生態(tài)環(huán)境與農(nóng)業(yè)氣象
Ecological Environment and Agrometeorology

生態(tài)與農(nóng)業(yè)氣象研究進展

1 生態(tài)和農(nóng)業(yè)氣象基礎(chǔ)理論與應(yīng)用技術(shù)研究

1.1 高時空分辨率數(shù)據(jù)融合模型研究

本研究對基于MODIS產(chǎn)品（高時間分辨率、低空間分辨率）和Landsat衛(wèi)星數(shù)據(jù)（高空間分辨率、低時間分辨率）光譜相似性、像元空間尺度和時間差異的數(shù)據(jù)融合模型STARFM進行如下改進：（1）將MODIS雙向反射率數(shù)據(jù)校正為天頂方向，根據(jù)地表覆蓋類型數(shù)據(jù)實現(xiàn)BRDF產(chǎn)品的缺測值填圖；（2）利用滑動窗技術(shù)，計算得到MODIS和Landsat數(shù)據(jù)的最大相關(guān)系數(shù)，進而實現(xiàn)MODIS和Landsat數(shù)據(jù)的幾何精度校正；在此基礎(chǔ)上構(gòu)建集成的數(shù)據(jù)融合模型框架（Integrated STARFM，ISTARFM），實現(xiàn)模型半自動化運行。該模型框架可以實現(xiàn)兩類數(shù)據(jù)在時間分辨率和空間分辨率的向上融合，為高時頻、精細化冬小麥區(qū)域干旱監(jiān)測服務(wù)。（王培娟）

1.2 Landsat地表參數(shù)反演及空間尺度轉(zhuǎn)換

利用LEDAPS（Landsat Ecosystem Disturbance Adaptive System）模型，將最新發(fā)射的Landsat8衛(wèi)星數(shù)據(jù)的DN值轉(zhuǎn)換為地表反射率，以地表反射率數(shù)據(jù)為基礎(chǔ)，計算得到歸一化植被指數(shù)（NDVI）、葉面積指數(shù)（LAI）和地表溫度（LST）。反演30 m分辨率地表溫度數(shù)據(jù)時，首先將Landsat8的30 m反射率數(shù)據(jù)進行降尺度轉(zhuǎn)換到100 m，得到與熱紅外波段空間分辨率相同的地表反射率數(shù)據(jù)，建立地表反射率和地表溫度的回歸模型T-R （Temperature-Reflectance），而后利用該模型和30 m分辨率的地表反射率數(shù)據(jù)，實現(xiàn)地表溫度的精細化空間尺度轉(zhuǎn)換（圖1）。（王培娟）

1.3 精細化逐日多層土壤墑情和灌溉預報系統(tǒng)研發(fā)

基于土壤水分、土壤質(zhì)地、蒸散、遙感水分虧缺指數(shù)等數(shù)據(jù)集，通過像元匹配、誤差分析和神經(jīng)網(wǎng)絡(luò)等方法，研究了土壤水分人工、自動、遙感監(jiān)測等多資料融合技術(shù)。利用中國103個站1961—2010年逐日總輻射和日照時數(shù)資料，分析了a、b系數(shù)時空變化規(guī)律，確定了中國不同區(qū)域各年代a、b系數(shù)參考值。在前期冬小麥農(nóng)田水量平衡簡化模型和精細化逐日多層土壤墑情和灌溉預報模型研究基礎(chǔ)上，構(gòu)建了3個省級預報系統(tǒng)（試用版），并在2014年冬小麥主要生長季進行試用。進一步開發(fā)了基于數(shù)據(jù)管理、數(shù)據(jù)處理、土壤墑情預報、灌溉量預報、產(chǎn)品制作、效果檢驗6個模塊的業(yè)務(wù)應(yīng)用平臺，并在河北省進行了初步本地化應(yīng)用，發(fā)布服務(wù)產(chǎn)品3次。（毛飛）

1.4 夏玉米生理生態(tài)與生長特性對干旱過程的響應(yīng)研究

夏玉米在七葉期灌水處理15天后發(fā)生輕旱，1個月后發(fā)生中旱，2個月后發(fā)生特旱。輕旱條件下，拔節(jié)期玉米葉片光合作用存在“午休”現(xiàn)象，頂部第一片展開葉光合速率、蒸騰速率、氣孔導度和水分利用效率對干旱響應(yīng)敏感，具有指示作用。全生育期內(nèi)玉米葉片光合速率、蒸騰速率和氣孔導度均隨干旱發(fā)生發(fā)展呈下降趨勢，葉片水分利用效率則呈上升趨勢。葉片光合速率、蒸騰速率和氣孔導度前期下降幅度較大，處理間差異顯著，隨干旱發(fā)展，土壤相對濕度差異減小，后期下降幅度減小，處理間差異也減小。葉片水分利用效率前期上升幅度較小，處理間差異也較小，干旱發(fā)展后期上升幅度較大，處理間差異也較大。夏玉米生育期內(nèi)葉片含水率隨干旱發(fā)生發(fā)展呈線性下降趨勢，并與土壤相對濕度顯著相關(guān)，可表征玉米的受旱程度。隨著干旱的發(fā)生發(fā)展，玉米光合限制因素由氣孔限制向非氣孔限制轉(zhuǎn)換，干旱強度越大、持續(xù)時間越長，葉片光合限制因素轉(zhuǎn)換時間越早，且轉(zhuǎn)換發(fā)生時葉片含水率越高。夏玉米莖含水率在拔節(jié)期達到最大值（93%左右），莖含水率對土壤水分脅迫的響應(yīng)沒有葉片含水率靈敏。土壤相對濕度較高處理的葉面積指數(shù)在抽雄期達到最大，土壤相對濕度較低處理的葉面積指數(shù)在灌漿期達到最大。比葉面積前期呈迅速下降趨勢，后期維持穩(wěn)定并略有下降趨勢；比葉重在開花期前呈明顯上升趨勢，開花期后呈穩(wěn)定并略有上升趨勢，比葉重在生長末期顯著高于生長初期。夏玉米地上生物量累積對土壤水分變化反應(yīng)靈敏，玉米各生育期的生物量干重均隨土壤相對濕度的下降而減少。拔節(jié)期和抽雄期干旱造成的夏玉米地上生物量增幅明顯下降。干旱條件下夏玉米地上生物量平均每天增幅速率在1.15～1.74 g/d之間，增幅速率隨土壤相對濕度的下降而下降（圖2）。（周廣勝）

2 農(nóng)業(yè)氣象防災減災技術(shù)研究

2.1 基于TIGGE和分布式水文模型的農(nóng)業(yè)干旱預警

采用全球三大TIGGE數(shù)據(jù)歸檔中心（中國氣象局、歐洲中期天氣預報中心和美國國家環(huán)境預報中心）的TIGGE資料，基于NOAH_LSM對TIGEE資料進行動力降尺度、神經(jīng)網(wǎng)絡(luò)ANN、TOPMODEL和新安江水文模型構(gòu)建了新的水文模型XXT。研究把土壤蓄水容量曲線和地下水水位線緊密關(guān)聯(lián)，形成新的蓄水容量曲線概念，并與TOPMODEL土壤分層結(jié)構(gòu)及地下產(chǎn)流方程結(jié)合，提取了新產(chǎn)流方程和新的水量平衡方程，在此基礎(chǔ)上構(gòu)建了新型降雨-徑流模型——XXT。雖然，此時的XXT模型通過大量試驗驗證了其比傳統(tǒng)的TOPMODEL和新安江水文模型性能優(yōu)越，但由于該模型的本質(zhì)是基于物理過程的模型，所以在預測精度上仍然比經(jīng)典的基于人工智能技術(shù)的統(tǒng)計模型（如神經(jīng)網(wǎng)絡(luò)，支持向量機等）要差。為了提高其模擬精度，我們首次將神經(jīng)網(wǎng)絡(luò)模塊嵌入到XXT模型的產(chǎn)流方案中，這與傳統(tǒng)方法顯著不同，因為傳統(tǒng)方法是將神經(jīng)網(wǎng)絡(luò)模塊集成到基于物理過程的水文模型的匯流方案中。（趙俊芳）

2.2 冬小麥干旱預測預警模式研發(fā)

在山東農(nóng)業(yè)大學國家重點實驗室試驗站利用美國Li-COR公司生產(chǎn)的Li-6400便攜式光合作用測定系統(tǒng)測定了干旱條件下作物光合生理生態(tài)參數(shù)變化情況，用英國Hansatech公司生產(chǎn)的FMS-II脈沖式調(diào)制熒光測定儀測定了葉片葉綠素熒光參數(shù)。研究表明：隨著干旱脅迫的加劇，фPSII呈現(xiàn)出明顯的下降趨勢，表明嚴重干旱脅迫時，葉片PSII的實際量子效率的影響顯著，F(xiàn)v/Fm與Fv/F0均與фPSII一致；隨著干旱脅迫程度加劇呈持續(xù)降低，或呈現(xiàn)降低—升高—降低趨勢，表明夏玉米在經(jīng)受適度水分脅迫時，可以作出一定的適應(yīng)性調(diào)節(jié)反應(yīng)，其調(diào)節(jié)程度與品種有關(guān)?；?，將農(nóng)業(yè)干旱識別模式中的相關(guān)模塊進行了相應(yīng)改進，在此基礎(chǔ)上利用C++與Fortran混合編程技術(shù)，建立起了適用于我國山東、河北及河南3省的農(nóng)業(yè)干旱預測預警模式，小麥干旱綜合預測預警準確率達到85%以上。（劉建棟）

2.3 華南地區(qū)龍眼寒害時空分布規(guī)律及演變趨勢

基于華南4?。ǜ＝?、廣東、廣西、海南）64個氣象站點1961—2012年逐日資料，根據(jù)《龍眼寒害等級標準》，選取影響龍眼寒害的主要影響因子，計算龍眼寒害指數(shù)，確定各級別寒害的發(fā)生頻率；通過對龍眼寒害指數(shù)矩陣、主要發(fā)生月份（11月至次年3月）寒害指數(shù)矩陣的經(jīng)驗正交分解（EOF），提取第一時間分量，分析了近52年華南龍眼寒害時空分布特征。結(jié)果表明：受氣候變暖影響，華南地區(qū)平均積寒隨年代呈減少趨勢，有利于龍眼寒害的減少；時間分布特征為1—2月為寒害發(fā)生頻率最高的月份，體現(xiàn)了年寒害分布的主要特征，是年寒害發(fā)生的最主要時段，但發(fā)生頻率隨年代呈現(xiàn)明顯減少趨勢。11—12月和3月不是年寒害發(fā)生的主要時段，發(fā)生頻率相對低，但個別年份仍有可能出現(xiàn)較重寒害；龍眼寒害空間分布呈明顯緯向分布，寒害發(fā)生頻率由南至北增加，寒害還與地理位置有關(guān)，發(fā)生頻率由沿海至內(nèi)陸增加，研究結(jié)果與歷史記載情況基本一致。對閩東地區(qū)龍眼寒害風險分析表明，1981—2010年閩東沿海龍眼寒害發(fā)生概率較大（0.42～0.43），風險概率為0.10～0.16。（趙俊芳）

2.4 西南地區(qū)農(nóng)業(yè)與水稻洪澇災害發(fā)生等級評價指標構(gòu)建

針對西南地區(qū)農(nóng)業(yè)與水稻洪澇災害監(jiān)測預警的需求，基于西南農(nóng)業(yè)區(qū)341個氣象站1961—2010年的逐日降水量、分省農(nóng)業(yè)洪澇災情、分縣水稻產(chǎn)量等資料，以重慶市單站洪澇等級指標（過程降水量）為原型，通過對原型指標中不同等級洪澇的降水量臨界值進行幅度為-50～+50 mm、步長為1 mm的增減，得到各單站共101個洪澇指標，在此基礎(chǔ)上，分別構(gòu)建單站、分省逐年洪澇指數(shù)以及平均每站分省逐年洪澇指數(shù)，采用灰色關(guān)聯(lián)度方法確定不同等級洪澇對實際災情的影響權(quán)重，綜合考慮由101個洪澇指標計算得到的洪澇指數(shù)與農(nóng)業(yè)洪澇實際受災程度的相關(guān)性、指標與歷史洪澇災害記錄的吻合性以及分省指標的可比性，優(yōu)選構(gòu)建了分省農(nóng)業(yè)洪澇等級指標。基于分縣水稻減產(chǎn)率與洪澇指數(shù)、洪澇過程日數(shù)、過程降水量，構(gòu)建了水稻洪澇等級指標。依據(jù)構(gòu)建的指標，揭示了西南地區(qū)農(nóng)業(yè)與水稻洪澇災害的時空分布特征。近50年云南、貴州、四川、重慶農(nóng)業(yè)洪澇發(fā)生較嚴重的年代分別為20世紀80年代、20世紀90年代、20世紀80年代和21世紀初。近50年洪澇的多發(fā)區(qū)分別位于云南西南和東南部、貴州西南部和四川盆地的西部和東北部。一季稻洪澇危險性指數(shù)高值區(qū)主要位于四川中北部、云南南部以及貴州西南部（圖3）。（霍治國）

2.5 重大病蟲害發(fā)生的氣候背景指示指標與長期預測、動態(tài)預警模型優(yōu)化

針對重大病蟲害（小麥白粉病、稻飛虱）發(fā)生氣象條件監(jiān)測、預警與評價的業(yè)務(wù)應(yīng)用需求，基于1971—2010年病蟲害發(fā)生面積、發(fā)生程度、74項大氣環(huán)流特征量、北太平洋海溫場格點資料，采用因子膨化、空間拓撲分析、最優(yōu)化處理、因子獨立性檢驗等方法，補充構(gòu)建了全國小麥白粉病發(fā)生面積率等級的大氣環(huán)流、北太平洋海溫指示指標；經(jīng)驗證檢驗，指示因子對小麥白粉病發(fā)生流行等級指示效果較好。補充研發(fā)了全國稻飛虱發(fā)生程度、小麥白粉病發(fā)生面積率等級的大氣環(huán)流預測模型、北太平洋海溫預測模型，對應(yīng)的模型等級預測正確率分別為97.5%、82.5%、82.5%、80%。補充研發(fā)了基于fisher判別分析的廣西桂林地區(qū)稻飛虱發(fā)生程度等級、河北地區(qū)小麥白粉病發(fā)生程度等級的動態(tài)預警模型，桂林地區(qū)稻飛虱發(fā)生程度等級的回代檢驗、預測檢驗基本一致，準確率分別為84.6%、88.2%，河北地區(qū)小麥白粉病發(fā)生程度等級回代檢驗、預測檢驗基本一致，準確率分別為97.8%、95.0%。（霍治國）

2.6 南方雙季稻低溫災害時空分布特征及風險分析

針對中國南方雙季稻區(qū)早稻播種移栽期低溫災害、晚稻抽穗揚花期寒露風災害，基于169個氣象站1981—2010年的逐日平均氣溫資料、氣象行業(yè)標準規(guī)定的雙季稻低溫災害等級指標，綜合考慮不同低溫災害等級及其出現(xiàn)的風險概率，構(gòu)建了低溫災害綜合風險指數(shù)，揭示了雙季稻不同低溫災害等級發(fā)生次數(shù)的時空變化以及等級風險和綜合風險的地理分布特征。近30年南方早稻低溫災害、晚稻寒露風（粳稻、秈稻）輕、中、重度發(fā)生次數(shù)以及總次數(shù)總體均呈減少趨勢，但部分區(qū)域呈增加趨勢，按災害呈增加趨勢覆蓋面積的大小依次為粳稻寒露風、早稻低溫災害和秈稻寒露風。中重度災害風險高值區(qū)：早稻低溫災害主要位于湖南、江西、福建3省的部分地區(qū)，發(fā)生概率中度為20%～40%、重度在10%以下；粳稻寒露風主要位于云南種植區(qū)中部、陜西種植區(qū)、四川成都西部以及四川種植區(qū)東北部等地，發(fā)生概率中度為20%～30%、重度為20%～45%；秈稻寒露風主要位于云南、湖南、安徽、陜西、四川的部分地區(qū)，發(fā)生概率中度為20%～40%，重度為50%～95%（圖4）。（霍治國）

2.7 華南荔枝寒害風險評估

構(gòu)建了包括災害危險性、承災體暴露性和脆弱性3個方面的華南荔枝寒害風險動態(tài)評估指標體系。致災因子及孕災環(huán)境的危險性由災害強度乘以其月或年發(fā)生的頻率確定。寒害強度即由最大降溫幅度、極端最低氣溫、日最低氣溫≤5.0 ℃的持續(xù)日數(shù)、日最低氣溫≤5.0 ℃的積寒等要素構(gòu)造的綜合寒害指數(shù)，分為極重、重度、中度、輕度和無寒害5個等級。承災體的脆弱性由荔枝減產(chǎn)（相對氣象產(chǎn)量）等級與其發(fā)生頻率之積確定，減產(chǎn)等級分為4級。承災體的暴露性根據(jù)荔枝收獲面積與耕地面積的比值確定。運用自然災害風險指數(shù)法（風險指數(shù)=危險性×暴露性×脆弱性）構(gòu)建了華南荔枝寒害的月或年風險評估模型。據(jù)此模型對廣東和海南省的荔枝進行逐月或年的寒害風險區(qū)劃。結(jié)果表明，廣東荔枝從11月開始即有發(fā)生寒害的中度風險，至12月、1月及2月發(fā)展為較高風險及以上。其中，廣寧、高要到臺山一線為寒害的高風險區(qū)。另外，年度寒害風險區(qū)較月的范圍有所擴大。海南荔枝月的寒害風險在中度及以下，而年的風險級別達到了較高（圖5）。（馬玉平）

2.8 重大農(nóng)業(yè)氣象災害立體監(jiān)測與動態(tài)評估技術(shù)研究

針對西南地區(qū)農(nóng)業(yè)干旱、南方雙季稻低溫和黃淮海小麥干熱風災害，取得了以下幾個方面的研究成果。完善了災害立體監(jiān)測指標。建立了以遙感植被指數(shù)為指標的遙感監(jiān)測大面積干熱風災害的方法。建立了基于干熱風危害指數(shù)的黃淮海地區(qū)冬小麥干熱風災損評估模型。采用經(jīng)驗指標模型或?qū)⒅笜饲度隣RYZA2000模型作為水稻低溫冷害的動態(tài)評估方法，并可以從站點和區(qū)域尺度進行災害的動態(tài)評估。站點尺度的評估緊跟水稻生長過程，監(jiān)測到低溫影響后，及時作出冷害評估，提供定量的冷害評估結(jié)果。區(qū)域評估以遙感反演信息作為數(shù)據(jù)源代入指標模型或作物模型中計算或更正評估結(jié)果，并以MODIS為例，探索了低溫造成生育期延遲的評估方法。對于西南作物干旱，建立了基于干旱累積指數(shù)（DI）的產(chǎn)量損失評估模型和基于WOFOST模型的干旱影響作物評估模型。在前期研究成果的基礎(chǔ)上，初步完成了黃淮海干熱風和西南作物干旱省級災害監(jiān)測與評估業(yè)務(wù)平臺。（趙艷霞）

2.9 華北冬小麥灌漿期高溫熱害指標研究

在冬小麥灌漿中后期，短期（3天、6天或9天）高溫脅迫導致灌漿速率下降、灌漿持續(xù)時間縮短，千粒重與高溫脅迫期間午間平均冠層氣溫呈顯著性的負線性相關(guān)關(guān)系。通過對千粒重數(shù)據(jù)進行標準化處理，得到了不同年型間表現(xiàn)一致的千粒重相對值與高溫特征值的關(guān)系，并據(jù)此提出了冬小麥灌漿中后期短期高溫熱害的溫度指標。由3年不同長度的短期高溫脅迫試驗結(jié)果提煉出了一個反映灌漿中后期全階段高溫熱害當量的特征量-高溫有效度時，并得出適應(yīng)于不同年型的千粒重相對值與高溫有效度時的統(tǒng)計相關(guān)關(guān)系，并由此給出了不同減產(chǎn)程度的致災高溫積熱指標。（譚凱炎）

2.10 長江中下游地區(qū)雙季早稻冷害、熱害風險識別技術(shù)

以長江中下游地區(qū)48個站點1961—2012年的氣象資料為基礎(chǔ)，利用統(tǒng)計分析、Mann-Kendall和小波分析的方法研究早稻生長季前期冷害、后期冷害、后期熱害的時空變化。結(jié)果表明：近52年來早稻生長季氣溫呈極顯著升高趨勢（＞0.5 ℃/10a）；前期冷害和后期冷害均從21世紀初開始出現(xiàn)減弱趨勢，前期冷害變化存在2～4年的短周期，后期冷害變化存在2～4年和4～6年2個周期；后期熱害從21世紀初開始有顯著增強趨勢，無明顯周期特征；空間分布上，前期冷害高風險區(qū)基本與山脈地形特征相吻合而后期冷害與水系的分布相關(guān)性很高。全球變暖背景下長江中下游地區(qū)低溫災害減弱、高溫災害增強。（王春乙）

2.11 東北地區(qū)玉米主要氣象災害風險評價

利用東北地區(qū)35個農(nóng)業(yè)氣象站1961—2010年氣象資料、1981—2010年玉米發(fā)育期資料、1961—2010年產(chǎn)量面積資料、近50年東北3省的災情資料以及近10年東北3省各縣的社會經(jīng)濟統(tǒng)計資料，以玉米出苗-抽雄、抽雄-成熟兩個生長階段里發(fā)生的干旱及冷害為研究對象，基于水分虧缺指數(shù)和熱量指數(shù)分別建立了干旱指標和冷害指標，對東北地區(qū)玉米干旱、冷害進行風險分析。建立了包括危險性、脆弱性、暴露性和防災減災能力4個方面的東北地區(qū)玉米干旱、冷害風險評價模型，指出危險性和防災減災能力是風險評價模型中最重要的2個影響因子。給出了東北地區(qū)玉米干旱、冷害高風險值區(qū)位于黑龍江西南部和東北部，以及遼寧西部建平縣一帶，風險指標值在0.8以上；吉林西北部、東南部、遼寧東北部為次高值區(qū)，風險指標值在0.6～0.7之間；低值區(qū)位于遼寧中南部及遼東半島，風險值在0.3左右。（王春乙）

3 農(nóng)業(yè)應(yīng)對氣候變化研究

3.1 全球變化影響下主要作物的脆弱性及評價指標

研究表明，顯著增溫對春小麥生育期和關(guān)鍵發(fā)育期的影響是氣候變化背景下春小麥產(chǎn)量脆弱性的主要原因。春小麥生育階段氣溫升高，使拔節(jié)期、抽穗期、開花期和成熟期顯著提前，從而使春小麥生育期顯著縮短。然而，播種期調(diào)制、耕作方式改變和新品種引入等適應(yīng)性措施的不斷實施，彌補了由于生育期縮短和發(fā)育期變化等對小麥產(chǎn)量的可能影響。統(tǒng)計分析表明，由于增溫趨勢和適應(yīng)性措施實施水平的區(qū)域性差異影響，增溫對春小麥生育期縮短和發(fā)育期變化的貢獻率在23%～68%，平均達到40.3%；每增溫1 ℃，生育期可能縮短6～7天，產(chǎn)量可能減少522 kg/hm2。另外，應(yīng)該清晰地認識到，人類的適應(yīng)性措施伴隨氣候變化而持續(xù)實施，到目前為止的作物生育期和發(fā)育期變化是氣候變化和人類適應(yīng)氣候變化的共同結(jié)果。（俄有浩）

3.2 東北玉米生產(chǎn)對氣候變化的響應(yīng)與適應(yīng)

進一步細化了作物發(fā)育期，以平均資源適宜指數(shù)（Isr）、平均效能適宜指數(shù)（Ise）和平均資源利用指數(shù)（K）作為評價指標，評價了近50年東北地區(qū)玉米氣候資源的適宜情況和利用率。從年際、潛在生長季和作物生長期3個時間尺度方面，確定影響春玉米生長發(fā)育的關(guān)鍵氣象因子，分析了近30年來東北春玉米關(guān)鍵發(fā)育期的變化特征和演變趨勢，構(gòu)建了玉米生長期對氣候變化的響應(yīng)關(guān)系模型，探討了春玉米發(fā)育期對不同時間尺度氣象因子的響應(yīng)規(guī)律。提出了分離東北玉米氣候產(chǎn)量的最優(yōu)方法，分離出氣候變化對東北春玉米產(chǎn)量的影響，確定了影響東北春玉米氣候產(chǎn)量的關(guān)鍵氣象因子，建立了反映不同地區(qū)氣象因子和春玉米氣候產(chǎn)量關(guān)系的區(qū)域模型。（趙俊芳）

3.3 大氣CO2 濃度升高和增溫影響作物需水量變化機理研究

利用開頂式生長箱（OTC）與大田試驗，對比研究了增溫和CO2濃度升高對冬小麥需水量的影響。結(jié)果表明，冬小麥生長期間（2013年10月11日至2014年6月10日），日平均增溫3.1 ℃，并且日平均施加CO2濃度到760×10-6，冬小麥生育期縮短17天（7%）。日均耗水量增加0.7 mm（OTC日均2.7 mm，大田日均2.0 mm），全生育期增加需水量123.5 mm（25.5%）。另外，利用半封閉OTC（無CO2施加）與大田對比試驗表明，半封閉式OTC比大田日平均增溫0.7 ℃，冬小麥生育期縮短7天，日均耗水量增加0.65 mm，全生育期增加需水量138.8 mm（28.7%），與封閉式OTC中日均耗水量和全生育期增加的需水量接近。出現(xiàn)這種情況的原因可能有兩方面：一是雖然封閉式OTC增溫幅度明顯大于半封閉式OTC，但是封閉環(huán)境下空氣流動性比半封閉環(huán)境下差，不利于水汽流動，導致封閉式OTC中蒸散量較??；二是許多文獻資料提到增加CO2濃度能夠減少蒸散量，增加水分利用效率。由于該試驗觀測結(jié)果是增溫和CO2濃度增加的綜合表現(xiàn)，無法分離各自作用對需水量的影響及貢獻，因此，還需要進一步試驗觀測和分析。（俄有浩）

3.4 增溫和CO2增加對華北主要作物產(chǎn)量影響的試驗研究

偏冷年的增溫將促進小麥的分蘗，有效穗數(shù)顯著增加，籽粒產(chǎn)量較對照大幅度增加；偏暖年的增溫也顯著促進冬小麥有效穗數(shù)和穗粒數(shù)增加，盡管使得千粒重顯著降低，但并未導致產(chǎn)量下降。OTC小麥增溫+CO2的試驗也得出，增溫+CO2對產(chǎn)量沒有產(chǎn)生影響，與單純CO2增加試驗不同，即沒有預期的CO2施肥效應(yīng)在小麥增溫+CO2中表現(xiàn)出增產(chǎn)的結(jié)果。玉米大田增溫試驗得出：增溫可能降低玉米產(chǎn)量，可能與玉米生長期處于夏季高溫期，如再升溫就可能超出玉米適應(yīng)高溫的能力有關(guān)。預期的CO2施肥效應(yīng)在玉米增溫+CO2中表現(xiàn)出增產(chǎn)的效應(yīng)（圖6）。（房世波）

3.5 近30年東北春玉米發(fā)育期對氣候變化的響應(yīng)

基于1981—2010年東北地區(qū)55個農(nóng)業(yè)氣象觀測站發(fā)育期數(shù)據(jù)、16個氣象站逐日氣象資料，采用趨勢變率、秩相關(guān)分析、主成分分析和結(jié)構(gòu)方程模型等方法，分析了近30年東北春玉米關(guān)鍵發(fā)育期的變化特征，探討了春玉米發(fā)育期對不同時間尺度氣象因子的響應(yīng)規(guī)律。結(jié)果表明，1981—2010年春玉米關(guān)鍵發(fā)育期（播種期、抽雄期、成熟期）均有延后趨勢，大部分地區(qū)春玉米生長前期（播種期-抽雄期）日數(shù)減少，生長后期（抽雄期-成熟期）日數(shù)增加，全生育期日數(shù)增加。在絕大多數(shù)年份，春玉米播種期在溫度適播期之后，成熟期在初霜日之前。近30年對東北春玉米生育期日數(shù)影響最大的氣象要素為溫度，主成分分析結(jié)果顯示，年際尺度的升溫、溫度生長期的延長和作物生長期的高溫對生育期日數(shù)影響顯著；結(jié)構(gòu)方程模型指出，作物生長期的最高溫度和最低溫度對生育期日數(shù)影響有間接效應(yīng)，主導氣象要素能夠解釋生育期日數(shù)變異的44%。全球變暖背景下，東北春玉米發(fā)育期變化是作物響應(yīng)氣候變化和農(nóng)業(yè)生產(chǎn)適應(yīng)氣候變化的共同結(jié)果。（郭建平）

3.6 氣候變化和科技進步對中國玉米產(chǎn)量影響的定量評估

農(nóng)作物產(chǎn)量受氣候變化和科技進步的共同影響，客觀和定量評價氣候變化和科技進步對農(nóng)作物產(chǎn)量的影響有利于保障農(nóng)業(yè)的可持續(xù)發(fā)展。本研究利用農(nóng)業(yè)生態(tài)區(qū)域法計算了作物光合生產(chǎn)潛力、光溫生產(chǎn)潛力和氣候生產(chǎn)潛力，并結(jié)合主要作物主產(chǎn)省實際產(chǎn)量變化趨勢，系統(tǒng)分析了近50年來不同氣候資源要素變化對作物氣候生產(chǎn)潛力的影響，分析了近30年來氣候變化和科技進步對主要作物產(chǎn)量的影響。研究結(jié)果表明：近50年輻射減少對我國主要農(nóng)作物產(chǎn)生了顯著的不利影響；近50年溫度升高對玉米和雙季稻有不利影響，但對小麥有正效應(yīng)；近50年降水變化對農(nóng)作物的影響地區(qū)間有顯著差異，對不同作物也有差異，但以正效應(yīng)為主；近30年氣候變化對農(nóng)作物產(chǎn)量均以負面影響為主，農(nóng)業(yè)增產(chǎn)主要是科技進步的貢獻。如果沒有氣候變化的不利影響，我國糧食作物的實際增產(chǎn)幅度可能比現(xiàn)在更大。（郭建平）

3.7 東北種植制度變化下作物氣候生產(chǎn)潛力和氣候資源利用效率分析

利用已有的種植制度界限指標體系，綜合考慮水、熱資源對東北地區(qū)農(nóng)業(yè)種植制度的影響，分析了種植制度界限變化敏感區(qū)作物氣候生產(chǎn)潛力和氣候資源利用效率的變化。結(jié)果表明：隨著未來氣候變化，綜合考慮熱量和降水資源，東北一年兩熟的北界明顯的北移東擴，2071—2100年黑龍江哈爾濱附近的小部分區(qū)域可以實現(xiàn)一年兩熟，而黑龍江西南地區(qū)由于水資源的缺乏，仍只適宜一熟制作物種植；由于高溫脅迫，研究區(qū)域內(nèi)春玉米氣候生產(chǎn)潛力會隨時間下降，種植模式的調(diào)整是提高氣候生產(chǎn)潛力的一種有效方式。但在熱量相對較高的地區(qū)，過高的溫度導致冬小麥、夏玉米生育期的縮短，冬小麥與夏玉米兩熟模式下的氣候生產(chǎn)潛力總和也會降低，但仍然高于一熟種植方式；隨著氣候變暖，積溫相應(yīng)增加，一熟制種植方式出現(xiàn)了資源浪費的現(xiàn)象。改一熟為兩熟，延長了作物的生長季節(jié)，可以很好地利用氣候資源，有利于提高農(nóng)業(yè)氣候資源利用效率。（郭建平）

圖1 Landsat8數(shù)據(jù)地表參數(shù)反演結(jié)果(a：歸一化植被指數(shù)(NDVI)；b：葉面積指數(shù)(LAI)；c：地表溫度(LST))Fig. 1 Inversion maps of Landsat8 surface parameters (a: NDVI; b: LAI; c: LST)

圖2 玉米全生育期葉片光合作用限制因素Fig. 2 Factors limiting leaf photosynthesis of summer maize during growing season

圖3 1961—2010年中國西南地區(qū)農(nóng)業(yè)洪澇平均次數(shù)Fig. 3 Average agricultural f ood frequency in Southwest China during 1960–2010

圖4 南方地區(qū)晚稻（粳）寒露風風險概率及綜合風險指數(shù)地理分布（a：輕度寒露風；b：中度寒露風；c：重度寒露風；d：寒露風風險指數(shù)）Fig. 4 Geographical distribution of risk probability of each-level cold dew wind to japonica rice during 1981 to 2010 (a: the occurring probability of mild cold dew wind; b: the occurring probability of medium cold dew wind; c: the occurring probability of severe cold dew wind; d: risk index of cold dew wind)

圖5 廣東和海南省荔枝月及年寒害風險區(qū)劃Fig. 5 The monthly or annual risk zoning of litchi chilling injury in Guangdong and Hainan provinces

圖6 單獨增溫和增溫+CO2復合對玉米產(chǎn)量影響的差異Fig. 6 The different effects between warming treatments and warming conbined CO2 enrichment

Research Progress in Ecology and Agrometeorology

1 Ecological and agrometeorological theories and application techniques

1.1 A study on data fusion model for high temporal and spatial resolution remote sensing data

The Spatial and Temporal Adaptive Ref ectance Fusion Model (STARFM) was used to blend Landsat and MODIS surface ref ectance based on their spectral similarities and spatio-temporal differences. An operational data fusion framework was built by integrating STARFM (referred to as ISTARFM hereinafter). Compared with STARFM, several improvements have been incorporated in the ISTARFM. These include: (1) viewing angular correction on the MODIS daily bidirectional ref ectance to get daily NBAR (Nadir BRDF (Bidirectional Reflectance Distribution Function) - Adjusted Reflectance), and filling gap values in BRDF products with BRDF lookup table obtained from MODIS International Geosphere-Biosphere (IGBP) land cover and MODIS BRDF products, (2) precise and automated co-registration on MODIS and Landsat paired images by looking for maximum correlation coeff cient with the moving-window technique. The ISTARFM provides a feasible and cost-effective way to build dense time-series surface ref ectance at Landsat spatial resolution, which can serve as regional drought monitoring for winter wheat precisely at high spatial and temporal resolutions with remote sensing images. (Wang Peijuan)

1.2 Landsat surface parameters inversion and spatial down-scaling

Digital number (DN) was converted to surface reflectance based on LEDAPS (Landsat Ecosystem Disturbance Adaptive System) model for Landsat8. And then several surface parameters were inverted using surface reflectance at their own spatial resolution, which include normalized difference vegetation index (NDVI) at 30 m resolution, leaf area index (LAI) at 30 m resolution, and land surface temperature (LST) at 100 m resolution. A Temperature-Ref ectance (T-R) model was developed with LST and down-scaled surface ref ectance at 100 m resolution. Therefore, LST can be down-scaled spatially based on T-R model and surface ref ectance at 30 m resolution to get LST at 30 m resolution (Fig. 1). (Wang Peijuan)

1.3 Research on and development of f ne daily multi-layer soil moisture and irrigation forecasting system for winter wheat

Based on data sets of soil moisture, soil texture, evapotranspiration, water def cit indices of remote sensing and so on, the fusion technologies of soil moisture observation data by manual, automatic and remote sensing were studied with the methods of pixel matching, error analysis, neural network and others. Using daily total radiation and sunshine time data in 103 stations from 1961 to 2010 in China, the spatial and temporal distribution of the a coeff cients and b coeff cients were analyzed and the a coeff cients and b coefficients for each age (ten years) in every region of China were given. Based on earlier stage research on the simplified models of water balance and daily multi-layer soil moisture and irrigation forecast models for winter wheat, three provincial forecasting systems (probation) were established, and tried in main winter wheat growth period in 2014. The operational application platform based on six modules of data management, data processing, soil moisture forecasting, irrigation quantity forecasting, product making and results testing was developed and tried in Hebei Province preliminarily. Three service products were issued. (Mao Fei)

1.4 Study on responses of summer maize ecophysiological and growth characteristics to drought process

During the manipulation experiment of summer maize drought occurrence and development, light drought occurred after 15 days of water control, and moderate drought happened one month later, extreme drought appeared two months later. The midday depression phenomenon of leaf photosynthesis occurred during the jointing stage under the conditions of light drought. The photosynthetic rate, transpiration rate, stomatal conductance and water use eff ciency of the f rst maize leaf were sensitive to drought, implying they were able to be indicative indices. The photosynthetic rate, transpiration rate and stomatal conductance of maize leaves during growing season decreased signif cantly with the decrease of soil relative moisture. The difference of all these parameters between water treatments declined with the development of consecutive drought. The water use eff ciency showed a rising trend and its difference between water treatments was small at earlier stage, but enlarged at later stage with the aggravation of consecutive drought. The leaf water content decreased linearly with the drought occurrence and development during growing season and it was signif cantly correlated with soil relative moisture, implying that the leaf water content is able to be an index of crop drought. The limiting factor of maize photosynthesis was a conversion process from stomatal limitation to non-stomatal limitation during growing season. The occurrence time of the conversion process was in correspondence with drought intensity and duration. The leaf SPAD value was not as sensitive as the photosynthetic rate to drought, which ref ected its hysteretic nature. Electron transfer rate, photochemical quenching coeff cient and photochemical efficiency decreased significantly from filling stage to milk stage. The summer maize stem water content reached the maximum at jointing stage (about 93%) and it was not as sensitive as leaf water content to drought. The leaf area index reached the maximum at tasseling stage for the treatment under higher soil relative moisture, while for the treatment under lower soil relative moisture it reached the maximum at f lling stage. The specific leaf area declined rapidly in the early stage, while it showed a stable trend. The specific leaf weight showed an opposite trend, and it was signif cantly higher in the late stage than in the early stage. The summer maize aboveground biomass was sensitive to soil water content, and it decreased with the decrease of soil relative moisture at all stages. The reduction caused by drought at jointing and tasseling stages was obvious. The average daily increase of summer maize aboveground biomass was between 1.15 g to 1.74 g, and the increase rate declined with the decrease of soil relative moisture (Fig. 2). (Zhou Guangsheng)

2 Agrometeorological disaster prevention and mitigation

2.1 Agricultural drought warning based on TIGGE and distributed hydrological model

In this study, three global ensemble weather prediction systems from the China Meteorological Administration (CMA), the European Centre for Medium-Range Weather Forecasts (ECMWF), and the US National Centers for Environmental Prediction (NCEP) in THORPEX (The Observing System Research and Predictability Experiment) Interactive Grand Global Ensemble (TIGGE) archive were used. A new distributed hydrological model XXT was built based on TIGEE dynamical downscaling through NOAH_LSM, neural network ANN, TOPMODEL and Xin,anjiang hydrological models. It was based on the soil moisture storage capacity distribution curve (SMSCC), some essential modules of the Xin,anjiang model, together with the simple model framework of the TOPMODEL (a topography based hydrological model). The innovation of XXT was that the water table was incorporated into SMSCC to connect the surface runoff production with base f ow production. This improved the description of the dynamically varying saturated areas that produced runoff and also captured the physical underground water level. XXT was tested and found to have better performance against the TOPMODEL and the Xin,anjiang model. However, due to the fact that XXT model was based on physical processes, the prediction accuracy was still poorer than the classical statistical model which was based on artif cial intelligence techniques (such as neural networks, support vector machines, etc.). In order to improve its simulation accuracy, we f rstly embedded neural network module in the runoff program of XXT model, which was signif cantly different from the conventional method. Because the traditional method was that the neural network module was integrated into the scheme of hydrological model with physical processes. (Zhao Junfang)

2.2 Development of the agro-drought prediction model for the winter wheat in North China Plain

Variation of the physiological parameters of the winter wheat under drought conditions, including photosynthesis rate of the winter wheat leaves etc., was measured with Li-6400 portable photosynthesis analyzers in the National Key Laboratory of Shandong Agricultural University. Green f uorescence parameters of the leaves were also determined with FMS-II pulse modulated f uorescence detector. The results indicat that фPSII deceased obviously with the evolution of the drought in the f eld, which means that the real quantum efficiency of PSII would be seriously effected under severe drought conditions. Variations of the Fv/Fm and Fv/F0 agreed quite well with that of the фPSII. All of the parameters would decrease steadily with the development of the drought, or fluctuate like a wave of deceasing-increasing-decreasing. These variations indicat that the crop can somewhat regulate itself to adapt to the consecutive drought conditions. However, the adaptability changes with varieties. Based on these achievements mentioned above, the researchers have successfully improved some of the related modules in the agro-drought prediction model. An agro-drought prediction model was also established by using the hybrid programming technology of C++and Fortran, with comprehensive prediction accuracy as high as 85%. The model is quite suitable for agro-drought prediction in Shandong, Hebei and Henan provinces in North China Plain. (Liu Jiandong)

2.3 Spatial and temporal distribution characteristics of longan caused by chilling injury in South China

According to the meteorological service standards “Grade of Chilling Injury to Dimocarpus Longan Trees” (QX/T168-2012), the main inf uencing factors of chilling injury in longan production were selected, and the chilling injury indexes (Hi) of longan in the 64 stations in four provinces of South China from 1961 to 2012 were calculated. The occurrence frequencies of each level of chilling injury were determined. Based on the empirical orthogonal function (EOF), the matrix of longan chilling injury indexes of 52 years, and the matrix of chilling injury indexes in main months (November-March) were analyzed, while the f rst time components of the above matrix were extracted. The results show that under climate warming, the average accumulated harmful temperature (＜5 ℃) presented a signif cant decreasing trend along with decadal time in South China, and it was helpful to reduce disaster of chilling injury. Changes in temporal characteristics of chilling injury indicated the highest frequency was found in January and February, ref ecting the main distribution features of chilling injury. These two months were the most important period of chilling injury occurring in a year. November–December and March were not the main period for chilling injury, with a relatively low frequency. But in some years, it might still experience severe chilling injury. Spatial distributions of chilling injury presented obvious latitudinal distribution. The occurrence frequency of chilling injury increased from south to north and from coast to inland areas, which was related to geographic location. A comparative analysis of the historical records indicated that the spatial and temporal distributions of chilling injury could reflect the main actual characteristics of historical record. The frequency of chilling injury from 1981 to 2010 in the eastern coastal Fujian was high (0.42–0.43), with the risk probability standing at 0.10–0.16. (Zhao Junfang)

2.4 The evaluation indicators of agricultural and rice f ooding level

For the demand of agricultural and rice f ood monitoring and early warning in Southwest China, the single station f ood indicators, provincial f ood indicators year by year and provincial f ood indicators year by year in each station have been developed based on daily precipitation data, agricultural f ood disaster information and the rice yield by county from 341 weather stations in southwest agricultural areas from 1961 to 2010. In the progress of building the indicators, flood indicators of Chongqing as the prototype, through increasing and decreasing precipitation threshold by the amplitude which was -50～+50 mm and step size was 1 mm, 101 groups of flood indicators were obtained in each province. Agricultural flood level indicators of each province were constructed which used grey correlation method to determine weights of different levels of f ood impact on actual situation, comprehensive consideration of correlation between f ood intensity calculated by 101 indicators and crop flood real seriousness, coincidence between indicators and historical flood records and comparability of indicators of each province. Rice f ood level indicators were constructed based on yield reduction rate of each county and f ood index, days of f ood disaster process and process rainfall. According to the f ood level indicators, temporal-spatial distribution features of f ood disaster in southwest agricultural areas were revealed. The ages in which f ood disaster appears very serious are 1980s in Yunnan, 1990s in Guizhou and 1980s and 2000s in Sichuan and Chongqing in recent 50 years. The areas in which flood frequently occurred are located in southwest and southeast of Yunnan, southwest of Guizhou and west and northeast of Sichuan basin in recent 50 years. Regions of high value f ood dangerous index of single-season rice are mainly located in northern and central Sichuan, south of Yunnan and southwest of Guizhou (Fig. 3). (Huo Zhiguo)

2.5 The climate background indicator of major pests diseases occurrence and long-term prediction, dynamic early warning model optimization

For operational application needs of major pests diseases (powdery mildew, planthoppers) occurrence weather conditions monitoring, early warning and assessment, data on the national occurrence area and degree of rice planthopper and wheat powdery mildew from 1971 to 2010, 74 kinds of atmospheric circulation characteristics from 1970 to 2010, and North Pacif c SST data from 1969 to 2010 were reveiwed. Atmospheric circulation and North Pacific SST data level indicators of rice planthopper and wheat powdery mildew occurrence area level were also used including factor puffing, topological analysis, optimization process, and factor independence test. After the test, indication factor method showed a better result for the powdery mildew of occurrence epidemic level. The whole nation rice planthopper occurrence degree model, the rice planthopper occurrence rate of atmospheric circulation model and North Pacif c SST model were established, the accuracy of which are 97.5%, 82.5%, 82.5% and 80%, respectively. Early dynamic warning models were developed based on f sher discriminant analysis of rice planthopper occurrence degree in Guangxi Guilin and wheat powdery mildew in Hebei. The level evaluation accuracy of rice planthopper model in Guilin is 84.6%, while the prediction accuracy is 88.2%. The level evaluation accuracy of wheat powdery mildew model in Hebei achieves 97.8%, while the prediction accuracy is 95.0%. (Huo Zhiguo)

2.6 Low temperature disaster temporal/spatial distribution characteristics of double cropping rice in southern China and its risk analysis

For cold damage in early rice seeding transplanting period and cold dew wind damage of late rice in heading f owering stage in double cropping rice areas in southern China, the risk index of cold temperature damage was established which was based on the community standard and the data of daily average temperature from 169 weather stations during 1981–2010. It revealed the characteristics of geographical distribution of the occurrence probability and integrated risk index for double cropping rice from different low temperature disasters by level. In recent 30 years, the cold damage of early rice in southern China, the frequency of light, moderate and severe occurrence of late rice cold dew wind (japonica rice and indica rice), and the number of total occurrences showed a decreasing trend. But some regions showed an increasing trend. According to the disasters showing an increasing trend in terms of coverage, they are japonica cold dew wind, cold rice disasters and indica cold dew wind. High value areas of moderate to severe disaster risk: Cold damage of early rice mainly in parts of Hunan, Jiangxi and Fujian provinces, moderate probability of 20% to 40%, severe below 10%. Japonica rice cold dew wind is mainly located in the middle of Yunna, Shaanxi planting area, western Sichuan and Chengdu, and in northeastern of Sichuan growing areas, where the moderate probability of occurrence is 20% to 30%, and the severe probability of occurrence is 20% to 45%. Indica rice cold dew wind is located in parts of Yunnan, Hunan, Anhui, Shaanxi and Sichuan, where the moderate of occurrence is 20% to 40%, and the severe probability of occurrence is 50% to 95% (Fig. 4). (Huo Zhiguo)

2.7 Risk assessment of litchi chilling injury in South China

The risk evaluation index system of litchi chilling injury was f rstly constructed, which included risk of disasters, exposure and vulnerability of litchi in South China. The risk of the disaster factor and its environment was determined by the intensity of disaster multiplied by the frequency of its occurrence in a month or year. Chilling intensity was the integrative chilling damage index constructed by maximum temperature drop, extreme minimum temperature, persisting days of low temperature and harmful chilling accumulation in which the minimum temperature stays below 5. Chilling intensity was classif ed as severe, heavy, moderate, mild and no chilling injury. The vulnerability was determined by the grades of litchi yield decrease multiplied by its occurrence frequency. The grade was divided into 4 levels. Exposure was determined by the ratio of litchi harvest area to cultivated land area. The monthly or annual risk assessment model of litchi chilling in South China was then constructed by using natural disaster risk index method which was risk multiplied by vulnerability and exposure. Based on this model, the monthly or annual chilling injury risk zoning was carried out in Guangdong and Hainan provinces. The results show that moderate risks of litchi chilling injury have begun in November and gradually developed into higher risk from December, January to February in Guangdong Province, in which the high risk area of litchi is in Guangning, Gaoyao to Taishan regions. Moreover, annual chilling risk area is greater than that for monthly one. Monthly chilling risk grade is in moderate and below but the annual is rather high in Hainan Province (Fig. 5). (Ma Yuping)

2.8 Study on the stereo monitoring of and the dynamic assessment technology for major agrometeorological disasters

The f ndings about the dry-hot wind for winter wheat in Huanghuaihai areas, low-temperature damage for double cropping rice in southern China and the agricultural drought in Southwest China have been summarized as follows. Indicators for stereo monitoring were improved. Vegetation indexes were developed to monitor the large-scale dry-hot wind disastrous to wheat. The model based on the hazard index of dry-hot wind was constructed to assess the wheat loss due to dry-hot wind. For low-temperature damage to rice, loss was assessed at site and regional scale based on ORYZA2000 and empirical models. The assessment at site scale followed the rice growth process. Once the impact of low temperature was monitored, the quantitative evaluation results would be made in time. At regional assessment, remote sensing inversion was used as input data to ORYZA2000 or empirical model to calculate and modify the results. And taking MODIS data coupled with model as example, the methods were investigated for evaluating the impact of low temperature on delaying rice growth. For crop drought, yield loss was assessed based on drought accumulation index (DI) and WOFOST model. The operational platforms of monitoring and evaluation of dry-hot wind in Huanghuaihai areas and drought disasters in Southwest China at the provincial level have been established. (Zhao Yanxia)

2.9 Study on high temperature stress at grain f lling stage of winter wheat in North China

High temperature stress at the mid to late grain filling stage of winter wheat reduced the grain filling rate and shortened the grain f lling duration, resulting in a signif cant reduction of grain yield. After taking into account the other influencing factors, a significant negative linear relationship was found between the grain weight and the midday canopy air temperature in this period. This study suggest that the extent of the high temperature stress effects on winter wheat grain yield depends on both the strength and duration of the stress, which can be represented by the effective accumulated heat (summation of the hour’s temperature differences above a threshold value during the mid to late grain filling stage of winter wheat). There were signif cant negative linear relationships between the normalized grain weight of winter wheat and the effective accumulated heat above 30 °C during the mid to late grain f lling stage (P ≤0.05). This proved that the high temperature stress index is a useful parameter for quantitative evaluation of the impacts of high temperature at the grain f lling stage on winter wheat yield. (Tan Kaiyan)

2.10 Risk identif cation techniques for cold and heat damage to double-season early rice (DSER) in the middle and lower reaches of Yangtze River

Based on daily meteorological data of 48 meteorological stations during 1961–2012 in the middle and lower reaches of Yangtze River (MLRYR), spatial and temporal distribution of early cold damage, late cold damage and late heat damage of early rice was analyzed using such methods as statistical analysis, Mann-Kendall and wavelet analysis. Results show that over the past 52 years, average temperature increased signif cantly (＞0.5/10a) during early rice growing season in MLRYR. Both the early and late cold damage showed a downtrend from 2000. There was a short cycle of cold damage of two to four years in early period and two to four, four to six years in late period. Late heat damage showed a signif cant increase trend from 2000 with no obvious periodicity. High risk area of early cold damage was consistent with spatial distribution of mountain terrain while that of late cold damage was consistent with distribution of water system. Cold damage weakened and heat damage increased in MLRYR under global warming. (Wang Chunyi)

2.11 Risk assessment of main meteorological disasters to maize in Northeast China

The meteorological data from 1961 to 2010, maize growth data from 1981 to 2010, production data from 1961 to 2010 in 35 agro-meteorological stations, disaster record for the latest 50 years in three provinces of Northeast China, and social and economic statistics for the latest 10 years in the counties of these three provinces were employed in this study. The drought and chilling injury at two stages of the maize including emergence-tasseling and tasseling-mature were studied, and the risk of the drought and cold injury to corn in Northeast China was evaluated. The creation of index systems of drought and cold injury was based on water deficit index and heat index. The indices of drought and chilling injury risk assessment model of the maize included hazard, vulnerability, exposure, and emergency response and recovery. Hazard and emergency response and recovery were the two most important factors in the model. For the whole growth and development period of the maize, the high values of risk index were in the southwest and northeast of Heilongjiang Province and west of Liaoning Province, standing above 0.8. The moderate values were found in the northwest and southeast of Jilin Province and the northeast of Liaoning Province, standing between 0.6 and 0.7. The low values were in the central south of Liaoning Province and Liaodong Peninsula, standing around 0.3. (Wang Chunyi)

3 Response of agriculture to climate change

3.1 Vulnerability and evaluation index of main crops under global change

The research results indicate that the impact of climate warming on growth periods and critical growth stages of spring wheat was the primary reason for vulnerability of spring wheat yield under the background of climate change. Increased temperature during the growing season accelerated the phonological development, including internode elongation, heading, anthesis and stage of grain ripening, resulted in a shorter growth period. The application of adaptation strategies, such as adjusting the sowing dates, introducing new cultivars and changing the tillage practices, likely compensated some effects of shorter growth period on wheat yield. A correlation analysis reveales that the increase in temperature during the growing season was responsible by 23% to 68%, with an average of 40.3% for the shortening growth period and altered farming practices based on phonological events, in spite of regional differences in increased temperature trend and in application of adaptation strategies. The trend analyses reveal that rising temperature has shortened the growth period by 6 to 7 days per 1 ℃, resulting in the decrease in spring wheat yield of averagely 522 kg ha-1. In addition, the fact should be recognized that the adaptive strategies continue to be performed along with the changing climate since the changes in growth period and phenology of spring wheat up to now both result from climate change and human adaptive activities. ( E Youhao)

3.2 Response of maize production to climate change and its adaptation in Northeast China

The growth period of spring maize was further subdivided into four stages: germination to emergence, emergence to jointing, jointing to tasseling, and tasseling to maturity. The average resource suitability index (Isr), average eff cacy suitability index (Ise), and average resource utilization index (K) were used as indicators of agricultural climatic resource suitability and utilization for maize production. The utilization dynamics of agricultural climatic resource during spring maize cultivation from 1961 to 2010 in Northeast China were analyzed. The key climatic factors limiting the spring maize growth were identif ed from three time scales of interannual, potential growth season and crop growing period. The variations of key growth stages of spring maize in Northeast China over the past 30 years were analyzed. The regression models on the growing period of maize in response to climate change were also established. The response of growing period of maize to climate change at different time scales was investigated. Compared with the methods of moving average and harmonic average, logistic regression optimally decoupled the climate induced yield of spring maize. The contribution of climate change to spring maize yield over the past three decades in Northeast China was decoupled. The key meteorological factors limiting the climate-induced yield were determined. Finally, the models between climatic variables and climate-induced yield of spring maize in Northeast China were also established to ref ect geographical differences. (Zhao Junfang)

3.3 Study on the impact of extreme temperature and CO2 enrichment on crop water demand

The control experiments in open top chambers (OTC), side half closed chambers and field were performed to investigate the effects of elevated atmospheric carbon dioxide concentration [CO2] and increase in temperature on winter wheat water requirement. The results show that, compared to the f eld, in open top chambers the air temperature increased by mean daily temperature of 3.1 ℃ and [CO2] elevated to 760×10-6, the whole growing period was shortened by 17 days (7%), and the average daily water requirement increased by 0.7 mm. Consequently, the water requirement increased by 123.5 mm (25.3%) during the growth period. While in side half closed chambers in which using the ambient [CO2], the air temperature increased by mean daily temperature of 0.7 ℃, the whole growing period was shortened by 7 days and the average daily water requirement increased by 0.65 mm. So, the water requirement increased by 138.8 mm (28.7%) during the growth period, indicating the approximately equivalent water requirement both in open top chambers and in side half closed chambers. There would be two reasons to explain these results. For one thing, in OTC, the air exchange with outside is less than that in side half closed chambers, leading to less evapotranspiration though signif cantly warmer in OTC than in side half closed chambers. For another, the elevated [CO2] may decrease the evapotranspiration by raising water use efficiency as mentioned in literatures. However, we have not separated the positive effect by elevated [CO2] on evapotranspiration from increased air temperature, due to the compounded effects of both elevated [CO2] and increased air temperature. Further experiments still need to be made to separate the impacts of elevated [CO2] and increased air temperature. (E Youhao)

3.4 Experimental study on the impacts of extreme temperature and CO2 enrichment on crop yield in North China

The heated treatments of winter wheat showed higher values of panicles per square meter, and no signif cant difference in the 1000 grain weight, which had signif cantly increased the yield compared to CK in the colder year (2011). The warm year (2012) showed higher values of panicles per square meter and grain number in each panicle, while the 1000 grain weight is signif cantly reduced compared with CK, but the heated treatments did not lead to yield reduction. Warming and CO2enrichment could hardly show any difference with CK, and no fertilizer effects of CO2were shown in this experiment. The warmed treatment of maize had the potential to reduce the yield compared with CK, which maybe come from the negative effect of extreme temperature in summer. But warming and CO2enrichment could set off the negative effect of extreme temperature, which increased the yield in 2012 (Fig. 6). (Fang Shibo)

3.5 Response of growth stages of spring maize to climate change in Northeast China over the past30 years

Based on the observation data of spring maize from 55 agricultural meteorological stations, and daily meteorological data of 16 meteorological stations in Northeast China, the variations of key growth stages of spring maize in Northeast China over the past 30 years were analyzed, using such methods as trend rate, spearman correlation analysis, principal component analysis and structural equation modeling. Finally, the responses of growth period of spring maize to meteorological factors over the past 30 years were further analyzed at different time scales. The results show that the spring maize’s critical growth stages in Northeast China over the past 30 years were postponed. The number of days decreased during the early maize growth period (from sowing to tasselling), while both days during the late maize growth period (from tasselling to maturation) and days of whole growth period increased. In most years, the sowing date of spring maize was later than suitable planting date, and the maturating date was earlier than the f rst frost date. Temperature was the most notable meteorological factor responsible for the altered spring maize growth period during the past 30 years. A principal component analysis showed that the increased temperature at the inter-annual timescale, the prolonged temperature growth period and the high temperature in the crop growth period were more notable than other meteorological factors. While, in the structural equation modeling, the effects of temperature on days of growth period partly were indirect, and signif cant meteorological factors could explain 44% of variation in growth period’s days. Changes in the growth stages of spring maize were caused by the response of crops and the adaptation of agricultural production to climate change under global warming. (Guo Jianping)

3.6 Attribution of crop yield increase in China to climate change and technological advancement

Crop yields are affected by climate change and technological advancement. Objectively and quantitatively evaluating the attribution of crop yield change to these two factors will ensure sustainable development of agriculture under climate change. The relative impacts of different climate variables on climatic potential productivity of maize from 1961 to 2010 in China are then evaluated based on Agro-Ecological Zones (AEZ) model. In the AEZ model, the climatic potential productivity was examined through three steps or levels: photosynthetic potential productivity, photosynthetic thermal potential productivity, and climatic potential productivity. Combined with the observations of maize, the contributions of climate change and technological advancement to maize yield from 1981 to 2010 in China were separated. The results show that, from 1961 to 2010, decreased radiation was the main factor leading to the decrease of climatic potential productivity of crops. Increased temperature had a signif cant adverse impact on the climatic potential productivities of maize and double cropping rice in China. However, it had a positive effect on wheat. Climate change had a signif cant adverse impact on the crop yield in China from 1981 to 2010. This suggests that technological advancement had offset the negative effects of climate change on maize yield. Our f ndings highlight the fact that agronomic technological advancement has contributed dominantly to crop yield increases in China in the past three decades. (Guo Jianping)

3.7 Effects of adjusted cropping systems on climatic potential productivity and utilization eff ciency of climatic resources in Northeast China

Based on the existing limitation indicators of cropping systems, while considering the impact of water and thermal resources on cropping systems in Northeast China, changes in climatic potential productivity and utilization eff ciency of climatic resources in the sensitive areas of cropping systems were analyzed. The results show that the planting boundaries of different cropping systems in Northeast China obviously shifted toward the north and the east along with the heat and precipitation resources. However, in southwestern Heilongjiang Province, only one crop in a year could be planted because of the shortage of precipitation resources. Due to high temperature stress, the climatic potential productivity of spring maize would be reduced in the future. In certain small areas nearby Harbin in the Heilongjiang Province, two crops in a year could be planted from 2071 to 2100. Therefore, adjusting the cropping system was an effective way to improve the climatic potential productivity and climate resource utilization. In some thermal resource rich areas, the climatic potential productivity of summer maize or winter wheat decreased because of the shortened growth season caused by a warming and drying climate. However, if the one crop in one year pattern (spring maize) becomes a two crops in one year pattern (summer maize and winter wheat), the total climatic potential productivities of two crops would be higher than that of one crop. This f nding further illustrates that after a paradigm shift from one crop in one year to two crops in one year, the thermal resources, which were not originally used, would become fully utilized. (Guo Jianping)