1 農(nóng)業(yè)氣候資源與農(nóng)業(yè)氣象災(zāi)害

1.1 中國農(nóng)業(yè)氣象（小麥）模式（CAMM）初始版本的構(gòu)建

從機理過程（作物生長過程、區(qū)域特色過程）、構(gòu)造方法（插拔式建模、數(shù)據(jù)同化）和應(yīng)用方式（實時運轉(zhuǎn)、互聯(lián)網(wǎng)）等方面確定了中國農(nóng)業(yè)氣象（小麥）模式的基本框架。明確了模式的基本功能、產(chǎn)品類型和內(nèi)容。初步完成了模式基礎(chǔ)網(wǎng)格、種植區(qū)、作物生長、發(fā)育、土壤等22種參數(shù)或變量初值的區(qū)域化。建立了小麥發(fā)育初值對溫度的非線性響應(yīng)和品種改良模式、冬小麥株高變化模式。形成了中國農(nóng)業(yè)氣象（小麥）模式（CAMM）的初始版本。（馬玉平）

1.2 基于降水過程的玉米、油菜澇漬等級指標(biāo)構(gòu)建

創(chuàng)建了基于降水過程的玉米澇漬等級指標(biāo)，提出了油菜澇漬過程災(zāi)變判別方法。針對現(xiàn)有玉米、油菜澇漬指標(biāo)難以業(yè)務(wù)應(yīng)用的難題，提出了基于風(fēng)險逆過程分析的區(qū)域夏玉米、春玉米澇漬等級指標(biāo)構(gòu)建方法，解析了當(dāng)前過程和前期降水量對玉米澇漬形成的影響效應(yīng)，創(chuàng)建了基于澇漬當(dāng)前過程和前期降水量的江淮夏玉米、江漢和江南西部春玉米不同生育時段澇漬等級指標(biāo)，探索了玉米澇漬過程實時監(jiān)測評估的天氣學(xué)方法，為開展玉米澇漬監(jiān)測評估業(yè)務(wù)提供了技術(shù)支撐?；谟筒藵碀n臨災(zāi)、受災(zāi)狀態(tài)變化，采用降水量、連陰日數(shù)影響的動態(tài)累積模擬災(zāi)害發(fā)生過程，分生育期厘定臨災(zāi)狀態(tài)與受災(zāi)狀態(tài)的臨界線，構(gòu)建了基于油菜澇漬過程的逐日災(zāi)變判別指標(biāo)，確定了澇漬過程的災(zāi)變時間及災(zāi)害的后續(xù)影響時間，為區(qū)域油菜澇漬災(zāi)變過程的動態(tài)監(jiān)測提供了方法支撐。（霍治國）

1.3 海南辣椒氣象災(zāi)害綜合風(fēng)險評估研究

針對海南省辣椒干旱、洪澇、低溫災(zāi)害風(fēng)險綜合評價，提出了不同災(zāi)害的危險性、敏感性、脆弱性和防災(zāi)減災(zāi)能力量化表征方法，基于層次分析法和熵權(quán)法，厘定了不同災(zāi)害影響因子的權(quán)重，建立了區(qū)域災(zāi)害風(fēng)險綜合評價模型，編制了海南辣椒氣象災(zāi)害綜合風(fēng)險區(qū)劃，經(jīng)驗證與實際發(fā)生情況相吻合。研究發(fā)現(xiàn)干旱和低溫敏感性具有相似的空間分布，從中部向沿海減少，而洪澇敏感性相反；災(zāi)害的脆弱性主要集中在中部地區(qū)，與低防災(zāi)能力相似；海南東部遭受綜合損害風(fēng)險高（圖1）。（霍治國）

1.4 華北平原冬小麥品種變化對發(fā)育期影響模擬研究中的不確定性

發(fā)育期模型是評價品種變化對作物發(fā)育期影響的主要研究工具。目前模擬品種變化對作物發(fā)育期影響研究中由模型機理和參數(shù)化方法導(dǎo)致的不確定性認(rèn)識還不充分。針對此問題，收集了華北平原47個農(nóng)業(yè)氣象觀測站1986—2012年冬小麥的發(fā)育期資料和同期的逐日平均氣溫數(shù)據(jù)，研究了4個常用的發(fā)育期模型和2種參數(shù)化方法下模擬結(jié)果中存在的不確定性及其來源。首先利用1986—1988年的觀測資料校正模型參數(shù)，然后模擬1989—2012年的發(fā)育期，以定量研究冬小麥品種變化對發(fā)育期的影響。結(jié)果表明，經(jīng)過參數(shù)校正，選用的4個模型和2個參數(shù)化方法均能較好地模擬1986—1988年的發(fā)育期，模擬誤差（RMSE）都小于3 d。然而，在模擬品種變化對發(fā)育期的影響時，不同的模型和不同參數(shù)化方法間的結(jié)果差異卻比較大。在營養(yǎng)生長階段，影響的平均值為0.20 d/10a（95%的置信區(qū)間為-2.81～3.22 d/10a），在生殖生長階段，平均值為1.50 d/10a（置信區(qū)間為-1.03～4.02 d/10a）。進一步的分析表明，模型機理和參數(shù)化方法均可引入不確定性，且機理的作用大于參數(shù)化方法。在營養(yǎng)生長階段，模擬值的范圍與發(fā)育階段平均溫度呈顯著正相關(guān)關(guān)系，在生殖生長階段則呈多項式關(guān)系，這表明不同模型得到的模擬值之間差異較大。由于無法評價哪個模擬結(jié)果更接近真實值，我們建議開展不同年代培育品種的試驗研究，以進一步加深對此問題的了解。此外，為了降低不確定性，需要選擇溫度波動較大的年份對模型進行校正。（鄔定榮）

1.5 基于MODIS和SEBAL模型的黃淮海平原冬小麥水分生產(chǎn)力研究

以MODIS 多時相遙感影像產(chǎn)品、氣象數(shù)據(jù)和作物生育期為基礎(chǔ)，借助SEBAL模型估算了黃淮海平原冬小麥實際蒸散量；通過MODIS NDVI光譜曲線特征與冬小麥單產(chǎn)數(shù)據(jù)的耦合，將縣域尺度作物單產(chǎn)“降尺度”至基于像元的產(chǎn)量柵格圖，實現(xiàn)冬小麥產(chǎn)量柵格化。在完成作物實際蒸散量模擬和產(chǎn)量柵格化基礎(chǔ)上，對黃淮海平原冬小麥水分生產(chǎn)力進行估算。冬小麥水分生產(chǎn)力區(qū)域平均值為1.21 kg/m3，高值區(qū)主要位于北京、天津、山東北部和河北南部地區(qū)。在環(huán)渤海山東半島濱海外向型二熟農(nóng)漁區(qū)、海河低平原缺水水澆地二熟兼旱地一熟區(qū)和黃淮平原南陽盆地水澆地旱地二熟區(qū)，隨著冬小麥產(chǎn)量的增加，其水分生產(chǎn)力增加；在燕山太行山山前平原水澆地二熟區(qū)水分生產(chǎn)力將隨著實際蒸散量的減少和產(chǎn)量的增加而增大，同時產(chǎn)量增加對水分生產(chǎn)力提高的貢獻(xiàn)大于實際蒸散量的減少；在江淮平原丘陵麥稻兩熟區(qū)水分生產(chǎn)力的提高主要依靠實際蒸散量的減少。（楊建瑩）

1.6 華北平原冬小麥作物系數(shù)降尺度估算方法——以固城農(nóng)業(yè)氣象實驗站為例

作物系數(shù)（Kc）是計算作物實際蒸散的重要參數(shù)，也是估算作物需水量或耗水量的重要指標(biāo)。目前，獲取作物系數(shù)的主要方法中聯(lián)合國糧農(nóng)組織（FAO）推薦的“四段式”作物系數(shù)（Kc_FAO）是普遍采用的方法，但是對于冬小麥等越冬作物而言，因其在越冬前亦存在一個小的生長高峰期，故FAO推薦的“四段式”作物系數(shù)方法并不適用。本研究提出了基于冬小麥關(guān)鍵發(fā)育階段的作物系數(shù)方法（Kc_stage），與FAO推薦的“四段式”作物系數(shù)方法相比，Kc_stage可以捕獲冬小麥作物系數(shù)的雙峰特征，能夠更精細(xì)地刻畫冬小麥的生長規(guī)律。同時，基于固城站渦度相關(guān)測得的實際蒸散數(shù)據(jù)，采用留一交叉驗證的方法，評估了關(guān)鍵發(fā)育階段作物系數(shù)和傳統(tǒng)的FAO作物系數(shù)的優(yōu)劣。結(jié)果表明，與實測ETa相比，基于Kc_stage和Kc_FAO估算的實際蒸散的均方根誤差在逐日尺度上分別為0.07 mm/d和0.16 mm/d，在關(guān)鍵發(fā)育階段尺度上分別為0.01 mm/d和0.27 mm/d，且用Kc_FAO方法在冬小麥生長初期會低估實際蒸散而在快速生長階段和生長旺盛階段則存在高估的現(xiàn)象。（王培娟）

1.7 黃淮海平原冬小麥最大可能蒸散的估算

作物最大可能蒸散（ETc）考慮了作物及當(dāng)?shù)氐乇頎顩r，為當(dāng)?shù)氐乇韺嶋H覆蓋情況下實際蒸散的理論上限值，能客觀分析作物對水分的需求程度和農(nóng)業(yè)干旱狀況?；谶b感（葉面積指數(shù)和地表反照率）數(shù)據(jù)和逐日氣象數(shù)據(jù)，利用Penman-Monteith公式，計算黃淮海平原小麥種植區(qū)27個氣象站冬小麥生育期2000—2015年逐日蒸散，提取得到冬小麥生育期逐日最大可能蒸散數(shù)據(jù)集，并分析其時空變化特征及成因。結(jié)果表明，與聯(lián)合國糧農(nóng)組織（FAO）單作物系數(shù)法計算的最大可能蒸散ETc_kc對比，區(qū)域平均最大可能蒸散ETc的時間變化趨勢與ETc_kc一致，空間分布上ETc符合客觀實際。黃淮海平原冬小麥全生育期、越冬期和返青-拔節(jié)期ETc均呈北低南高的分布特征，日平均值分別為1.95 mm/d，0.46 mm/d和2.74 mm/d；其余3個生育期（越冬前、抽穗期、乳熟-成熟期）在空間分布上差異不大，日平均值分別為1.23 mm/d，4.71 mm/d和3.72 mm/d。冬小麥不同生育期（含全生育期）ETc的空間分布主要受葉面積指數(shù)分布特征的影響，二者呈顯著正相關(guān)關(guān)系。（王培娟）

1.8 夏玉米苗期主要生長指標(biāo)的土壤水分臨界點確定方法

土壤水分不足是引起作物干旱的主要因素。準(zhǔn)確確定作物生長變化響應(yīng)土壤水分的臨界點對客觀辨識、監(jiān)測作物干旱的發(fā)生發(fā)展具有重要意義。本研究基于6個初始土壤水分梯度的夏玉米持續(xù)干旱模擬試驗，利用多元方差分析方法確定了對土壤水分變化敏感的玉米生長指標(biāo)，提出了基于正態(tài)總體統(tǒng)計容忍下限確定玉米生長指標(biāo)的臨界土壤水分方法。結(jié)果表明，夏玉米苗期的莖含水率、葉含水率、蒸騰速率、光合速率、氣孔導(dǎo)度和葉面積是對土壤水分變化敏感的玉米生長指標(biāo)，其臨界土壤水分（0～30 cm平均土壤相對濕度）分別為72%、65%、62%、60%、58%、46%，反映出隨著土壤水分降低，玉米的莖含水率、葉含水率、蒸騰速率、光合速率、氣孔導(dǎo)度和葉面積依次受到影響。研究結(jié)果可為夏玉米苗期干旱發(fā)生發(fā)展的監(jiān)測和定量評估提供依據(jù)，也可為生態(tài)系統(tǒng)響應(yīng)氣候變化的閾值確定提供思路。（周廣勝）

1.9 干旱對夏玉米苗期葉片權(quán)衡生長的影響

葉片是植物對干旱響應(yīng)最敏感的器官之一，葉片性狀變化及其權(quán)衡關(guān)系能夠反映植物對資源的利用策略以及對干旱的適應(yīng)對策?；?014年6個初始土壤水分梯度的夏玉米持續(xù)干旱模擬試驗研究表明，隨著干旱的發(fā)展，夏玉米各葉片性狀均會受到影響，但不同干旱程度的影響不一致?；谒置{迫系數(shù)及干旱持續(xù)時間提出了干旱程度的定量表達(dá)，隨著干旱的發(fā)生發(fā)展，干旱程度在0～1之間變化。當(dāng)干旱程度小于0.21時，夏玉米葉片性狀不會受到顯著影響；0.21～0.76時，葉片性狀大小受到影響，但變化趨勢不會發(fā)生改變；0.76～0.91時，新葉形成補償不了老葉脫落，有效葉片數(shù)、葉干重、綠葉面積和葉含水量等性狀提前出現(xiàn)下降趨勢；大于0.91時，葉片生長幾乎停滯。夏玉米葉片性狀在干旱條件下的適應(yīng)性生長本質(zhì)上體現(xiàn)了其在快速生長與維持生存之間的權(quán)衡，但不同干旱程度下，夏玉米葉片性狀生長的權(quán)衡策略不同：未發(fā)生干旱時，夏玉米傾向于維持較高的代謝活性，一旦干旱程度大于0，夏玉米就會降低葉片代謝活性；當(dāng)干旱程度小于0.48時，夏玉米傾向于通過迅速增加葉面積來吸收較多的能量，以獲得較大的生長速率，為生殖器官的生長及產(chǎn)量形成儲備能量；當(dāng)干旱程度大于0.48時，夏玉米會減小單葉面積以減少水分散失，傾向于資源貯存以提高其生存能力。（周廣勝）

1.10 中國產(chǎn)區(qū)蘋果越冬凍害的風(fēng)險評估

越冬期凍害是影響蘋果種植分布和質(zhì)量形成的主要氣象災(zāi)害之一。利用1961—2014 年2084個氣象站點資料和凍害調(diào)查數(shù)據(jù)，基于二分類Logistic 回歸分別建立了越冬期初冬凍害和極端低溫凍害發(fā)生的概率預(yù)測模型，并按照風(fēng)險極低、低、中度、高、極高等5 個等級進行了空間劃分。結(jié)果表明，建立的初冬凍害和極端低溫凍害風(fēng)險概率預(yù)測模型均通過了Hosmer-Losmer檢驗，獨立樣本的預(yù)測準(zhǔn)確率分別達(dá)到了83.6%和91.4%。中國產(chǎn)區(qū)蘋果的越冬凍害主要以初冬凍害為主，覆蓋了除黃河故道和云南產(chǎn)區(qū)外的大部分果區(qū)，而極端低溫凍害主要分布在緯度或海拔較高的環(huán)渤海灣北部產(chǎn)區(qū)、黃土高原西北部和北疆。兩種凍害的高風(fēng)險區(qū)域面積基本相當(dāng)，中度以上風(fēng)險面積較大的省份依次為甘肅、遼寧、河北和山西，但蘋果種植面積最大的陜西和山東遭遇2種凍害特別是極端低溫凍害的風(fēng)險概率總體較低。（周廣勝）

1.11 東北春玉米積溫穩(wěn)定性影響因素與積溫模型的改進

以東北春玉米-四單19為例，獲取其生長發(fā)育資料和同期氣象觀測資料，應(yīng)用沈國權(quán)提出的非線性積溫模型（NLM）進行擬合，分析了參數(shù)選擇對積溫穩(wěn)定性的影響，提出使用平均溫度的二次函數(shù)對線性積溫模型（簡稱LM）進行修正（修正后模型稱TRM）及效果分析，并與NLM結(jié)果進行比較。為分析東北春玉米不同品種在應(yīng)用NLM時的適用性問題，結(jié)合東北地區(qū)春玉米生長發(fā)育的實際情況，以觀測年份較多、觀測地點較廣為原則增加選取3個東北春玉米品種“東農(nóng)248”“龍單13”和“丹玉13”，應(yīng)用NLM進行擬合，分析并討論參數(shù)的生物學(xué)意義及其與品種熟型的關(guān)系，對NLM進行了有效的改進及驗證。結(jié)果表明：NLM擬合時參數(shù)P越小，擬合的有效積溫越穩(wěn)定；NLM計算的積溫在年際間、地區(qū)間均存在差異，造成積溫不穩(wěn)定的主要原因是溫度強度與其他因子相關(guān)性較差；各生育期有效積溫與生育期平均溫度呈二次曲線關(guān)系，將LM的溫度二次方修正結(jié)果與NLM結(jié)果比較發(fā)現(xiàn)二次方修正的方法具有可行性；4個春玉米品種NLM均不存在無效參數(shù)，以積溫變異較小為原則確定參數(shù)P=0.5；參數(shù)K與參數(shù)Q存在顯著的相關(guān)性，說明參數(shù)K可能僅是一個統(tǒng)計參數(shù)，并沒有明確的生物學(xué)意義；積溫在品種間存在顯著差異，全生育期模型參數(shù)Q與多年站次平均有效積溫或活動積溫有較好的相關(guān)性，由于不同的積溫意味著不同的玉米品種熟型，說明參數(shù)Q與玉米品種的熟型有關(guān)，將模型參數(shù)Q和K用反映玉米品種熟型的參數(shù)（有效積溫、活動積溫）表示，建立了適用于不同品種的通用積溫模型，取得較好的應(yīng)用效果。研究成果對于農(nóng)業(yè)氣象指標(biāo)的修訂和服務(wù)能力的提升有指導(dǎo)意義。（郭建平）

1.12 不同遙感干旱指數(shù)在玉米全生育期的干旱監(jiān)測能力分析

在國家自然科學(xué)基金中英國際合作重點項目 “基于高分雷達(dá)遙感和快中子水分傳感技術(shù)發(fā)展近實時的高時空分辨的區(qū)域土壤濕度監(jiān)測方法”的支持下，對比了當(dāng)前主要的各個遙感干旱監(jiān)測指數(shù)的業(yè)務(wù)監(jiān)測能力，綜合分析當(dāng)前主要應(yīng)用的光學(xué)遙感MODIS數(shù)據(jù)干旱指數(shù)。通過遼寧西北部2014年玉米全生育期10次遙感監(jiān)測，并與地面15個土壤水分站的觀測數(shù)據(jù)做相關(guān)分析，綜合對比得出在玉米播種初期（5月）ATI指數(shù)監(jiān)測效果最好，其他時期與土壤水分（0～10 cm）相關(guān)系數(shù)最高的是TVDI指數(shù)。所以綜合來看，TVDI干旱指數(shù)遙感監(jiān)測精度較高，適合大范圍和玉米苗期以外的玉米各個生育期的干旱業(yè)務(wù)監(jiān)測（圖2）。（房世波）

1.13 雷達(dá)遙感在多云區(qū)水稻種植面積提取方法及精度評價

光學(xué)遙感在多云區(qū)的作物生長季很難獲取遙感數(shù)據(jù)，限制了南方作物種植面積等的提取，雷達(dá)遙感可以穿透云等而獲取數(shù)據(jù)，是多云區(qū)陸面信息提取的有效方法。然而如何利用作物本身的雷達(dá)回波的時間序列，分析獲取準(zhǔn)確的作物分布需要進一步研究。項目組基于早稻全生育期長時間序列的后向雷達(dá)回波的差異，獲取了江西南昌縣的2016年早稻分布面積，經(jīng)過實地GPS調(diào)查和分類混淆矩陣分析得出，VV極化下，Kappa系數(shù)為0.79。水稻的用戶制圖精度為92.42%（圖3）。（房世波）

2 農(nóng)業(yè)對氣候變化的響應(yīng)與適應(yīng)

2.1 多模式集合模擬氣候變化對玉米產(chǎn)量的影響

利用1981—2009 年東北地區(qū)海倫、長嶺、本溪3 地區(qū)農(nóng)業(yè)氣象站的歷史氣象資料和玉米作物數(shù)據(jù), 分別建立作物統(tǒng)計模型并驗證了APSIM 機理模型在研究區(qū)域的適用性。在此基礎(chǔ)上，與CMIP5計劃在RCP4.5 情景下的8個全球模式結(jié)合，基于多模式集合評估了未來2010—2039 年時段和2040—2069 年時段氣候變化對玉米產(chǎn)量的可能影響（相對于1976—2005年基準(zhǔn)時段）。研究結(jié)果表明，APSIM模型對玉米生長發(fā)育和產(chǎn)量形成有很好的模擬能力。玉米生育期的模擬誤差（RMSE）為3～4 d，產(chǎn)量的RMSE 為0.6～0.8 t/hm2。建立的產(chǎn)量統(tǒng)計模型表明，玉米出苗階段（5月中旬）的溫度增加對產(chǎn)量增加有積極作用，而開花到成熟階段（7月中旬到9月上旬）的溫度和降水的增加、光照的不足均不利于產(chǎn)量增加。與1976—2005年基準(zhǔn)時段相比，氣候因素影響下2010—2039 年玉米減產(chǎn)3.8%（海倫）～7.4%（本溪），減產(chǎn)的概率為64%（長嶺）～73%（本溪）；2040—2069 年時段減產(chǎn)6.4%（海倫）～10.5%（本溪），減產(chǎn)的概率為74%（海倫）～83%（本溪）。未來2010—2039 年時段和2040—2069年時段基于機理模型模擬的產(chǎn)量降低分別為6.6%（海倫）～8.9%（本溪）和9.7%（海倫）～13.7%（本溪），均高于相應(yīng)時段基于統(tǒng)計模型得到的0.9%（海倫）～6.0%（本溪）和2.0%（長嶺）～7.3%（本溪）減產(chǎn)結(jié)果。（張祎）

2.2 未來氣候變化對我國小麥生育期和產(chǎn)量的可能影響

利用中國農(nóng)業(yè)氣象（小麥）模式（CAMM）探討了未來氣候變化對中國小麥生育期和產(chǎn)量的可能影響。結(jié)果表明，未來氣候變化對中國小麥生育期長度的影響并不明顯，而品種改良才是更重要的影響因素。原因在于隨著播種期的調(diào)整和返青期的變化，作物整個生育期的溫度并未發(fā)生明顯改變。生育期的延長并不必然導(dǎo)致產(chǎn)量升高，這可能是由于營養(yǎng)生長期旺長導(dǎo)致生殖生長期弱長的原因。由于氣候預(yù)測和模型本身的不足，這一評估結(jié)果仍存在較大的不確定性（圖4）。（馬玉平）

2.3 我國北方春玉米生產(chǎn)對氣候變化的響應(yīng)識別研究

基于內(nèi)蒙古土默特左旗農(nóng)業(yè)氣象試驗站春玉米多年田間試驗數(shù)據(jù)和逐日氣象數(shù)據(jù)，分析了農(nóng)業(yè)生產(chǎn)系統(tǒng)模型APSIM在內(nèi)蒙古玉米產(chǎn)區(qū)的適應(yīng)性，確定影響玉米發(fā)育期的關(guān)鍵環(huán)境因子，探討發(fā)育期對環(huán)境因子的響應(yīng)規(guī)律。結(jié)果表明：（1）驗證后的APSIM玉米模型在內(nèi)蒙古土默特左旗具有較好的適應(yīng)性。（2）不同發(fā)育期，對環(huán)境影響因子的響應(yīng)不同；影響開花期和成熟期日序遲早的共同主控因子首先是溫度，其次為相對濕度和潛在蒸散，最后為風(fēng)速；成熟期日序還與生育期內(nèi)總降水量的變化成顯著正相關(guān)；開花期日序?qū)ζ骄畹蜏囟茸兓淖兓顬槊舾校怀墒炱谌招驅(qū)ζ骄罡邷囟茸兓顬槊舾?，依次為平均溫度，土壤表層平均溫度和平均最低溫度。?）分別建立了反映春玉米關(guān)鍵發(fā)育期對環(huán)境因子響應(yīng)關(guān)系的統(tǒng)計模型。以上結(jié)果為今后在內(nèi)蒙古玉米產(chǎn)區(qū)開展解析春玉米生長發(fā)育進程及產(chǎn)量形成的限制因素等研究提供了技術(shù)支撐。（趙俊芳）

2.4 東北玉米適應(yīng)氣候變化措施對生產(chǎn)潛力的影響

為探求東北玉米未來如何更好地適應(yīng)氣候變化，本研究采用抗逆品種和推遲播種期2種適應(yīng)措施，結(jié)合區(qū)域氣候模式模擬的2010—2099年間RCP4.5、RCP8.5這2種濃度路徑逐日氣象資料，分析不同氣候變化情景下東北玉米適應(yīng)措施的生產(chǎn)潛力變化。結(jié)果表明，2010—2099年，東北區(qū)玉米氣候生產(chǎn)潛力的空間分布特征基本為東南向西北減小的趨勢，RCP4.5情景下東北玉米生產(chǎn)潛力高于RCP8.5情景，且RCP8.5情景下出現(xiàn)極低值年份明顯多于RCP4.5情景。所有抗逆品種的玉米生產(chǎn)潛力均高于原有品種，RCP4.5情景下耐高溫品種的玉米生產(chǎn)潛力更高，在RCP8.5情景下，耐旱品種表現(xiàn)更好，雙耐（耐高溫、耐旱）品種的玉米生產(chǎn)潛力在2種氣候變化情景下均最高。RCP4.5情景下，推遲播種均出現(xiàn)增產(chǎn)情況，其中，推遲30～40 d播種的玉米增產(chǎn)率達(dá)到最大；RCP8.5情景下，部分地區(qū)出現(xiàn)減產(chǎn)情況。由此說明了適當(dāng)推遲播種期有利于提高玉米氣候生產(chǎn)潛力，但地區(qū)間存在差異。（郭建平）

2.5 華北平原冬小麥生育期溫度條件變化趨勢分析

統(tǒng)計分析了華北平原42個農(nóng)業(yè)氣象觀測站1981—2010年的月平均氣溫、冬小麥發(fā)育日期和各主要生育階段平均氣溫的變化特征。結(jié)果表明，研究期內(nèi)華北平原冬小麥生長季的10月、12月和2—6月增溫趨勢顯著，氣候傾向率在0.44～1.05 ℃/10a（P＜0.05），2月平均氣溫上升線性傾向率最大。氣候變暖使冬小麥越冬階段（50%的站點表現(xiàn)為顯著增溫，增溫線性傾向率在0.45～1.18 ℃/10a）和返青—拔節(jié)階段（64%的站點增溫顯著，增溫速率達(dá)0.49～1.57 ℃/10a）的平均溫度顯著升高，從而導(dǎo)致冬小麥拔節(jié)—成熟日期顯著提前，但冬小麥冬前生長階段和拔節(jié)—成熟階段的平均溫度則未呈現(xiàn)上升趨勢。冬小麥冬前生長階段的溫度環(huán)境因播種期適應(yīng)性推遲而保持基本穩(wěn)定，拔節(jié)—成熟階段平均溫度變化不明顯則歸因于發(fā)育期前移和當(dāng)?shù)貧鉁氐募竟?jié)性變化特點。（譚凱炎）

2.6 甘肅農(nóng)業(yè)對氣候變化的適應(yīng)與風(fēng)險評估

系統(tǒng)闡述了甘肅省氣候變化、農(nóng)業(yè)氣候資源、農(nóng)業(yè)氣象災(zāi)害、農(nóng)業(yè)種植制度、農(nóng)業(yè)氣象災(zāi)損及其風(fēng)險的時空演變，探討了甘肅省農(nóng)業(yè)適應(yīng)氣候變化的對策措施，可為推動甘肅農(nóng)業(yè)健康可持續(xù)發(fā)展及氣象精準(zhǔn)扶貧脫貧提供決策依據(jù)。

（1）氣候變化趨勢。1961年以來，甘肅省平均氣溫、平均最高氣溫和平均最低氣溫均呈上升趨勢，且平均最高氣溫上升幅度最大、平均氣溫上升最小。各季增溫明顯，其中冬季升溫幅度最大，夏季升溫幅度最小。平均年降水量和年日照時數(shù)均呈不顯著的減少趨勢。

（2）極端氣候事件變化趨勢。1961年以來，除祁連山西段呈下降趨勢外，甘肅省其他地區(qū)極端最高氣溫均呈升高趨勢；日最高氣溫≥35 ℃日數(shù)除祁連山區(qū)和甘南高原北部外均呈增加趨勢，增加速率為0.3 d/10a。年極端最低氣溫除酒泉市西北部外均呈升高趨勢，其中甘南高原增溫最為明顯。最長連續(xù)無降水日數(shù)在河西中東部、隴中中北部和西南部、隴東東部、甘南中部均呈增加趨勢。

（3）農(nóng)業(yè)氣候資源演變。1961年以來，甘肅省日均氣溫≥0 ℃和≥10 ℃的初日均呈不同程度的提前趨勢，終日呈不同程度推遲趨勢。日均氣溫穩(wěn)定通過0 ℃和10 ℃期間的降水量分布呈東南向西部遞減趨勢，且在河西大部呈增多趨勢、河?xùn)|大部呈減少趨勢；積溫均呈顯著增加趨勢；日均氣溫穩(wěn)定通過0 ℃期間的日照時數(shù)在甘肅大部呈增加趨勢，日均氣溫穩(wěn)定通過10 ℃期間的日照時數(shù)在全省各地變化趨勢不明顯。

（4）農(nóng)業(yè)氣候資源變化對作物產(chǎn)量的影響。1980—2014年，冬麥區(qū)生育期平均氣溫升高使單產(chǎn)減少7.5%；氣溫日較差增大使單產(chǎn)減少6.7%；降水量減少使單產(chǎn)減少0.4%。春小麥平均氣溫升高使單產(chǎn)減少4.3%，氣溫日較差減小使單產(chǎn)增加0.7%，降水量增加使單產(chǎn)增加0.1%。玉米生育期平均氣溫升高使單產(chǎn)減少5.0%，氣溫日較差增大使單產(chǎn)減少1.2%，降水量減少使單產(chǎn)減少0.2%。1985—2014年，馬鈴薯生育期平均氣溫升高使單產(chǎn)減少1.8%，氣溫日較差減小使單產(chǎn)增加0.3%，降水量減少使單產(chǎn)減少0.3%。

（5）農(nóng)業(yè)氣象災(zāi)害演變。1961年以來，甘肅省氣象干旱發(fā)生頻率與強度呈明顯增加趨勢，春旱與伏旱發(fā)生范圍呈明顯擴大趨勢，春末夏初旱與秋旱發(fā)生范圍呈明顯縮小趨勢。全省大風(fēng)（站）日數(shù)總體呈減少趨勢，但2007年后呈增加趨勢；暴雨日數(shù)呈不顯著減少趨勢，主要出現(xiàn)在河?xùn)|地區(qū)；霜凍（站）日數(shù)呈先增后減趨勢，20世紀(jì)80年代后減少趨勢顯著，其中初霜凍在2005年以來進入歷史上低值時段。

（6）農(nóng)業(yè)氣象災(zāi)損時空演變。1961年以來，農(nóng)業(yè)干旱災(zāi)害發(fā)展具有面積增大和危害程度加劇的趨勢，干旱受災(zāi)、成災(zāi)和絕收率（25.2%、14.1%和2.2%）均明顯高于全國平均（15.0%、8.1%和1.7%），增加速率（0.16%/10a、0.15%/10a和0.05%/10a）也高于全國平均。風(fēng)雹災(zāi)害、暴雨洪澇災(zāi)害和低溫冷害的綜合損失率亦均呈增加趨勢，增加速率分別為0.29%/10a、0.45%/10a和0.72%/10a。

（7）農(nóng)業(yè)病蟲害演變趨勢及其影響。1981—2015年，氣候變化總體有利于甘肅省農(nóng)業(yè)病蟲草鼠害發(fā)生面積擴大，危害程度加劇。病蟲草鼠害、病害、蟲害、草害和鼠害發(fā)生面積率分別以0.31/10a、0.20/10a、0.08/10a、0.06/10a、-0.03/10a的速率變化。農(nóng)區(qū)病害、蟲害和鼠害的發(fā)生面積率主要受溫度影響，草害發(fā)生面積率主要受降水日數(shù)影響。無論是單產(chǎn)還是總產(chǎn)，病蟲害危害可能損失遞增率均是馬鈴薯＞玉米＞小麥，馬鈴薯病害＞蟲害，玉米蟲害＞病害，小麥病害＞蟲害。因此，未來需高度關(guān)注馬鈴薯和玉米的病蟲害，尤其是馬鈴薯病害和玉米蟲害的影響，同時也需注意小麥病害對單產(chǎn)的影響，進行重點防控治理。

（8）農(nóng)業(yè)種植制度演變及其影響。與1951—1980年相比，1981—2013年一年兩熟制作物可種植北界不同程度地北移，北移最大的地區(qū)有隴南、隴東和甘南高原。冬小麥種植北界不同程度西擴，西擴最大的地區(qū)為河西地區(qū)和甘南高原。冬小麥、玉米、春小麥、馬鈴薯等一年一熟種植模式轉(zhuǎn)變?yōu)槎←湣挠衩滓荒陜墒旆N植模式的變化可使單產(chǎn)大幅增加，隴南地區(qū)的冬小麥、玉米、馬鈴薯增產(chǎn)率分別達(dá)153.53%、65.13%和149.69%；隴中地區(qū)的冬小麥、玉米、春小麥、馬鈴薯增產(chǎn)率分別達(dá)84.56%、91.27%、76.42%和83.02%。

（9）農(nóng)業(yè)適應(yīng)氣候變化的對策措施。針對氣候變化背景下甘肅省農(nóng)業(yè)生產(chǎn)面臨的農(nóng)業(yè)氣候資源新特點與農(nóng)業(yè)氣象災(zāi)害的新形勢，本研究（書）提出了一系列氣候資源高效利用的新模式，以有效緩解氣候變化對農(nóng)業(yè)生產(chǎn)的不利影響，甚至將不利影響轉(zhuǎn)變?yōu)橛行У馁Y源充分利用，提升甘肅農(nóng)業(yè)生產(chǎn)水平，服務(wù)于甘肅綠色扶貧脫貧。主要包括：優(yōu)化土地利用格局，充分利用光熱資源；調(diào)整作物種植制度，主動適應(yīng)氣候變化；選育適宜作物品種，科學(xué)應(yīng)對暖干化與病蟲害影響；調(diào)整作物復(fù)種指數(shù)，提高耕地資源利用效率；調(diào)整作物品種布局，充分利用水熱資源優(yōu)勢；針對氣候變化分異，調(diào)整農(nóng)區(qū)生產(chǎn)管理方式。（周廣勝）

3 生態(tài)氣象監(jiān)測

3.1 典型草原植物響應(yīng)干旱的指標(biāo)及閾值

以CO2濃度和溫度升高為標(biāo)志的全球氣候變化導(dǎo)致干旱等極端天氣氣候事件發(fā)生頻率增加，對植物的影響已經(jīng)產(chǎn)生且將持續(xù)到可預(yù)見的未來。開展植物響應(yīng)干旱過程的研究有利于監(jiān)測植物干旱發(fā)生、發(fā)展趨勢，評估植物干旱狀況，具有較強的實用價值。本研究基于典型草原植物外場模擬試驗和降水變化與CO2濃度交互作用的開頂式生長箱模擬試驗，從植物響應(yīng)干旱過程的敏感指標(biāo)受到脅迫的先后順序、臨界閾值、可塑性和關(guān)聯(lián)度以及光合生理生態(tài)機制方面進行了定量分析及復(fù)水后各指標(biāo)的變化研究；并探究了干旱指標(biāo)及其閾值在高CO2濃度情景下的變化。主要研究結(jié)果如下：

（1） 羊草和克氏針茅生物學(xué)指標(biāo)都是光合速率（Pn）和葉含水量最先受到干旱脅迫，土壤相對濕度閾值在49%～54%之間。羊草單葉面積先于葉數(shù)受到干旱脅迫，但克氏針茅葉數(shù)先于單葉面積受到干旱脅迫。

（2） 羊草和克氏針茅的形態(tài)指標(biāo)干旱可塑性顯示，兩者植株葉面積在干旱過程中可塑性較大，羊草株高的可塑性大于葉數(shù)，但是克氏針茅葉數(shù)可塑性大于株高。羊草總?cè)~面積、株高、葉數(shù)和地上生物量4個指標(biāo)與干旱關(guān)聯(lián)度差異不大；克氏針茅葉數(shù)與干旱關(guān)聯(lián)度最大。在葉片的生理特征中，羊草蒸騰速率（Tr）與干旱關(guān)聯(lián)度最大，克氏針茅飽和水汽壓虧缺（VPD）與干旱關(guān)聯(lián)度最大。

（3） 干旱脅迫下，首先飽和水汽壓虧缺（VPD）增加，植物葉片氣孔部分關(guān)閉，阻力增加，氣孔導(dǎo)度（Gs）下降，從而使蒸騰（Tr）降低，減少失水速率，凈光合速率（Pn）下降以氣孔限制為主，但隨著干旱的持續(xù)，葉綠素含量降低，胞間CO2濃度（Ci）增加，Ci/Ca（大氣CO2濃度）值增大，非氣孔限制上升為主導(dǎo)因素。

（4） 干旱復(fù)水后羊草和克氏針茅光合速率（Pn）和葉含水量以及羊草的葉綠素含量可以迅速恢復(fù)至充分供水水平，脅迫效應(yīng)是可逆的；總?cè)~面積和單葉面積復(fù)水后逐漸增大，但最終仍遠(yuǎn)小于充分供水水平。羊草的光合速率（Pn）、氣孔導(dǎo)度（Gs）、飽和水汽壓虧缺（VPD）和蒸騰速率（Tr）等指標(biāo)的變化程度要大于克氏針茅；光合速率（Pn）和葉片含水量受到干旱脅迫的時間比克氏針茅早，土壤相對濕度閾值比克氏針茅高，復(fù)水后，羊草光合速率（Pn）和地上生物量恢復(fù)的程度也大于克氏針茅，表現(xiàn)出羊草對干旱和復(fù)水的敏感性要大于克氏針茅。

（5） 在當(dāng)前和高CO2濃度情境下，總?cè)~面積/葉片數(shù)、葉水勢/葉片含水量和地上生物量可分別從形態(tài)、生理和生物量積累3方面反映克氏針茅水分狀態(tài)。CO2濃度升高條件下干旱指標(biāo)的敏感性會發(fā)生變化。CO2升高將降低上述5個干旱指標(biāo)的降水閾值。（周廣勝）

3.2 中國草地固碳的時空動態(tài)及其控制機制

中國草地占國土面積的40%以上，分布范圍廣泛，是氣候敏感區(qū)和生態(tài)脆弱區(qū)，也是受人為活動影響最為嚴(yán)重的區(qū)域。準(zhǔn)確評估中國草地固碳的時空動態(tài)、揭示其控制機制是草地碳收支評估的難點與熱點。本研究基于2011—2013年中國草地碳收支野外調(diào)查數(shù)據(jù)以及相關(guān)文獻(xiàn)資料，校準(zhǔn)并驗證陸地生態(tài)系統(tǒng)模型（TEM 5.0），模擬了1961—2013年中國草地固碳的時空格局特征，探討了氣象要素（溫度、輻射和降水）的年際變化及其空間分布對中國草地固碳的影響，分析了未受干擾情況下，當(dāng)前氣候背景、RCP4.5和RCP8.5氣候情景的中國草地的固碳潛力。主要結(jié)論如下：

（1） 1961—2013年間，面積為3.95×106km2的中國草地表現(xiàn)為碳匯，強度為19.06 Tg C/a，53年間共固碳1.01 Pg C。其中，內(nèi)蒙古草原（5.77 Tg C/a）、新疆草地（5.44 Tg C/a）和青藏高原草地（5.16 Tg C/a）是中國草地碳匯的主體，三者對碳匯的貢獻(xiàn)達(dá)85.9%。由于草地植物凈初級生產(chǎn)力（NPP）和土壤呼吸（RH）對年際間溫度變化的敏感性不同，中國草地碳收支（NEP）表現(xiàn)出顯著的年際變化趨勢。在年代際總量上，中國草地各年代際均表現(xiàn)為碳匯，其中2001—2010年的碳匯強度最大（63.33 Tg C/a），占1961—2010年5個年代總碳吸收值的57.6%。研究區(qū)域內(nèi)碳匯增加表現(xiàn)為植被碳庫和土壤有機碳庫的增加。

（2） 中國草地單位面積多年平均NEP為4.84 g C/（m2a），研究區(qū)域整體表現(xiàn)為碳匯。從空間分布來看，碳匯主要分布在青藏高原西部、內(nèi)蒙古草原東北部和新疆草地北部，而碳源主要分布在青藏高原東南邊界、東北邊界以及西部邊界，碳源區(qū)僅占研究區(qū)域總面積的13.5%。研究時段內(nèi)，中國草地大部分區(qū)域NEP呈增加趨勢，占研究區(qū)域總面積的82.5%，主要集中在青藏高原西北部、新疆草地北部和內(nèi)蒙古草原東北部等碳匯區(qū)，NEP減少的區(qū)域主要分布在內(nèi)蒙古草原西北部、青藏高原西南部和東北部以及南方草地區(qū)等碳匯能力相對較弱的地區(qū)，占研究區(qū)域總面積的17.5%。

（3） 研究時段內(nèi)，中國草地區(qū)域年平均溫度呈顯著增加趨勢，降水量年際波動較大，在空間格局上自東南向西北遞減，年代際間變化呈干燥（60年代、70年代）—濕潤（80年代、90年代）—干燥（21世紀(jì)初期）趨勢。區(qū)域年均NEP與年均溫度、年降水量呈正相關(guān)關(guān)系，與輻射相關(guān)性不顯著。年際間溫度和降水變化是中國草地NEP年際波動的關(guān)鍵原因。

（4） 1961—2013年中國草地年均固碳潛力為18.41 Pg C/a，其中土壤年均固碳潛力為18.12 Pg C/a，植被年均固碳潛力為0.29 Pg C/a。從單位面積固碳潛力的空間分布上看，各年代草地固碳潛力的空間格局大體一致，青藏高原和新疆準(zhǔn)噶爾盆地的草地固碳潛力較大，固碳潛力一般大于10000 g C/（m2a）；南方草地固碳潛力一般在1000～5000 g C/（m2a）之間；其他草地固碳潛力較小。

（5） 無人為干擾影響的RCP4.5和RCP8.5氣候情景下，中國草地碳儲量在研究時段內(nèi)均呈顯著增加趨勢，平均增加速率分別為61.0 Tg C/（m2a）（RCP4.5）和234.2 Tg C/（m2a）（RCP8.5）；且2種氣候下碳密度變化的空間分布較一致，2011—2040年間中國草地生產(chǎn)系統(tǒng)碳密度以增加為主，增加較顯著的區(qū)域主要是青藏高原草地和內(nèi)蒙古草原。

（6） RCP4.5氣候情景下，2011—2040年中國草地年均固碳潛力為18.01 Pg C/a，其中草地土壤固碳潛力約17.78 Pg C/a，草地植被固碳潛力約0.24 Pg C/a。RCP8.5氣候情景下，2011—2040年中國草地年均固碳潛力為18.00 Pg C/a，其中草地土壤固碳潛力約17.79 Pg C/a，草地植被固碳潛力約0.20 Pg C/a。兩種氣候情景下，中國草地固碳潛力的分布格局相似。以2030年代為例，新疆準(zhǔn)噶爾盆地的草地單位面積固碳潛力最大；其次是新疆中部天山和南部昆侖山、部分內(nèi)蒙古草地，以及青藏高原南部、中部和東南部草地；余下草地的單位面積固碳潛力很小，一般不超過5000 g C/（m2a），有些地方近似為零。（周廣勝）

4 成果應(yīng)用

4.1 基于集合卡爾曼濾波的PyWOFOST模型在玉米估產(chǎn)中的應(yīng)用

改進并評估了基于集合卡爾曼濾波（EnKF）構(gòu)建的遙感信息-作物模型耦合模型（PyWOFOST）在東北玉米估產(chǎn)中的適用性。該成果已于2017年底在中國農(nóng)業(yè)科學(xué)院農(nóng)業(yè)信息研究所農(nóng)作物生長機理和空間遙感同化技術(shù)的縣域產(chǎn)量風(fēng)險預(yù)測模型的服務(wù)中得到了應(yīng)用。（趙艷霞）

圖1 海南省辣椒災(zāi)害綜合風(fēng)險區(qū)劃Fig.1 Distribution of integrated risk for major disasters in paprika pepper in Hainan Province

圖2 玉米全生育期不同遙感干旱指數(shù)與土壤水分（0～10 cm）相關(guān)系數(shù)及其與云量(%)的關(guān)系Fig.2 Comparison of the correlational relationship (r value) of each index and the average percentage of cloud coverage (%) in the whole study area from 9 May to 22 September, 2014

圖3 2016年江西南昌縣早稻分布信息Fig.3 Spatial distribution of early rice information in 2016 in Nanchang County, Jiangxi Province

圖4 未來氣候變化對我國小麥生育期和產(chǎn)量的可能影響Fig.4 The possible influence of future climate change on the growth period and yield of wheat in China

Progress in Ecology and Agrometeorological Research

1 Agroclimatic resources and agrometeorological disasters

1.1 Construction of the initial version of the Chinese AgroMeteorological (wheat) Model (CAMM)

The basic framework of Chinese Agrometeorological (wheat) Model (CAMM) was determined from the aspects of mechanism processes (crop growth process, regional characteristic process), construction method(insert and pull type modeling, data assimilation), and application mode (real-time operation and Internet).The basic function, product type, and content of the model were clarified.We had preliminarily completed the regionalization of 22 parameters or initial values of the model.Such as base grid, planting area, crop growth,development, soil parameters and so on.The models of the nonlinear response of the initial value of wheat development to the temperature, the variety improvement, the change of winter wheat height were established.The initial version of the Chinese agrometeorological (wheat) model was constructed.(Ma Yuping)

1.2 Establishment of waterlogging level indices based on the rainfall process with corn and rape

We have constructed the corn waterlogging level indices based on the rainfall process and presented a method to judge the catastrophe of rape waterlogging process.On the puzzle of existing corn and rape indices difficult to be applied to meteorological work, a method of constructing the regional summer corn and spring corn waterlogging grading indices based on risk reverse course analysis has presented, the effect of current precipitation process and antecedent precipitation on corn waterlogging formation has been parsed and the level indicators of corn waterlogging in the different growing phases based on current precipitation process and antecedent precipitation in Jianghuai summer corn, Jianghan and Jiangnan West Region spring corn has been built up.The study explored a synoptic method for real-time monitoring and assessment of corn waterlogging and provided technical support for monitoring and assessment business of corn waterlogging.Based on the changes of impending hazard state and affected state in the rape waterlogging process, dynamic accumulative effect of precipitation and continuous cloudy days has been used to simulate the development of waterlogging and the critical line between impending hazard and affected in different growing periods has been found.A day-to-day distinction index based on the rape waterlogging catastrophe process has been built then.It can be used to determine the occurrence time of disaster and duration of catastrophe and the disaster duration after the precipitation, which provided methodological support for dynamic monitoring of catastrophe of rape waterlogging.(Huo Zhiguo)

1.3 Integrated risk assessment of meteorological disasters with paprika pepper in Hainan Province

For the integrated risk assessment of meteorological disasters, including drought, flood and chilling with paprika pepper in Hainan Province, the methods to quantify the hazard, sensitivity, vulnerability and prevention capability with different disasters have been put forward.Based on the Analytic Hierarchy Process and Entropy method, the combined weight was given to each disaster factor, thus an integrated disaster risk assessment model was developed and applied at regional level, and then mapped the distribution of integrated risk for major disasters in paprika pepper in Hainan.Drought and chilling sensitivity had a similar spatial distribution, which decreased from the central to the coastal regions while flood sensitivity was the opposite.High vulnerability of the disasters mainly occurred in the central regions, similar to low prevention capability.Eastern Hainan suffered from high integrated damage risk (Fig.1).(Huo Zhiguo)

1.4 Uncertainty in simulating the impact of cultivar improvement on winter wheat phenology in the North China Plain

The phenology model is one of the major tools in evaluating the impact of cultivar improvement on crop phenology.Understanding uncertainty in simulating the impact is an important prerequisite for reliably interpreting the effect of cultivar improvement and climate change on phenology.However, uncertainty induced by different temperature response functions and parameterization methods have not been properly addressed.Based on winter wheat phenology observations during the years 1986–2012 in 47 agro-meteorology observational stations in the North China Plain (NCP), the uncertainty of the simulated impacts caused by four widely applied temperature response functions and two parameterization methods were investigated.The functions were firstly calibrated using observed phenology data during 1986–1988 from each station using two parameterization methods and were then used to quantify the impact of cultivar improvement on wheat phenology during 1986–2012.Results showed that all functions and all parameterization methods could achieve acceptable precision (RMSE＜3 days for all functions and parameterization methods), however,substantial differences still exist in the simulated impacts between different functions and parameterization methods.For vegetative growth period, the simulated impact is 0.20 day 10yr–1(95% conf i dence interval –2.81 – 3.22 day 10yr–1) across the NCP, while for reproductive period, the value is 1.50 day 10yr–1(–1.03 – 4.02 day 10yr–1).Further analysis showed that uncertainty can be induced by both different functions and parameterization methods, while the former has greater influence than the latter.During vegetative period,there is a significant positive linear relationship between ranges of the simulated impact and growth period average temperature, while during reproductive period, the relationship is polynomial.This highlights the large inconsistency that exists in most impact quantifying functions and the urgent need to carry out field experiment to provide realistic impacts for all functions.Before applying a simulated effect, we suggest that the function should be calibrated over a wide temperature range.(Wu Dingrong)

1.5 Investigation on water productivity of winter wheat based on MODIS and EBAL in the Huang-Huai-Hai Plain, China

Understanding how the productivity of water can be increased is widely accepted to be a high priority where water resources are currently scarce and/or over-exploited in China.As the primary data source, MODIS remote sensing, statistics, meteorological data, crop growth period data, and ground truth-data from Jan.2011 to Dec.2012 were used in actual evapotranspiration estimation, yields rasterizing, and water productivity calculation for winter wheat in the Huang-Huai-Hai Plain (3H Plain).The statistical data were synthesized to calculate district-state-level land productivity, which is then further extrapolated to pixel-level values using a MODIS NDVI image based on a crop dominance map.Spatial variation of crop water productivity was investigated in order to reveal the key factors of crop water productivity.The regional average value of water productivity for winter wheat was detected to be 1.21 kg ? m–3, with the higher value in Beijing, Tianjin,northern part of Shandong Province and southern part of Hebei Province.The pronounced relationship of spatial correlation of the yield and water productivity for winter wheat indicated that the increasing yield governed the increment of water productivity for winter wheat in the coastal land-farming-f i shing area, low plain-hydropenia irrigable land and dry land zone and basin- irrigable land and dry land.The increasing water productivity for winter wheat was recognized to be controlled by rather the increment of yield than the reduction of actual evapotranspiration for winter wheat in the piedmont plain-irrigable land.Whereas, the pronounced relationship of negative correlation of water productivity and actual evapotranspiration for winter wheat described that only the reduction of actual evapotranspiration was responsible for the increment of water productivity for winter wheat in hill-wet hot paddy-paddy field.The results provide a basis information for agricultural water management, improvement of crop water productivity and choice of adaptive mechanism under climate change in Huang-Huai-Hai plain.(Yang Jianying)

1.6 Temporal downscaling of crop coefficients for winter wheat in the North China Plain: A case study at the Gucheng Agro-Meteorological Experimental Station

The crop coefficient (Kc) is widely used for operational estimation of actual evapotranspiration (ETa)and crop water requirements.The standard method for obtainingKcis via a lookup table from FAO-56 (Food and Agriculture Organization of the United Nations Irrigation and Drainage Paper No.56), which broadly treatsKcas a function of four crop-growing stages.However, the distinctive physiological characteristics of overwintering crops, such as winter wheat (Triticum aestivumL.), which is extensively planted in the North China Plain (NCP), are not addressed in this method.In this study, we propose a stage-wise method that accounts forKcvariations for winter wheat at each critical phenological stage, thereby estimatingKcat finer temporal scales.Compared with the conventional FAO method, the proposed stage-wise method successfully captures the bimodal pattern inKctime series for winter wheat, which is shown at both tenday and phenological time scales.In addition, the accuracies of the proposed stage-wiseKcmethod and the FAO method were evaluated using micro-meteorological measurements of ETacollected at the Gucheng agrometeorological experimental station in the NCP.Using a leave-one-out strategy, the evaluation revealed that the stage-wise method significantly outperformed the FAO method at both daily and critical phenological time scales, with root-mean-square errors in ETafor the stage-wise method and the FAO method being 0.07 mm day?1and 0.16 mm day?1, respectively, at the daily time scale, and 0.01 mm day?1and 0.27 mm day?1at the critical phenological time scale.Generally, the FAO method underestimates ETaduring the initial stage and overestimates ETaduring both the development and mid-season stages.It is shown that the proposed stagewise method is important for the water-stressed NCP where precision irrigation is highly desirable, especially during the critical phenological stages.Results from this study provide insight into accurate estimation of water requirements for winter wheat at phenological time scales.(Wang Peijuan)

1.7 Estimation of crop evapotranspiration under standard conditions for winter wheat in the Huang-Huai-Hai Plain

Crop evapotranspiration under standard conditions (ETc) is defined as the evapotranspiration from disease-free, well-fertilized crops grown in large fields, under optimum soil water conditions, and achieving full production under the given climatic conditions.The calculation of ETcconsiders crop and local surface conditions.ETcis the theoretical upper limit of actual evapotranspiration for actual local surface coverage,ensuring objective analysis on crop water requirements and agricultural drought.To summarize the spatial and temporal distribution characteristics and their causes of ETc, daily ETcis calculated based on Penman-Monteith method using meteorological data and satellite remote sensing data from 2000 to 2015.The meteorological data are provided by 27 meteorological stations in the winter wheat growing area of the Huang-Huai-Hai Plain.The satellite remote sensing data area extracted from NASA MODIS products (LAI (MOD15A2) and Albedo (MCD43C3)) at the corresponding location of 27 meteorological stations.ETc_kcis calculated based on single crop coefficient approach recommended by FAO.Results show that daily dynamic changes of ETcand ETc_kcare consistent in the regional time scale.However, compared with ETc_kc, ETchas a spatial distribution corresponding to the objective reality.The growth period of winter wheat is divided into five stages: Before wintering stage, wintering stage, returning green-jointing stage, heading stage and milky maturity-maturity stage.With the spatial distribution characteristic of higher in the south and lower in the north, the average daily ETcin the whole winter wheat season, wintering stage and returning green-jointing stage are 1.95 mm,0.46 mm and 2.74 mm, respectively.The average value of ETcis 1.23 mm before wintering stage, and the whole fluctuation of ETcin the Huang-Huai-Hai Plain is small.There is no significant fluctuation in ETcin heading stage and milky maturity-maturity stage except for the central part of the Huang-Huai-Hai Plain.The average values of ETcare 4.71 mm and 3.72 mm in these two growth stages, respectively.In terms of spatial distribution, extremely significant positive correlation is shown between LAI and ETcin all growth period.In wintering stage, returning green-jointing stage, and milky maturity-maturity stage, ETcalso shows a higher significant negative correlation coefficient with albedo.During the whole growth period of winter wheat, ETchas a higher partial correlation coefficient with LAI and water vapor pressure.These results can provide basic data for drought monitoring and wet or dry climate zoning in China, and also provide a new idea for the actual evapotranspiration estimation.(Wang Peijuan)

1.8 Developing a method to determine critical soil water contents of growth indicators of summer maize in the seedling stage

Deficit in soil moisture is the main cause of crop drought.Accurate determination of the thresholds of crop growth change responding to soil water is of great importance for identification and monitoring of the occurrence and development of crop drought.Based on the data of a progressive drought simulation experiment with 6 primary soil water gradients on summer maize, one-way MANOVA method was performed to identify maize growth indicators sensitive to soil water variations, and the method of normal population statistic tolerance limits was proposed to identify the threshold of maize growth indicator responding to the change of soil water content.The results showed that the stem and leaf moisture content, transpiration rate,net photosynthetic rate, stomatal conductance, and leaf area of summer maize were maize growth indicators responding to the change of soil water content in the seedling stage, and their corresponding critical soil water contents at the top of 30 cm were 72%, 65%, 62%, 60%, 58%, and 46%, respectively, indicating the changing order of stem moisture content, leaf moisture content, transpiration rate, net photosynthetic rate, stomatal conductance, and leaf area responding to the decrease of soil water content.These results could provide reference to the monitoring and quantitative assessment of the development of drought to summer maize during the seedling stage, and also a way to determine the thresholds of ecosystems responding to climate change.(Zhou Guangsheng)

1.9 Revealing the effects of drought on the trade-off growth of leaf traits of summer maize in the seedling stage

Changes in leaf traits and their trade-offs are reflections of how plants use resources as well as their adaption strategies to environmental changes.Drought is known as one of the greatest threats to the growth and development of plants worldwide, while leaf is one of the most sensitive parts of plants to drought.At present,the influences of drought on leaf traits have been widely studied, but less involved in the trade-off growth between different leaf traits during the progress of drought, which limits the understanding of plant adaptation strategies in the arid environment.Maize was one of the leading crops in the world, while it was also the most susceptible one to drought.Understanding how the leaf trait of maize changes during the progress of drought and their trade-off growth would help promote the understanding of the adaptation strategies of maize plant to drought and contribute to make targeted drought-resisting and drought-relief measures.Therefore, based on the consecutive drought simulation experiments with 6 primary soil water levels on maize that conducted in 2014, we investigated the changes of 5 key leaf traits of maize, they were the green leaf area, the leaf dry mass,the specific leaf weight, the leaf water content, and the effective leaf number, as well as the trade-off growth between green leaf area and effective leaf number, green leaf area and specific leaf weight, leaf dry mass and leaf water content, respectively.Besides, we newly developed an indicator to quantitatively evaluate drought degree, which consisted of available soil water content, readily available soil water content, and drought duration time, comprehensively involved factors that related to intensities and progresses of water deficit, such as soil texture, meteorological environment, plant water requirement, and so on.Based on this indicator, we quantitatively characterized how these leaf traits change and their trade-off strategies under different drought degrees.The results showed that every leaf trait investigated here would be affected by drought, however,the extent to which they were affected relied on the drought degree.The leaf traits would not be affected significantly when the drought degree was below 0.21, and their quantities would be affected while their tendencies would not when the drought degree was between 0.21–0.76.When the drought degree increased to 0.76–0.91, the effective leaf number, dry mass of leaf, green leaf area, and leaf water content would decrease prematurely due to the fact that the growth of new leaves failed to compensate the losses caused by senescence of old leaves.When the drought degree exceeded 0.91, it would lead to stagnation in leaf growth, and leaf traits would barely change.The adaptive growth of leaf under drought embodied its trade-off between rapid growth and survival, which varied along with drought degree.Maize would maintain high metabolic activities of leaves until the drought degree was greater than 0.When the drought degree was below 0.48, maize tended to rapid growth by investing more resources for leaf expansion so that more energy could be absorbed.When the drought degree exceeded 0.48, maize would tend to conserve resource for survival by maintaining relatively lower metabolic activities and smaller size of leaves to minimize water loss.(Zhou Guangsheng)

1.10 Risk assessment of winter injury in the main apple-producing regions of China

Winter injury is one of the main meteorological disasters that affect apple growing distribution and quality.The existing research mainly takes into account winter injury in deep dormancy caused by extremely low temperatures, mostly the occurrence frequencies of the coldest monthly mean temperature, extreme minimum temperatures as meteorological indicators for risk assessment and division of distribution boundaries.But fruit trees cold resistances during different dormant phases have significant differences and are closely related to the process of cold-resistance exercises in fall.The probability forecasting model of early winter injury and extreme winter cold injury are established by the binary Logistic regression, on the basis of data from 2084 meteorological stations and the disaster caused by winter injury from 1981 to 2010.The spaces are divided into five risk degrees as very low, low, moderate, high and very high.Winter injury probability classification standards are referred to the IPCC report of the assessment of possibilities of classification criteria.The results showed that the probability forecasting model of early winter injury and extreme winter cold injury passed the Hosmer-Losmer test, the independent sample prediction accuracy reached 83.6% and 91.4%, respectively.Early winter injury depends mainly on the parameters as first frost and the interval between the beginning date and the date of first frost of the lowest temperature ≤–15 °C.The extreme winter cold injury was affected by the average temperature of the coldest month and the cold accumulated temperature of the lowest temperature≤–20 °C.The freezing injury caused by extreme short-term low temperature is of small occurrence probability and limited influence scope.The extreme low-temperature injury in winter is mainly caused by the continuous intimidation of strong and low temperature, which is compatible with the existing observation facts.Early winter injury is main freezing injury in the main apple-producing regions of China, including most regions in addition to ancient Yellow River and Yunnan.Wherein, most parts of Around Bohai Bay Area and Loess Plateau Area are within the low-risk area.In addition to Dalian, all parts of Liaoning are within the moderaterisk area.Meanwhile, Baoding in Hebei, Changzhi in Shanxi, Yan’an in Shaanxi, and northern area of Jingning in Gansu are within the moderate or severe risk area.Areas with the extreme winter cold injury are mainly distributed in high latitude or high altitude, including the northern Bohai Gulf, the northwest of the Loess Plateau and northern Xinjiang.Within main producing areas of big apple, areas with higher risks from extreme low-temperature freezing injury include Benxi and Fushun in Liaoning, southern part of Zhangjiakou in Hebei,Datong and southern part of Shuozhou in Shanxi as well as Yining in Xinjiang.High risk areas of early winter injury and extreme winter cold injury are almost equal.Gansu province tops the list of above moderate risk,followed by Liaoning, Hebei and Shanxi.On the other hand, there are low risks of winter injury in Shaanxi and Shandong, especially for extreme winter cold injury.(Zhou Guangsheng)

1.11 Impact factors of accumulated temperature stability and improvement of accumulated temperature model of northeast spring maize

Based on the observations on the growth and development of spring maize and meteorological observational data in Northeast China, the spring maize “Sidan19” is firstly taken as an example.The nonlinear accumulated temperature model proposed by Shen Guoquan (or simply as NLM) with good stability is adopted to fi t, and the influence of the selection of parameters on the stability of accumulated temperature is analyzed.The quadratic function of the mean temperature to the linear model (or simply as LM) is revised (which is called TRM after revision) and analyzed and compared with the nonlinear model.In order to analyze the applicability of different northeast spring maize varieties in the application of NLM, three other varieties with more observable years and stations are selected, “Dongnong 248”, “Longdan 13” and “Danyu 13”.Biological significance of parameters and the relationship between parameters and varieties or mature period are analyzed.In addition, the NLM has been improved effectively and verified.

The results show that the stability of accumulated temperature is related to the parameterP, more stable with the smallerP.Accumulated temperature calculated by the nonlinear model of Shen Guoquan shows interannual and inter-regional differences.The main reason for the instability is different temperature strength and its less correlation with other meteorological factors.For each growth period, the fi tted curves between the accumulated temperature and mean temperature are quadratic.The fitting curves of the accumulated temperature calculated by the revised linear model is better than that of Shen Guoquan nonlinear model.Therefore, the revised linear model considering the mean temperature for spring maize in Northeast China is feasible.There are no invalid parameters in the fi tting equations of the four maize varieties.The parameterPis determined as 0.5 based on the smallest variation coefficient of accumulated temperature.There is a significant correlation between parameterKand parameterQ, indicating that the parameterKmay be only a statistical parameter with no clear biological significance.There is a significant difference of accumulated temperature among varieties.The relationship between the parameterQand the mean value of effective accumulated temperature or active accumulated temperature during the whole growth period is found to be statistically significant, indicating thatQis related to the mature period types of different maize varieties because different accumulated temperature means different mature period types.Therefore, a general model applicable to different varieties is proposed whose parameterQandKare represented by effective accumulated temperature or active accumulated temperature.Our findings give good application results and have important implications for revising agro-meteorological indices and improving agriculture service capacity.(Guo Jianping)

1.12 Analysis of drought monitoring ability of different remote sensing drought indices in the whole growth period of Maize

Many indices based on MODIS data have been used to monitor the process of agricultural drought.In this study, one whole growth period of spring maize using MODIS-based drought indices were tested in Songliao Plain, China, in 2014 when a typical serious drought process happened.The main purpose is to examine the applicability of these indices.The correlation coefficients between each index (ATI, VCIndvi, VCIevi, NDWI,VSWI, TVDI) and the soil moisture data were calculated to show the ability of each index in retrieving the drought condition of spring maize in the study area.The results showed that (1) in each part of the growth,each index has its own applicability and characteristics; (2) ATI is more sensitive for the bare soil monitoring,TVDI is more robust and practicable to monitor drought in the semi-arid region (Fig.2).(Fang Shibo)

1.13 Extraction method and accuracy evaluation of rice planting area in multi cloudy area by radar remote sensing

Optical remote sensing is difficult to get remote sensing data in the crops growing season of southern China where there are much more cloudy days, which restricts the application of optical remote sensing to the extraction of crop planting area and distribution.Synthetic aperture radar (SAR) can penetrate clouds and is an effective way to extract landuse information from cloudy area.However, it is necessary to further study how to use the time series of SAR data to get crop distribution and what precision of the classification.Based on time series of SAR images from transplantation to harvest in the whole growth period, the distribution area of early rice was obtained in Nanchang County, Jiangxi in 2016.After field investigation with GPS positioning, the confusion matrix analysis was done.It was concluded that the Kappa coefficient of VV was 0.79 under the VV polarization.The accuracy of the user mapping for early rice distribution is 92.42% (Fig.3).(Fang Shibo)

2 Response and adaptation of agriculture to climate change

2.1 Multi-model ensemble for simulation of the impact of climate change on maize yield

To estimate uncertainties in the study of the impacts of climate change on crop yield, we used 8 climate projections by GCMs under RCP4.5 in the CMIP5 (which represented the uncertainties in the projected climate change) and a statistical and process-based crop model (which represented the uncertainties in the different structures or different formulations of physiological processes of crop models).Historical data of crop and meteorological data during 1981–2009 from agro-meteorological stations of China Meteorological Administration in Hailun, Changling and Benxi in Northeast China were used to establish and evaluate statistical and process-based APSIM models, respectively.Then the two crop models were linked with 8 climate projections to evaluate the impact of climate change on maize yield during 2010–2039 and 2040–2069,using 1976–2005 as the baseline period.In total, 2 crop models under 8 climate projections for a period of 30 years (a total of 480 simulations) were generated for both the baseline and two future climate periods.The results showed that the APSIM model well simulated the growth and yield of maize.The root mean square error (RMSE) for the growth progress (flowering and maturity) simulation was 3–4 days and that for the yield simulation was 0.6–0.8 t hm–2.The established statistical model suggested that temperature during emergence(mid May) had a positive effect on maize yield.However, the increase of temperature and rainfall, and lack of solar radiation during flowering and grain-filling periods (mid July to early September) had a negative impact on maize yield.Compared with 1976–2005, the resulting probability distributions indicated that due to climate change, maize yield in 2010–2039 could decrease on average by 3.8% (Hailun) –7.4% (Benxi), at a probability of 64% (Changling) –73% (Benxi).During 2040–2069, maize yield could decrease by 6.4% (Hailun) –10.5%(Benxi), at a probability of 74% (Hailun) –83% (Benxi).The simulated yield decrease by the APSIM model was 6.6% (Hailun) –8.9% (Benxi) during 2010–2039 and 9.7% (Hailun) –13.7% (Benxi) during 2040–2069.These were higher relative to those simulated by the statistical model, which were 0.9% (Hailun) –6.0% (Benxi)during 2010–2039 and then 2.0% (Changling) –7.3% (Benxi) during 2040–2069.(Zhang Yi)

2.2 The possible influence of future climate change on the growth period and yield of wheat in China

The possible effects of future climate change on the growth period and yield of wheat in China were discussed by using the Chinese agrometeorological (wheat) model (CAMM).The results showed that the influence of future climate change on the growth period of Chinese wheat is not obvious, but the variety improvement is the more important factor.The reason is that with the adjustment of the sowing date and changes of turn green date, the temperature of the whole growth period of wheat did not change obviously.The prolongation of the growth period does not necessarily lead to the increase of yield of wheat, which may be due to the weak growth in the reproductive period caused by the flourishing growth in the vegetative period.However, due to the lack of reliable climate prediction and the discrepancy of the model itself, the results of the assessment still have greater uncertainty (Fig.4).(Ma Yuping)

2.3 Response of spring maize to climate change in Inner Mongolia based on APSIM model

Based on the spring maize field trial data from Tumotezuoqi Agro-Meteorological Experiment Station in Inner Mongolia and daily meteorological data, the adaptability of agricultural production system model APSIM in maize producing areas of Inner Mongolia was analyzed.The key environmental factors impacting the development stage of spring maize were determined.Response of spring maize development stage to environmental factors was explored.The results implied three main conclusions.First, the validated APSIM model had better adaptability in Tumetezuoqi, Inner Mongolia.Second, the response of different developmental stages to the environmental impact factor was different.The first factor affecting the DOY (day of the year) of flowering and maturity at the same time was temperature, followed by relative humidity, potential evapotranspiration, and wind speed.Moreover, the DOY of maturity had significant positive correlation with total precipitation in spring maize growth period.The DOY of flowering was the most sensitive to the average minimum temperature variations.However, the DOY of maturity was the most sensitive to the average maximum temperature variations, followed by the average temperature, soil surface average temperature,and average minimum temperature.Finally, a statistical model reflecting the relationship between critical development stage of spring maize and environmental factors was established.These results can provide technical support for analyzing the limiting factors of spring maize growth and yield formation in the future maize-producing areas in Inner Mongolia.(Zhao Junfang)

2.4 Impacts of adaptive measures to climate change on maize climatic potential productivity in Northeast China

In order to adapt to future climate change, this study adopted two measures of stress resistant varieties and delayed sowing date, combined with the daily meteorological data of RCP4.5 scenario and RCP8.5 scenario from 2010 to 2099 simulated by a regional climate model, to analyze the maize climatic potential productivity changes in Northeast China under different climate change scenarios.The results showed that in 2010–2099, the spatial characteristics of maize climatic potential productivity in Northeast China decreased from southeast to northwest.The maize climatic potential productivity under RCP4.5 scenario was higher than that under RCP8.5 scenario, while years with the lowest values in RCP8.5 scenario was more than that in RCP4.5 scenario.Maize climatic potential productivity for stress resistant varieties was higher compared with the original varieties.Under RCP4.5 scenario, the variety with heat resistant had higher productivity, but under RCP8.5 scenario, the variety with drought resistant performed better.The variety with both heat and drought resistant achieved the highest productivity under both RCP4.5 scenario and RCP8.5 scenario.Under RCP4.5 scenario, yield increased with postponed sowing and 30–40 days delay achieved the highest yield.Under RCP8.5 scenario, the yield reduction occurred in some areas.The results thus imply that the appropriate delay in sowing is conducive to improving the maize productivity but varying with regions.(Guo Jianping)

2.5 Trend analysis of temperature conditions over different growth periods of winter wheat in North China Plain

The changes of monthly averaged temperature and the changing trends of the dates of critical developmental stages and averaged temperatures during the growing periods of winter wheat in North China Plain were analyzed based on the meteorological and phonological data collected from agrometeorological observational stations from 1981 to 2010 and warming experiments in the region.The results showed that there was a significant warming trend in October, December and between February and June during the growing season in the plain.Furthermore, the average temperature in February exhibited the maximum linear trend of temperature increase.Climate warming significantly increased the average temperatures over overwintering and the period from reviving to jointing stage of winter wheat.At the same time, the warming caused the jointing to maturity dates of winter wheat significantly advanced.However, the average temperatures over two periods did not show the same increasing trend as the corresponding monthly averaged temperatures.The temperature condition over the period before winter was stable because of the delayed sowing, an insignificant change in the average temperature during the period from jointing to maturity was attributed to the advance of developmental stages and local climatic characteristics of seasonal variation.(Tan Kaiyan)

2.6 Assessment on adaptation and risk of agriculture to climate change in Gansu Province

The temporal and spatial evolution of climate change, agricultural climate resources, agricultural meteorological disaster, agricultural cultivation system, crop yield loss from agricultural meteorological disasters and agrometeorological disaster risk in Gansu Province were systematically evaluated, and the countermeasures of agricultural adaptation to climate change were put forward, in order to promote the healthy and sustainable development of agriculture and to alleviate accurately the poverty.The main results were given as follows.

(1) Climate change.Since 1961, the average temperature, average maximum temperature, and average minimum temperature in Gansu Province had been rising.The highest rising was the average maximum temperature, and the lowest was the average temperature.The temperature increased obviously in all seasons.Among the seasons, the temperature rising was the highest in winter and the lowest in summer.Both average annual precipitation and average annual sunshine hours showed no significant trend.

(2) Trends in extreme climate events.Since 1961, the extreme maximum temperature showed increasing trend in Gansu Province except in the western part of Qilian.The day number of the daily maximum temperature over 35 °C showed a rising trend except in the Qilian mountains and the northern part of Gannan plateau, with the increasing rate being about 0.3 d/10a.The annual extreme minimum temperature showed an increasing trend except in the northwestern part of Jiuquan city, and the most obvious warming happened in Gannan plateau.The longest continuous days without rainfall also showed an increasing trend in the centraleastern part of Hexi, the central and northern parts and the southwestern part of Longzhong, the eastern part of Longzhong and the central part of southern Gansu.

(3) Trend of agricultural climate resources.Since 1961, the beginning dates with average daily temperatures of more than 0 °C and 10 °C showed an advance trend to some degree, and the ending dates showed different degrees of delay trend.The precipitation with average daily temperature of more than 0 °C and 10 °C showed a decreasing trend from the southeast to the west.The precipitation showed an increasing trend in most part of Hexi and a decreasing trend in most part of Hedong.The accumulated temperature showed an obvious increasing trend.The sunshine hours with average daily temperature of more than 0 °C showed an increasing trend in most part of Gansu Province, and the sunshine hours with average daily temperature of more than 10 °C did not show obvious change.

(4) Effects of changes in agricultural climate resources on crop yield.During 1980–2014, the increasing average temperature of winter wheat region resulted in the decrease of yield by 7.5%; the increasing daily range of temperature resulted in the decrease of yield by 6.7%; the decreasing average precipitation resulted in the decrease of yield by 0.4%.In the region of spring wheat, the increasing average temperature resulted in the decrease of yield by 4.3%; the decreasing daily range of temperature resulted in the increase of yield by 0.7%; the increasing average precipitation resulted in the increase of yield by 0.1%.In the region of maize,the increasing average temperature resulted in the decrease of yield by 1.2%; the increasing daily range of temperature resulted in the decrease of yield by 1.2%; the decreasing average precipitation resulted in the decrease of yield by 0.2%.During 1985–2014, the increasing average temperature in the potato region resulted in the decrease of yield by 1.8%; the decreasing daily range of temperature resulted in the increase of yield by 0.3%; the decreasing average precipitation resulted in the decrease of yield by 0.3%.

(5) Trend of agricultural meteorological disasters.Since 1961, the occurrence frequency and intensity of meteorological drought in Gansu Province showed an obvious increasing trend.The area of spring and summer drought showed an obvious increasing trend, while the area of drought occurred in late spring and early summer and autumn showed an obvious decreasing trend.The number of days (stations) with gale showed a decreasing trend but showed an increasing trend after 2007.The number of days with heavy rain showed no significant trend, mainly in the region of Hedong.The number of frost days increased first and then decreased,particularly significant after 1980s.The day number of the first frost became the minimum value since 2005.

(6) Temporal and spatial evolution of agricultural meteorological disasters.Since 1961, the area and harm of the agricultural drought disaster increased significantly.The rates of drought disaster, drought hazard and total crop failure (25.2%, 14.1% and 2.2%) were significantly higher than the national average (15.0%, 8.1%and 1.7%), their increase rates (0.16%/10a,, 0.15%/10a, and 0.05%/10a,) were also higher than the national average.The comprehensive loss rates of hail disaster, flood disaster and chilling injury also increased with the increase rate of 0.29%/10a,, 0.45%/10a, and 0.72%/10a,.

(7) Trends and impacts of agricultural pests and diseases.During 1981–2015, climate change was conducive to the expansion of agricultural pests and diseases in Gansu Province, increasing the negative effects.The occurrence areas of agricultural diseases, pests, weeds and rat damage increased at rates of 0.20/10a,0.08/10a, 0.06/10a, –0.03/10a, respectively.The occurrence area rates of agricultural diseases, pests and rat damages were mainly affected by temperature, and the occurrence area rate of weed was mainly affected by the number of days of precipitation.The loss rate of both yield per unit and total production from agricultural pests and diseases ranged as potato ＞ maize ＞ wheat.The diseases harmed more loss for potato and wheat than the pests, while the pests caused more loss for maize than the diseases.Therefore, more attention should be paid to the pests and diseases of potato and maize, especially the diseases of potato and the pests of maize in the future, and the diseases of wheat should also draw more attention in the future.These effects from pests and diseases should be taken as the focus of prevention and control.

(8) Trend and effects of the agricultural planting system.Compared with 1951–1980, the planting boundary of two crops one year moved northward in different degrees.The regions with obvious changes included Longnan, Longdong and Gannan plateau.The northern boundary of winter wheat expanded westward in different degrees, and the regions with obvious changes included Hexi and Gannan plateau.The change from one crop one year cultivation pattern of winter wheat, spring wheat, maize, potato to two crops one year of winter wheat-summer maize could make the yield increase.The increasing yield rate reached about 153.53%,65.13%, 149.69% for winter wheat, spring wheat and potato in Longnan and 84.56%, 91.27%, 76.42%,83.02% for winter wheat, spring wheat, maize, potato in Longzhong.

(9) Countermeasures of agricultural adaptation to climate change.In the light of the new characteristics of agricultural climate resources and new situation of agrometeorological disasters in Gansu Province under the background of climate change, this study presented a series of new patterns of climate resource efficient utilization, in order to effectively alleviate the adverse effects of climate change on agricultural production,to promote the level of agricultural production and alleviate effectively poverty in Gansu Province as soon as possible.The countermeasures mainly included: the optimization of land use pattern, making full use of solar radiation and heat resources; adjustment of crop cultivation system, taking the initiative to adapt to climate change; breeding crop cultivars with high yield, good quality and strong resistance, coping scientifically with climate warming and pests and diseases; adjusting the multiple cropping index, improving the utilization efficiency of cultivated land resources; adjustment of crop planting area and species distribution, making full use of water and heat resources; emphasizing regional differentiation of climate change, scientific adjustment in production management.(Zhou Guangsheng)

3 Ecometeorological monitoring

3.1 Developing drought index and threshold of typical steppe

In response to the global climate change, marked by elevated CO2concentration and temperature, the frequency of extreme weather events, such as drought has been increasing, and the influences on plants have appeared and will continue into the foreseeable future.The research on the process of plant response to drought is helpful in monitoring the plant drought occurrence and development and evaluating the plant drought conditions.Based on the field manipulation experiments of typical grassland plants, this study quantitatively analyzed the sensitivity indicators in the process of plant response to drought, critical threshold, plasticity and correlation, and photosynthetic eco-physiology mechanism, and investigated the change in drought indicators and their thresholds under elevated CO2conditions.The results are as follows:

(1) Photosynthetic rate (Pn) and leaf water content ofLeymus chinensisandStipa kryloviiwere firstly affected by drought stress, and the relative humidity threshold was between 51% –54%.The single leaf area ofL.chinensiswas earlier subjected to drought stress than the leaf number.S.kryloviijust showed the opposite situation.

(2) Drought plasticity of total leaf area.ForL.chinensisandS.kryloviiwas greatest in the morphological characteristics.The plasticity of plant height ofL.chinensiswas greater than that of leaf number,S.kryloviishowed an opposite result.ForL.chinensis, the correlation of total leaf area, plant height, leaf number,and biomass with drought had little difference; but leaf number had the maximum correlation with drought among those 4 indicators inS.krylovii.In the physiological characteristics of leaves, the correlation between transpiration rate (Tr) ofL.chinensisand drought was the highest; the correlation between vapor pressure deficit (VPD) ofS.kryloviiand drought was the highest.

(3) Under drought stress conditions, the vapor pressure deficit (VPD) increased, the leaf stomata closed and resistance increased, the stomatal conductance (Gs) decreased, then the transpiration (Tr) decreased,the water loss rate was reduced, the net photosynthetic rate (Pn) decreased mainly by stomatal limitation in this stage.However, as the drought continued, the chlorophyll content decreased, the intercellular CO2concentration (Ci) andCi/Ca(atmospheric CO2concentration) increased, the increase of non-stomatal limitation was the leading factor.

(4) After re-watering,Pnand leaf water content ofL.chinensisandS.krylovii, and chlorophyll content ofL.chinensiscould quickly recover to the level of sufficient water supply, the effect of drought was reversible.The total leaf area and single leaf area ofL.chinensisandS.kryloviiincreased gradually after re-watering,still significantly lower than at the sufficient water supply.The change degrees ofPn,Gs, VPD andTrofL.chinensiswere larger than those ofS.kryloviiandPn, and the change degrees of leaf water content ofL.chinensisaffected by drought were earlier than inS.krylovii.The drought threshold (soil relative moisture) ofL.chinensiswas higher than that ofS.krylovii.After re-watering, the recovery degree ofPnand above ground biomass ofL.chinensiswere greater than that ofS.krylovii, all these results indicated that the sensitivity ofL.chinensisto drought and re-watering was greater than that ofS.krylovii.

(5) Total leaf area or leaf number (morphology), leaf water potential or leaf water content (physiology),and above ground biomass better reflected the water status ofS.bungeanaunder ambient and elevated CO2than the 13 other analyzed variables.The sensitivity of drought indicators changed under the elevated CO2condition.By quantifying the relationship between precipitation and the 5 most sensitivity indicators, we found that the thresholds of precipitation decreased under elevated CO2concentration.(Zhou Guangsheng)

3.2 Revealing the spatiotemporal dynamics of carbon sequestration and its driving mechanism in China’s grasslands

China’s grasslands constitute approximately more than 40% of the national land area, they’re climatesensitive and ecologically fragile area, and severely affected by human activities.Accurately assessing the spatiotemporal dynamics of carbon sequestration and the corresponding driving mechanism in China’s grasslands is one of the hot topics in the study on global change and carbon budget.In this study, we used the field observational data from 2011 to 2013 and relevant literatures with the process-based ecosystem model(Terrestrial Ecosystem Model, TEM 5.0) to examine the spatiotemporal dynamics of carbon sequestration of China’s grasslands during 1961 to 2013.We first parameterized TEM 5.0 with observational data for China’s grasslands.Second, we validated the model and extrapolated parameterization to the whole study region.We then analyzed the temporal and spatial variations of carbon sequestration and its driving mechanism.At last, we estimated the carbon sequestration potential under the current and future climate scenarios.The main conclusions are listed as follows:

(1) During 1961 to 2013, the China’s grasslands acted as a carbon sink, with annual mean value of 19.06 Tg C a–1.The Inner Mongila grasslands (5.77 Tg a–1), Xinjiang grasslands (5.44 Tg a–1) and Tibetan Plateau grasslands (5.16 Tg a–1) contributed most (85.9%) to the sink.Significant interannual variability trend of the net ecosystem productivity (NEP) was mainly attributed to different sensitivities of net primary productivity (NPP)and heterotrophic respiration (RH) to annual temperature variability.The gross NEP of each decadal acted as a carbon sink.The early stage of 21st century contributed most to the sink at 66.33 Tg C a–1, which accounted for 57.6% of the gross NEP in five decades.The vegetation carbon and soil organic carbon of the study region showed an increasing tendency during the study period.

(2) Annual NEP of China’s grasslands was 4.84 g C m–2a–1, and acted as a carbon sink in the whole region.The major carbon sink occurred in western Tibetan Plateau, northeastern Inner Mongolia, and northern Xinjiang.The southeastern, northeastern, and western boundaries of Tibetan Plateau acted as carbon sources.The carbon source area accounted for only 13.5% of the total study area.During the study period, most of the study region showed an increasing tendency of NEP, with 82.5% of China’s grassland, and distributed mainly in the grasslands of northwestern Tibetan Plateau, northern Xinjiang, and northeastern Inner Mongolia.The decreasing tendency of NEP occurred in northwestern Inner Mongolia, southwestern and northeastern Tibetan Plateau, and southern grasslands, which accounted for 17.5% of the total study region.

(3) The mean annual temperature showed a significantly increasing trend during 1961 to 2013.Annual precipitation varied among years and decreased from southeast to northwest in the spatial pattern, with the decadal trend of dry–wet–dry.The annual net ecosystem productivity showed a positive correlation with temperature and precipitation, but it showed no significant correlation with solar radiation.Annual variations of temperature and precipitation were main driving forces of carbon sequestration in China’s grasslands.

(4) The mean annual carbon sequestration potential was 18.41 Pg C a–1during 1961 to 2013.Among them,there were 18.12 Pg C a–1in soil and 0.29 Pg C a–1in vegetation.The spatial distribution patterns of carbon sequestration potential per unit area in different ages were roughly the same, the grasslands in the Qinghai-Tibetan Plateau and the Junggar basin, Xinjiang had the larger carbon sequestration potential, with mean annual carbon sequestration values generally above than 10000 g C m–2a–1; the southern China grasslands had mean annual carbon sequestration between 1000～5000 g C m–2a–1.

(5) Under RCP4.5 and RCP8.5 climate scenarios, without the influence of human disturbance, the results indicated that the total carbon storage showed an increasing trend, with the average annual growth rate of 61.0 Tg C m–2a–1(RCP4.5) and 234.2 Tg C m–2a–1(RCP8.5).The spatial pattern of carbon density under the two climate scenarios was consistent, showing an increasing trend in carbon density during 2011 to 2040, with significant increase in the Qinghai-Tibetan Plateau and Inner Mongolia grasslands.

(6) Under RCP4.5 climate scenario, the mean annual carbon sequestration potential was 18.01 Pg C a–1during 2011 to 2040, with 17.78 Pg C a–1in soil and 0.24 Pg C a–1in vegetation.Under RCP8.5 climate scenario, the mean annual carbon sequestration potential was 18.00 Pg C a–1from 2011 to 2040, with 17.79 Pg C a–1in soil and 0.20 Pg C a–1in vegetation.The spatial distribution of carbon sequestration potential in China’s grasslands was similar in the two climate scenarios.In 2030s, the carbon sequestration potential in Junggar basin, Xinjiang was the largest, followed by the central (Tianshan Mountain) and southern (Kunlun Mountain)part in Xinjiang, the part of Inner Mongolia grasslands, as well as the southern, central and southeastern parts of Qinghai-Tibetan Plateau grasslands.The carbon sequestration potential in the remaining grasslands was small, generally less than 5000 g C m–2a–1, close to zero in some places.(Zhou Guangsheng)

4 Application of results

4.1 Applicability of PyWOFOST based on EnKF in simulating maize yields

A modified and improved crop model (PyWOFOST), which coupled remote sensing information and a crop model (WOFOST) with Ensemble Kalman Filter (EnKF), was evaluated to simulate maize yields in Northeast China.The work has been adopted by the Agricultural Information Institute of CAAS for assessing regional yield risk services at the end of 2017.(ZhaoYanxia)