壤含盐量反演模型,使用集合卡尔曼滤波同化算法将遥感数据应用于HYDRUS-1D模型中,开展区域尺度不同深度土壤含盐量的同化研究。结果表明,基于不同深度土壤含盐量的岭回归法模型,其R²均在0.64以上,RE为0.14～0.22,反演精度较高,得到的反演值较为准确;在单点尺度上,与模拟值、反演值相比,同化值更接近实测值,其EFF为0.84～0.93,NER为0.61～0.73,均为正数,且RMSE降低到0.006%～0.011%,提高了HYDRUS-1D模型模拟精度;在区域尺度上,不同深度同化值的r均为0.94以上,NER为0.61以上,优于模拟值和反演值,且同化精度随着深度的增加而降低。本文基于遥感数据和HYDRUS-1D模型的集合卡尔曼滤波同化研究,提高了土壤含盐量的模拟精度,对提高监测区域尺度土壤含盐量时空信息的精度具有一定的参考价值。     

贵州最高气温预报方法研究

最高气温预报一直以来是贵州最为棘手的问题,近年来在国家气象局预报质量通报中,成绩较为靠后。为改变这一现象,笔者基于各预报中心常规气温预报资料及地面观测资料,在贵州境内展开最高气温的多模式集合预报研究。结果表明,多模式集合预报技术有效地改进了预报的准确率,在对2013年1月1日—2014年4月30日120 h的逐24 h预报中,各预报中心的多模式集合预报结果明显降低了预报的均方根误差,效果远优于最好的单个预报中心(ECMWF)和多模式的集合平均,不仅很好地改善了贵州最高气温的预报效果,还给当地预报及决策气象服务提供更有效的参考。     

基于作物生长模型与机器学习算法的区域冬小麦估产

为精准、高效、实时地实现区域冬小麦产量估算,以河南省鹤壁市淇县桥盟乡石桥村为研究区,基于分辨率10 m的Sentinel-2多时相光学遥感影像,利用集合卡尔曼滤波(Ensemble Kalman filter, EnKF)算法同化PROSAIL辐射传输模型反演的多期叶面积指数(Leaf area index, LAI)到PyWOFOST作物生长模型中实现一定数量不同长势单点产量的估测,最后利用建立的机器学习模型和面域数据反演区域冬小麦产量,实现作物生长模型与机器学习算法的应用耦合及一种新的区域冬小麦估产模式。研究基于Sobol参数敏感性分析法量化对贮藏器官总干重质量(Total dry weight of storage organs, TWSO)与LAI_max的敏感性参数,并基于反演的多期LAI和粒子群优化(Particle swarm optimization, PSO)算法优化与LAI_max相关的TDWI、TBASE、CVS、CVL敏感性参数,将其输入到PyWOFOST模型中,利用EnKF算法和时序LAI数据调整对TWSO相关的AMAX...     

基于高光谱遥感和集成学习方法的冬小麦产量估测研究

【目的】利用2种灌溉处理下不同发育阶段的冬小麦冠层高光谱信息,通过机器学习方法对小麦籽粒产量进行估测精度研究,明确产量最佳估测模型,对于育种工作有着重要应用价值。【方法】以黄淮麦区207个主栽小麦品种为材料,于2018—2019和2019—2020年度连续2个生长季在河南省新乡基地的正常灌溉和节水处理下种植,并调查开花期、灌浆前期和灌浆中期的冠层高光谱数据,分别以6种机器学习方法和集成方法建立光谱指数产量估测模型。【结果】2种灌溉处理下,3个生育期各光谱指数均与产量呈极显著相关（P<0.0001）,且表现出较高的遗传力（0.61-0.85）,主要受遗传因素控制。在正常灌溉处理下,与传统机器学习方法表现最佳的模型相比,集成学习方法在3个生育期的平均决定系数(R²) 分别由0.610、0.611和0.640提高至0.649、0.612和0.675,平均均方根误差 (RMSE) 分别降低至0.607、0.612和0.593 t·hm^-2;节水处理下,3个生育期的平均R²分别由0.461、0.408和0.452提高至0.467、0.433和0.498,平均RMSE分别降低至0.519、0.559和0.504 t·hm^-2。【结论】利用集成方法将不同模型估测结果进行结合,能够有效地提高产量估测精度,2种灌溉处理下均在灌浆中期估测精度最佳,可为冬小麦育种工作中产量估测提供参考。     

A model ensemble approach to determine the humus building efficiency of organic amendments in incubation experiments

Organic amendments are important to sustain soil organic matter (SOM) and soil functions in agricultural soils. Information about the contribution of organic amendments to SOM can be derived from incubation experiments. In this study, data from 72 incubated organic amendments including plant residues, digestates and manure were analysed. The incubation data was compiled from three experimental setups with varying incubation times, soils and incubation temperatures, in which CO₂ release was measured continuously. The analysis of the incubation data was performed with an approach relying on conceptual parts of C-TOOL, CCB, Century, ICBM, RothC and Yasso which are all well-approved first-order carbon models that differ in structure and abstraction level. All models are an approximation of reality, whereby each model differs in understanding of the processes involved in soil carbon dynamics. To accumulate the advantages from each model a model ensemble was performed for each substrate. With the ability of each carbon model to compute the distribution of carbon into specific SOM pools a new approach for evaluating organic amendments in terms of humus building efficiency is presented that, depends on the weighted model fit of each ensemble member. Depending on the organic substrate added to the soil, the time course of CO₂ release in the incubation studies was predicted with different accuracy by the individual model concepts. Averaging the output of the individual models leads to more robust prediction of SOM dynamics. The E_HUM value is easy to interpret and the results are in accordance with the literature.     

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.     

Neural network ensemble residual kriging application for spatial variability of soil properties

High quality, agricultural nutrient distribution maps are necessary for precision management, but depend on initial soil sample analyses and interpolation techniques. To examine the methodologies for and explore the capability of interpolating soil properties based on neural network ensemble residual kriging, a silage field at Hayes, Northern Ireland, UK, was selected for this study with all samples being split into independent training and validation data sets. The training data set, comprised of five soil properties: soil pH, soil available P, soil available K, soil available Mg and soil available S,was modeled for spatial variability using 1) neural network ensemble residual kriging, 2) neural network ensemble and 3) kriging with their accuracies being estimated by means of the validation data sets. Ordinary kriging of the residuals provided accurate local estimates, while final estimates were produced as a sum of the artificial neural network (ANN) ensemble estimates and the ordinary kriging estimates of the residuals. Compared to kriging and neural network ensemble,the neural network ensemble residual kriging achieved better or similar accuracy for predicting and estimating contour maps. Thus, the results demonstrated that ANN ensemble residual kriging was an efficient alternative to the conventional geo-statistical models that were usually used for interpolation of a data set in the soil science area.