陸汝華 王文軒 曹 強(qiáng) 田永超 朱 艷 曹衛(wèi)星 劉小軍,*

稻麥復(fù)種模式下氮肥與稻秸互作對(duì)小麥產(chǎn)量和N2O排放影響及推薦施肥研究

陸汝華1王文軒2曹 強(qiáng)1田永超1朱 艷1曹衛(wèi)星1劉小軍1,*

1南京農(nóng)業(yè)大學(xué)國(guó)家信息農(nóng)業(yè)工程技術(shù)中心 / 智慧農(nóng)業(yè)教育部工程研究中心 / 農(nóng)業(yè)農(nóng)村部農(nóng)作物系統(tǒng)分析與決策重點(diǎn)實(shí)驗(yàn)室 / 江蘇省信息農(nóng)業(yè)重點(diǎn)實(shí)驗(yàn)室, 江蘇南京 210095;2人文與社會(huì)發(fā)展學(xué)院, 江蘇南京 210095

優(yōu)化氮肥施用和秸稈還田技術(shù)為途徑的農(nóng)業(yè)管理措施被認(rèn)為是提升農(nóng)業(yè)可持續(xù)性的有效手段, 然而當(dāng)前關(guān)于氮肥和秸稈還田對(duì)小麥產(chǎn)量和N2O排放影響的研究仍十分有限。為此, 本研究基于2000—2022年發(fā)表的關(guān)于長(zhǎng)江中下游流域氮肥和秸稈投入下小麥產(chǎn)量和N2O排放變化的文獻(xiàn), 運(yùn)用隨機(jī)森林建模, 定量分析氮肥和秸稈還田對(duì)小麥產(chǎn)量和N2O排放的影響, 并結(jié)合情景設(shè)置進(jìn)行了特定地點(diǎn)的小麥產(chǎn)量和N2O排放模擬, 同時(shí)評(píng)估了碳排放強(qiáng)度(CEE)和凈生態(tài)系統(tǒng)經(jīng)濟(jì)效益(NEEB)。結(jié)果表明, 建立的區(qū)域尺度小麥產(chǎn)量與N2O排放對(duì)氮秸互作響應(yīng)的隨機(jī)森林模型, 驗(yàn)證結(jié)果2分別為0.66和0.65, RMSE分別為0.70和1.11。結(jié)果表明施氮量和土壤有機(jī)質(zhì)是影響小麥產(chǎn)量和N2O排放的重要因素。綜合來(lái)看, 達(dá)到最大產(chǎn)量所需的氮肥量為208～212 kg hm–2, 達(dá)到最小CEE所需的氮肥量為113～130 kg hm–2, 達(dá)到最高的NEEB所需的氮肥量為202～205 kg hm–2, 其中在6.75 t hm–2的秸稈投入下施用202 kg hm–2的氮肥可以獲得最高的生態(tài)收益1.37萬(wàn)元。優(yōu)化氮肥和秸稈投入具備減少作物碳排放強(qiáng)度并獲得最大凈生態(tài)環(huán)境效益的潛力。

施氮量; 秸稈投入; 小麥; N2O; 排放模型; 推薦施肥

小麥對(duì)糧食安全做出了巨大貢獻(xiàn)[1], 提供了人類飲食中約20%的熱量和蛋白質(zhì)[2]。農(nóng)業(yè)作為一氧化二氮(N2O)的最大排放源, 占全球人為N2O排放量的56%～81%, 對(duì)全球氣候變化具有重大影響[3]。農(nóng)田管理實(shí)踐(如秸稈投入和施肥)會(huì)影響土壤中的生化和水文條件, 進(jìn)而調(diào)節(jié)作物生長(zhǎng)和N2O排放[4]。長(zhǎng)江中下游地區(qū)是我國(guó)重要的小麥主產(chǎn)區(qū), 研究不同的氮肥和秸稈投入如何影響小麥產(chǎn)量和N2O排放對(duì)于制定區(qū)域農(nóng)業(yè)可持續(xù)施肥管理至關(guān)重要。

作物產(chǎn)量的增加依賴于基因型、環(huán)境因素(包括氣候和土壤條件)和農(nóng)業(yè)管理之間復(fù)雜的相互作用。近年來(lái), 作物生產(chǎn)依賴于越來(lái)越多的氮肥投入來(lái)追求更高的產(chǎn)量, 但作物生長(zhǎng)速度和產(chǎn)量的增加并沒(méi)有與氮肥施用量的增加相匹配[5]。與此同時(shí), 施用尿素等含氮化肥在維持作物產(chǎn)量的同時(shí), 也為N2O的合成提供了基質(zhì)。已有研究表明, 氮素含量與直接N2O排放之間呈線性函數(shù)關(guān)系或者非線性關(guān)系[6-8], 這都表明不斷增加氮肥施用量會(huì)導(dǎo)致N2O排放量迅速增加[9]。此外, 秸稈還田對(duì)土壤N2O排放的影響也十分復(fù)雜, 秸稈還田顯著影響生物地球化學(xué)C和N循環(huán), 從而影響N-痕量氣體的產(chǎn)生和排放, 作物秸稈滯留可能通過(guò)影響相對(duì)C和N的可用性以及在土壤通氣狀態(tài)調(diào)節(jié)土壤N2O通量方面發(fā)揮多種作用[10-11]。一般來(lái)說(shuō), 秸稈還田會(huì)刺激N2O的產(chǎn)生, 因?yàn)樽魑锝斩挿纸鉃橄趸?反硝化菌提供了底物, 促進(jìn)了反硝化進(jìn)程[12]。此外, 秸稈還田對(duì)N2O排放有負(fù)面影響或沒(méi)有顯著影響[13]。最近的研究表明, 氮肥與秸稈配合使用是提高氮素利用效率和土壤肥力的有效策略[14], 因?yàn)榻M合施用可以克服施用單一肥料的缺點(diǎn), 可以提高作物產(chǎn)量, 提高土壤肥力并緩解溫室氣體排放。Huang等[15]的研究發(fā)現(xiàn), 麥玉輪作中優(yōu)化的氮肥施用量與秸稈還田相結(jié)合, 大大增加了產(chǎn)量和降低了產(chǎn)量尺度的N2O排放。Akhtar等[16]的研究也發(fā)現(xiàn)9000 kg hm–2的玉米秸稈還田配合減氮20%被認(rèn)為是提高作物產(chǎn)量和減少土壤N2O排放的可行性技術(shù)。在長(zhǎng)江中下游地區(qū), 氮肥和秸稈合理配施, 是否可以平衡土壤碳氮, 改善農(nóng)田生態(tài)效應(yīng), 并顯著提高產(chǎn)量, 是本文的主要研究目標(biāo)。

小麥產(chǎn)量形成和N2O排放通常受到區(qū)域尺度下不同土壤、氣象和管理措施的綜合影響, 而秸稈和氮肥投入廣泛應(yīng)用的前提, 是揭示其在不同地區(qū)的小麥增產(chǎn)與N2O減排效應(yīng)的驅(qū)動(dòng)因素, 才能更好地權(quán)衡小麥的養(yǎng)分需求和環(huán)境影響并進(jìn)行指導(dǎo)施肥管理, 從而實(shí)現(xiàn)N2O排放減少和作物產(chǎn)量穩(wěn)定。目前探索產(chǎn)量措施響應(yīng)的田間試驗(yàn)通常只設(shè)置固定的氮肥和秸稈投入梯度和選用特定地點(diǎn)的小麥品種, 因此結(jié)果很難外推到所有的生長(zhǎng)環(huán)境中?；诖笮蛿?shù)據(jù)集的統(tǒng)計(jì)方法可以更全面地了解產(chǎn)量-投入關(guān)系, 機(jī)器學(xué)習(xí)(Machine learning, ML)模型提供了一種強(qiáng)大的方法來(lái)研究作物管理和生長(zhǎng)環(huán)境之間的復(fù)雜相互作用, 因?yàn)樗鼈兡軌蚧诩蓪W(xué)習(xí)方法處理預(yù)測(cè)變量與目標(biāo)變量之間的復(fù)雜關(guān)系, 被廣泛應(yīng)用在N2O排放量預(yù)測(cè)[17]、產(chǎn)量預(yù)測(cè)[18]和施肥推薦[19]。然而, 目前鮮少關(guān)于應(yīng)用ML模型預(yù)測(cè)氮秸互作下的麥田N2O排放和產(chǎn)量并用于推薦施肥措施的報(bào)道。鑒于此, 本研究首先基于隨機(jī)森林(Random forest, RF)算法對(duì)長(zhǎng)江中下游區(qū)域不同氮肥和秸稈投入對(duì)小麥產(chǎn)量和N2O排放變化響應(yīng)進(jìn)行建模、驗(yàn)證和評(píng)估。然后利用驗(yàn)證的模型和不同氮肥和秸稈投入情景設(shè)置對(duì)小麥產(chǎn)量和N2O累積排放量進(jìn)行估計(jì), 并評(píng)估不同氮肥和秸稈投入下的碳排放強(qiáng)度(Carbon emission efficiency, CEE)和凈生態(tài)系統(tǒng)經(jīng)濟(jì)效益(Net ecosystem economic benefits, NEEB), 為根據(jù)不同目標(biāo)的環(huán)境友好型的氮肥和秸稈投入措施提供技術(shù)支持。

1 材料與方法

1.1 數(shù)據(jù)來(lái)源

本研究檢索了2000—2022年在Web of Science和中國(guó)知網(wǎng)關(guān)于稻麥輪作區(qū)下麥田N2O排放、氮肥處理、秸稈投入、產(chǎn)量的中英文論文。以下搜索詞作為關(guān)鍵詞: “秸稈還田” “氮肥” “N2O” “溫室氣體” “產(chǎn)量”, “Straw return” “nitrogen fertilizer” “greenhouse gas” “GHG” “GWP”。不同的關(guān)鍵詞之間使用“AND”表明并列關(guān)系。符合標(biāo)準(zhǔn)的研究被納入并進(jìn)行分析: (1) 所有摘錄文獻(xiàn)必須是在野外開展的田間試驗(yàn), 且用的是靜態(tài)箱-氣相色譜法; (2) 還田的秸稈必須是水稻秸稈且麥田研究必須在長(zhǎng)江中下游地區(qū); (3) 文獻(xiàn)的結(jié)果都可以直接從表或圖或文本中獲取。

搜集相關(guān)文獻(xiàn)后, 需要對(duì)錄入的數(shù)據(jù)進(jìn)行異常值剔除的預(yù)處理, 數(shù)據(jù)庫(kù)從超過(guò)323篇下載的文獻(xiàn)中, 篩選了30篇符合標(biāo)準(zhǔn)的文獻(xiàn)[20-49], 涉及產(chǎn)量觀測(cè)數(shù)據(jù)為80條, 氧化亞氮觀測(cè)數(shù)據(jù)為184條。文獻(xiàn)數(shù)據(jù)地點(diǎn)主要分布在江蘇省、安徽省和湖北省。其中, 江蘇省的數(shù)據(jù)地點(diǎn)主要包括: 南京市(32.0°N, 118.8°E)、(32.48°N, 118.6°E)、(31.93°N, 118.98°E), 鎮(zhèn)江市(31.97°N，119.3°E), 揚(yáng)州市(32.58°N, 119.2°E)、(32.58°N, 119.7°E), (蘇州市(31.4°N, 120.42°E)、(31.53°N, 120.7°E)、 (31.55°N, 120.62°E)、(31.53°N, 120.92°E)、(31.53°N, 120.68°E), 南通市(31.68°N, 121.9°E), 無(wú)錫市(31.45°N, 120.42°E)。安徽省的數(shù)據(jù)點(diǎn)主要包括: 滁州市(32.0°N, 118.13°E), 蕪湖市(31.15°N, 118.13°E), 巢湖市(31.65°N, 117.68°E)、(31.6S°N, 117.67°E)。湖北省的數(shù)據(jù)點(diǎn)主要包括: 黃岡市(30.02°N, 115.57°E), 荊門市(30.88°N, 112.8°E), 襄陽(yáng)市(32.02°N, 112.07°E)。

1.2 數(shù)據(jù)分析

1.2.1 CEE和NEEB的計(jì)算 CEE是單位產(chǎn)量下的碳排放強(qiáng)度變化, 又稱碳排放強(qiáng)度, 由于碳排放強(qiáng)度變化的計(jì)量單元主要是旱地作物化肥施用引起的N2O直接排放, CEE計(jì)算如下:

CEE = N2O/Yield(1)

N2O表示麥田N2O累積排放量并轉(zhuǎn)化為二氧化碳當(dāng)量(kg CO2-eq hm–2), Yield表示小麥產(chǎn)量(kg hm–2)。

凈生態(tài)系統(tǒng)經(jīng)濟(jì)預(yù)算(NEEB)是根據(jù)糧食產(chǎn)量成本、農(nóng)業(yè)投入成本和碳成本計(jì)算出來(lái)的[50], 如下所示:

NEEB = 糧食產(chǎn)量成本–農(nóng)業(yè)活動(dòng)成本–碳成本(2)

在這項(xiàng)研究中, 糧食產(chǎn)量成本是根據(jù)當(dāng)前糧食價(jià)格(小麥, 2286 CHY t–1)和糧食產(chǎn)量計(jì)算得出的[51]。農(nóng)業(yè)活動(dòng)成本包括氮肥成本(3.6 CHY kg–1N)[52], 碳成本是當(dāng)前碳貿(mào)易價(jià)格(56.38 CHY t–1CO2-eq)和GWP的乘積[53]。在本研究中, 不考慮機(jī)械耕作、小麥種子、其他肥料、灌溉、化學(xué)農(nóng)藥(除草劑+殺蟲劑+殺菌劑)、秸稈處理和機(jī)械收割等的農(nóng)業(yè)活動(dòng)成本。

1.2.2 皮爾遜相關(guān)分析 利用IBM SPSS Statistic 26進(jìn)行皮爾遜相關(guān)分析, 本研究的重點(diǎn)是使用皮爾遜相關(guān)系數(shù)(公式3)分析2個(gè)變量之間的關(guān)系, 其值介于–1和+1之間。

式中,是總樣本大小,x和y是用索引I的單個(gè)樣本點(diǎn),、分別表示是樣本的平均值。

1.2.3 RF模型的構(gòu)建 RF是經(jīng)典的監(jiān)督學(xué)習(xí)算法, 本研究中的RF算法使用Python軟件(版本3.8)實(shí)現(xiàn), 并通過(guò)計(jì)算每個(gè)特征的平均不純度減少量對(duì)特征重要性進(jìn)行排序[54]。

1.2.4 模型的檢驗(yàn)與評(píng)價(jià) 本研究采用標(biāo)準(zhǔn)差(Standard deviation, SD)和變異系數(shù)(Coefficient of variation, CV)被用來(lái)表征總試驗(yàn)數(shù)據(jù)的分離分散程度。CV越大, 則所有數(shù)據(jù)包含的可能性就越多。SD和CV的計(jì)算如下:

本研究通過(guò)比較實(shí)際值和模擬值之間1∶1線的決定系數(shù)(2)、均方根誤差(RMSE), 并結(jié)合1∶1圖來(lái)描述模型對(duì)小麥生產(chǎn)力和溫室氣體排放的預(yù)測(cè)效果。具體計(jì)算如下:

1.3 數(shù)據(jù)處理

利用Microsoft Excel 2016軟件建立數(shù)據(jù)庫(kù), Origin pro 2022軟件繪制圖表和數(shù)值化。

2 結(jié)果與分析

2.1 區(qū)域尺度小麥產(chǎn)量和N2O排放效應(yīng)模型的構(gòu)建

2.1.1 N2O排放與模型輸入?yún)?shù)之間的關(guān)系 從表1發(fā)現(xiàn), 長(zhǎng)江中下游搜集的數(shù)據(jù)中pH、土壤有機(jī)質(zhì)(Soil organic matter, SOM)、年降雨量(LT_ Prec)、土壤總氮(Total nitrogen, TN)和N2O排放量(Cum N2O)的CV值均較大, 其中N2O排放量的CV值達(dá)到81.82%。這些結(jié)果表明數(shù)據(jù)可以代表實(shí)際生產(chǎn)場(chǎng)景中的大多數(shù)可能情況, 因此數(shù)據(jù)可用于分析更通用的小麥N2O排放情況。通過(guò)皮爾遜相關(guān)分析, 從相關(guān)系數(shù)可以看出, 在區(qū)域尺度上, N2O排放量與施氮量(N rate)、施肥次數(shù)(Split N)和年平均溫度(LT_Temp)呈顯著正相關(guān), 相關(guān)系數(shù)在0.16～0.23之間, 隨著施氮量的增加, 麥田引起的N2O排放越多。其中pH和土壤有機(jī)質(zhì)(SOM)與N2O排放量呈現(xiàn)負(fù)相關(guān), 相關(guān)系數(shù)分別為–0.20和–0.29。然而年降雨量(LT_Prec)與N2O排放量之間的相關(guān)性并不明顯, 這可能與年降雨量與當(dāng)季降雨量之間存在誤差, 增加了對(duì)N2O排放的響應(yīng)的不確定性。在長(zhǎng)江中下游地區(qū)中, 秸稈還田量對(duì)的小麥N2O排放量會(huì)有一定的抑制作用, 但并不顯著(圖1)。

2.1.2 產(chǎn)量和N2O排放效應(yīng)模型構(gòu)建與驗(yàn)證

基于收集的文獻(xiàn)數(shù)據(jù)和RF方法, 將數(shù)據(jù)集根據(jù)無(wú)秸稈投入(=0)、小于4.50 t hm–2的秸稈投入(0<<4.5)和大于4.50 t hm–2的秸稈投入(>4.50)進(jìn)行劃分, 構(gòu)建了小麥產(chǎn)量和N2O排放效應(yīng)模型。結(jié)果表明, RF模型可以預(yù)測(cè)不同區(qū)域氮肥和秸稈投入下的小麥產(chǎn)量和累積N2O排放量。小麥產(chǎn)量模型訓(xùn)練和測(cè)試集2介于0.66～0.96, RMSE介于0.35～0.70 t hm–2, 驗(yàn)證結(jié)果達(dá)到顯著水平(< 0.001)。產(chǎn)量在各秸稈投入下差異不大。對(duì)于溫室氣體而言, N2O模型訓(xùn)練和測(cè)試集2介于0.65～ 0.85, RMSE介于0.81～1.11 kg hm–2, 驗(yàn)證結(jié)果達(dá)到顯著水平(<0.001), N2O累積排放隨著秸稈投入的增加而增加(圖2)。

2.1.3 特征重要性分析 本研究中小麥產(chǎn)量和N2O排放效應(yīng)模型分別考慮不同的自變量, 不同自變量對(duì)模型重要性的量化結(jié)果如圖3所示。經(jīng)過(guò)模型選擇, 對(duì)應(yīng)于產(chǎn)量模型, N rate對(duì)于產(chǎn)量- RF模型的平均不純度減少最多, 達(dá)到0.43, 而SOM、LT_Prec、LT_Temp對(duì)于產(chǎn)量-RF模型的平均不純度減少達(dá)到0.12～0.14, 在區(qū)域尺度上N rate、SOM、LT_Prec、LT_Temp和土壤pH是影響產(chǎn)量形成的最重要變量。對(duì)于N2O模型, SOM、N rate和TN對(duì)于N2O-RF模型的平均不純度減少達(dá)到0.17～0.22, 而LT_Prec、LT_Temp在N2O-RF模型的貢獻(xiàn)度較低, 并不能區(qū)分作物種植的季節(jié)性, 很大程度上弱化了不同種植季節(jié)之間的溫度和降水量差異。

表1 田間管理措施、生態(tài)因子與N2O排放量的描述統(tǒng)計(jì)

SOM: 土壤有機(jī)質(zhì); TN: 土壤總氮; LT_Temp: 年平均溫度; LT_Prec: 年平均降雨量; Cum_N2O: N2O排放量。

SOM: soil organic matter; TN: total nitrogen; LT_Temp: the annual average temperature; LT_Prec: the annual average precipitation; Cum_N2O: cumulative N2O emission.

圖1 田間管理措施、生態(tài)因子與N2O排放量之間的皮爾遜相關(guān)性熱圖

縮寫同表1。Abbreviations are the same as these given in Table 1.

圖2 N2O排放和小麥產(chǎn)量模型的訓(xùn)練與驗(yàn)證結(jié)果

圖3 各個(gè)模型中的特征重要性

a: 一氧化二氮; b: 產(chǎn)量?？s寫同表1。a, N2O; b, yield. Abbreviations are the same as those given in Table 1.

2.2 不同目標(biāo)下的氮肥和秸稈投入措施推薦

2.2.1 不同情景下的小麥產(chǎn)量和N2O排放量模擬及函數(shù)構(gòu)建 通過(guò)在如皋的田間試驗(yàn)中獲取實(shí)測(cè)數(shù)據(jù)(SOM=13.34 g kg–1、TN=1.5 g kg–1、Split N=2次、LT_Prec =1145.28 mm、LT_Temp =17.49℃、pH 8.2), 并設(shè)置秸稈投入水平為0、2.25、4.50和6.75 t hm–2和氮肥投入水平為0到300 kg hm–2作為情景參數(shù), 基于所構(gòu)建的產(chǎn)量和N2O效應(yīng)模型對(duì)該地不同情景下的小麥產(chǎn)量和N2O累積排放量進(jìn)行估算, 并對(duì)反演的產(chǎn)量采用一元二次方程進(jìn)行擬合以及對(duì)反演的N2O累積排放量采用線性方程和指數(shù)方程進(jìn)行擬合, 擬合方程如表2所示, 其中N2O排放與施氮量之間的方程的斜率隨秸稈投入先增加后減少, 表現(xiàn)秸稈的增加可能會(huì)對(duì)礦化氮進(jìn)行固持, 減少了N2O排放生成的底物。而產(chǎn)量函數(shù)中負(fù)值的系數(shù)隨著秸稈投入而降低, 秸稈投入在較低氮肥投入下會(huì)抑制小麥的前期生長(zhǎng)從而抑制了產(chǎn)量形成, 而第二個(gè)系數(shù)為正值, 則隨著秸稈投入的增加而逐漸遞增, 表現(xiàn)為隨施肥量增加, 可以有效地滿足小麥生長(zhǎng)所需要的氮素營(yíng)養(yǎng), 從而形成合理的群體保證產(chǎn)量, 并且隨著施肥量的增加, 秸稈后期的養(yǎng)分效益逐漸顯現(xiàn)出來(lái), 表現(xiàn)出函數(shù)在較大的氮肥投入下產(chǎn)量會(huì)表現(xiàn)為秸稈投入高于無(wú)秸稈投入, 并且一定程度上秸稈的養(yǎng)分可以替代化學(xué)肥料的養(yǎng)分用于生產(chǎn)。

2.2.2 面向不同目標(biāo)下的氮肥和秸稈投入推薦及效益評(píng)估 基于表2的模型計(jì)算出不同秸稈還田量和氮肥投入下的CEE和NEEB如圖4所示, 對(duì)于N2O排放隨氮肥投入無(wú)論是呈現(xiàn)線性增長(zhǎng)還是指數(shù)增長(zhǎng), 在不同的秸稈投入下, 小麥的CEE變化均表現(xiàn)為隨氮肥量增加而呈現(xiàn)先降低后升高的趨勢(shì), 在CEE評(píng)估下, 不同秸稈投入下施用氮肥以113～130 kg hm–2為邊界線, 低于該氮肥投入水平N2O排放的增加會(huì)低于產(chǎn)量的增加, 直到CEE的邊界施氮量, 后面繼續(xù)增加施肥, 產(chǎn)量的增長(zhǎng)速度是低于N2O排放量的增長(zhǎng)的, 表現(xiàn)為N2O排放-氮肥投入呈現(xiàn)凹形的增長(zhǎng)曲線, 而產(chǎn)量-氮肥投入呈現(xiàn)凸性的增長(zhǎng)曲線。而NEEB則表現(xiàn)為先升高后下降的趨勢(shì), 這體現(xiàn)出將環(huán)境影響通過(guò)計(jì)算環(huán)境經(jīng)濟(jì)成本的形式進(jìn)行考慮, 增加氮肥投入所增加的產(chǎn)量收益是遠(yuǎn)大于其帶來(lái)的環(huán)境成本, 因此以NEEB最佳為目標(biāo)下推薦氮肥用量更側(cè)重于保證作物產(chǎn)量的穩(wěn)定, 同時(shí)減少了為追求產(chǎn)量最高下施肥所引起的環(huán)境影響。從表3可以看到, 觀察到在不同秸稈投入下產(chǎn)量最高、CEE最低、NEEB最高下的氮肥施用量以及經(jīng)濟(jì)收益。在0～6.75 t hm–2的秸稈投入下, 達(dá)到產(chǎn)量最高所需的氮肥量為208～212 kg hm–2, 并且隨秸稈投入的增加, 達(dá)到產(chǎn)量最高所需的氮肥量有所下降。在不同的秸稈投入中, 小麥達(dá)到CEE最低所需氮肥量在113～130 kg hm–2, 其凈生態(tài)收益也是最少的, 而達(dá)到最高NEEB所需氮肥則在202～205 kg hm–2, 其中6.75 t hm–2的秸稈投入下施用202 kg hm–2的氮肥可以獲得最高的生態(tài)收益13,669.18元。

表2 不同秸稈還田量情景的N2O累積排放和小麥產(chǎn)量模擬模型

S0、S2.25、S4.50、S6.75分別代表秸稈投入水平為0、2.25、4.50和6.75 t hm–2。

S0, S2.25, S4.50, and S6.75 represent the straw input levels of 0, 2.25, 4.50, and 6.75 t hm-2, respectively.

圖4 不同氮肥和秸稈投入下的CEE和NEEB響應(yīng)

處理同表2。Treatments are the same as those given in Table 2.

表3 不同目標(biāo)下小麥?zhǔn)┑亢徒斩捔客度爰捌鋬羯鷳B(tài)環(huán)境效益

(續(xù)表3)

3 討論

3.1 麥田N2O排放的主要驅(qū)動(dòng)因素

麥田的N2O排放受到管理措施(施氮量、秸稈還田、施肥次數(shù))的影響。本研究中, 氮肥施用與N2O排放通量呈現(xiàn)正相關(guān), 同時(shí)也是N2O-RF模型中的主要貢獻(xiàn)因子。其次, 氮肥的分次施用通常被認(rèn)為比一次性施用排放更低[55]。但本研究中Split N與N2O排放量呈現(xiàn)正相關(guān)關(guān)系, 與前人的研究有點(diǎn)差異, 可能是由于在長(zhǎng)江中下游地區(qū)多次施肥容易碰上降雨, 從而激發(fā)更高的N2O排放。此外, 秸稈的摻入和表面覆蓋在長(zhǎng)期試驗(yàn)中極大地影響了N2O排放量[56]。前人研究也表明, 對(duì)于土壤N2O排放, 秸稈還田具有抑制作用[57]、促進(jìn)[58]或無(wú)顯著的影響效果[59]。本研究結(jié)果表明, 秸稈還田量與N2O排放的相關(guān)分析呈現(xiàn)負(fù)相關(guān), 這些不一致的發(fā)現(xiàn)可能歸因于秸稈還田量和土壤類型的差異[50]。本研究的土壤N2O排放量在構(gòu)建的指數(shù)模型中也發(fā)現(xiàn)了隨著秸稈投入呈現(xiàn)先增加后減少的趨勢(shì)(表2)。施氮肥和秸稈還田都可能會(huì)刺激N2O排放[60]。此外, 本研究中基于Pearson相關(guān)分析發(fā)現(xiàn), N2O排放受到了年平均溫度的顯著影響, 較高的溫度通常會(huì)通過(guò)反硝化作用增加農(nóng)田生態(tài)系統(tǒng)中土壤N2O的排放量[61]。前人研究表明, SOM和土壤N含量也是影響土壤N2O排放變化的重要因素[62]。在考慮土壤N含量的作用時(shí), pH對(duì)于N2O排放量的預(yù)測(cè)并沒(méi)有具有較高的特征重要性, 這與前人的研究是一致的[63]。本研究揭示了土壤N2O排放的控制因素是復(fù)雜的, 雖然氣候因素(如年平均氣溫、年降雨量)、土壤理化性質(zhì)(如土壤pH值、SOM)顯著影響土壤N2O排放, 但作為硝化和反硝化作用底物的土壤N含量(土壤總氮), 在土壤N2O排放中也起到關(guān)鍵作用。

3.2 RF預(yù)測(cè)產(chǎn)量和N2O排放的可行性

了解和量化中國(guó)長(zhǎng)江中下游流域的產(chǎn)量和N2O排放量對(duì)于保證產(chǎn)量和減少環(huán)境污染至關(guān)重要。由于區(qū)域內(nèi)氣候、土壤和管理制度的變異性, 綜合探究不同氮肥和秸稈投入的響應(yīng)是復(fù)雜的。最近, 機(jī)器學(xué)習(xí)技術(shù)已成功應(yīng)用于農(nóng)業(yè)生產(chǎn), 以研究各種農(nóng)藝指標(biāo)[18]。作為一種流行的基于決策樹的集成機(jī)器學(xué)習(xí)算法, RF可以處理變量之間的非線性效應(yīng)和復(fù)雜的相互作用。通過(guò)RF模型的模擬, N2O排放和產(chǎn)量的估算是基于數(shù)據(jù)驅(qū)動(dòng)的, 而不依賴于預(yù)先指定的方程或函數(shù)形式。在這里, 我們將有關(guān)天氣、管理和土壤狀況的大型數(shù)據(jù)集與RF算法相結(jié)合, 以識(shí)別中國(guó)長(zhǎng)江中下游流域的稻麥輪作區(qū)的N2O排放量和產(chǎn)量。對(duì)于N2O排放來(lái)說(shuō), 使用線性模型僅靠基礎(chǔ)土壤數(shù)據(jù)和施氮量無(wú)法準(zhǔn)確預(yù)測(cè), 而RF模型對(duì)稻田下N2O總排放量的預(yù)測(cè)是顯著優(yōu)于其他模型的[64]。RF模型也已被用于預(yù)測(cè)玉米田的N2O排放和N淋溶[65]。而對(duì)于產(chǎn)量來(lái)說(shuō), 大量Meta分析的研究已經(jīng)表明土壤理化性質(zhì), 人為管理措施以及氮肥和秸稈使用量是主導(dǎo)產(chǎn)量變異的主要因素[55,66]。因此, 本研究中使用RF模型模擬不同氮肥和秸稈投入下麥田的產(chǎn)量和N2O總排放量是可行的, 該模型考慮了變量的非線性響應(yīng), 獲取優(yōu)異的性能(2= 0.65～0.96)。在本研究中, 模型模擬由于忽略了幾個(gè)因素而存在一些局限性。在這里, 我們搜集的數(shù)據(jù)中多是來(lái)自田間試驗(yàn)數(shù)據(jù), 大多是基于小麥的基本苗在225萬(wàn)公頃進(jìn)行試驗(yàn), 從而簡(jiǎn)化了現(xiàn)實(shí)生產(chǎn)中的密度響應(yīng), 這意味著改進(jìn)氮肥和秸稈的投入并不是孤立地進(jìn)行, 而是要與其他管理措施相結(jié)合, 例如增加密度, 以擴(kuò)大氮肥和秸稈投入的養(yǎng)分利用率。其次是需要通過(guò)考慮本研究無(wú)法解決的其他可能因素(例如倒伏、病蟲害風(fēng)險(xiǎn))來(lái)改進(jìn)N2O排放和產(chǎn)量預(yù)測(cè)。此外, 我們的N2O排放和產(chǎn)量模型是使用RF算法開發(fā)的, 但仍可能存在一些不足, 進(jìn)一步可以通過(guò)結(jié)合更多生物物理因子和機(jī)器學(xué)習(xí)建模來(lái)改進(jìn)N2O排放和產(chǎn)量的評(píng)估。盡管存在這些限制, 模型的良好性能表明, 氣候-土壤數(shù)據(jù)集與機(jī)器學(xué)習(xí)技術(shù)相結(jié)合是研究影響N2O排放和產(chǎn)量變化的有效方法。

3.3 麥田優(yōu)化施肥方案推薦

化肥施用不當(dāng)導(dǎo)致養(yǎng)分失衡、利用效率低下, 對(duì)環(huán)境造成大量損失, 已成為我國(guó)小麥生產(chǎn)系統(tǒng)的普遍現(xiàn)象。然而, 確定合適的施肥量仍然是基于科學(xué)的養(yǎng)分管理的基礎(chǔ)。以往農(nóng)民通常根據(jù)當(dāng)?shù)靥镩g試驗(yàn)或過(guò)去的經(jīng)驗(yàn)使用固定的氮肥施用量, 較容易造成環(huán)境污染[67]。

早期的施肥研究是基于糧食產(chǎn)量目標(biāo)出發(fā), 基于經(jīng)驗(yàn)分析的方法對(duì)產(chǎn)量/投入之間的關(guān)系進(jìn)行定量, 從而推薦施肥策略, 相較于農(nóng)戶方案降低了氮肥施用量并提高作物產(chǎn)量, 但忽略了環(huán)境問(wèn)題[68-70]。在考慮環(huán)境成本方面, 前人研究通過(guò)預(yù)測(cè)油菜籽產(chǎn)量、種植者的利潤(rùn)和EONR值, 從而為加拿大東部的油菜籽生產(chǎn)提供環(huán)境經(jīng)濟(jì)最佳的N推薦[19]。而優(yōu)化施氮量并不能完全滿足最佳氮肥管理, 一次性撒肥、施肥不足、施肥過(guò)量、秸稈不合理使用等傳統(tǒng)田間管理方式在中國(guó)許多地區(qū)的小農(nóng)戶中仍然存在。在本研究中, 通過(guò)結(jié)合特定地點(diǎn)的情景設(shè)置反演數(shù)據(jù)構(gòu)建的產(chǎn)量一元二次模型發(fā)現(xiàn), 隨氮肥投入增加, 產(chǎn)量呈現(xiàn)先增后減趨勢(shì)。當(dāng)秸稈還田量大于4.5 t hm–2時(shí), 本研究構(gòu)建的產(chǎn)量模型反演的產(chǎn)量高于無(wú)秸稈投入處理, 因此秸稈還田可以是提高作物產(chǎn)量的有效措施之一, 但同時(shí)會(huì)受到氮肥配施量的影響與調(diào)控, 通過(guò)改善管理實(shí)踐是實(shí)現(xiàn)以更少的投入生產(chǎn)穩(wěn)定的糧食和減少溫室氣體排放的重要策略。

目前追求優(yōu)化策略組合以提高糧食產(chǎn)量和降低環(huán)境成本, 仍然缺乏對(duì)不同目標(biāo)下的施肥措施策略組合進(jìn)行綜合的評(píng)估。本研究中, 選用碳排放強(qiáng)度作為排放最低下的生態(tài)環(huán)境效益的定量評(píng)估, 而NEEB則結(jié)合作物產(chǎn)量、農(nóng)業(yè)活動(dòng)和全球變暖潛能值的成本[50], 進(jìn)而考慮生態(tài)環(huán)境效益。經(jīng)濟(jì)效益通常是農(nóng)民改善農(nóng)藝管理和政策制定者提出有效農(nóng)業(yè)政策的主要?jiǎng)恿? 本研究表明, 在不同秸稈投入下達(dá)到CEE最低,會(huì)隨著秸稈投入的增加氮肥施用也有所增加, 并獲得更高的NEEB。這可能與CEE評(píng)估下的施肥推薦中, 為保證排放最低的前提是以犧牲產(chǎn)量為代價(jià)的。而相較于CEE評(píng)估, NEEB評(píng)估有助于以貨幣為基礎(chǔ)的方式讓農(nóng)民更容易評(píng)估科學(xué)決策, 并鼓勵(lì)他們采用環(huán)境友好型管理。與低秸稈投入相比, 在較高的秸稈投入下減少了氮肥施用并獲得了更高的NEEB, 結(jié)果有利于鼓勵(lì)農(nóng)民采用秸稈還田技術(shù), 有益于人們減少購(gòu)買氮肥的成本并有助于減少人們焚燒秸稈的行為?？偠灾? 秸稈還田在實(shí)際農(nóng)業(yè)生產(chǎn)中具備固碳減排的潛力, 配施合適的氮肥可以獲得更好的NEEB, 前人研究也與本試驗(yàn)的結(jié)果相似[16]。

4 結(jié)論

在稻麥復(fù)種模式下, 麥田N2O排放因秸稈投入和施肥不同存在顯著差異, 表現(xiàn)為N2O的累積排放量隨施氮量和秸稈還田量的增加而顯著增加。在區(qū)域尺度上, 通過(guò)搜集長(zhǎng)江中下游地區(qū)上氮肥和秸稈投入對(duì)小麥產(chǎn)量與N2O排放的相關(guān)文獻(xiàn)數(shù)據(jù), 并建立了RF效應(yīng)模型, 驗(yàn)證結(jié)果表明模型擬合情況良好, 結(jié)果表明氮肥和秸稈投入的小麥產(chǎn)量和N2O排放會(huì)受到人為管理因素、土壤因素和氣候因素的影響?；陂_發(fā)的模型結(jié)合情景設(shè)置進(jìn)行了試驗(yàn)地點(diǎn)的小麥產(chǎn)量和N2O排放模擬, 并評(píng)估了碳排放強(qiáng)度和凈生態(tài)環(huán)境經(jīng)濟(jì)效益。若追求高產(chǎn), 所需氮肥量為208～212 kg hm–2, 若以達(dá)到最小碳排放為目標(biāo), 所需氮肥量在113～130 kg hm–2, 若要實(shí)現(xiàn)最大生態(tài)經(jīng)濟(jì)效益, 所需氮肥則是在202～205 kg hm–2, 其中在6.75 t hm–2的秸稈投入下施用202 kg hm–2的氮肥可以獲得最高的生態(tài)收益13,669.18元。優(yōu)化氮肥和秸稈投入能在順應(yīng)綠色生產(chǎn)的前提下收獲理想的經(jīng)濟(jì)效益, 未來(lái)應(yīng)用前景廣闊。

[1] Reynolds M, Foulkes J, Furbank R, Griffiths S, King J, Murchie E, Parry M, Slafer G. Achieving yield gains in wheat., 2012, 35: 1799–1823.

[2] Van Dijk M, Morley T, Rau M L, Saghai Y. A meta-analysis of projected global food demand and population at risk of hunger for the period 2010–2050., 2021, 2: 494.

[3] Smith K A. Changing views of nitrous oxide emissions from agricultural soil: key controlling processes and assessment at different spatial scales., 2017, 68: 137–155.

[4] Cowan N, Levy P, Maire J, Coyle M, Leeson S R, Famulari D, Carozzi M, Nemitz E, Skiba U. An evaluation of four years of nitrous oxide fluxes after application of ammonium nitrate and urea fertilisers measured using the eddy covariance method., 2020, 280: 107812.

[5] Meng Q F, Yue S C, Hou P, Cui Z L, Chen X P. Improving Yield and Nitrogen Use Efficiency Simultaneously for Maize and Wheat in China: a review., 2016, 26: 137–147.

[6] Millar N, Urrea A, Kahmark K, Shcherbak I, Robertson G P, Ortiz-Monasterio I. Nitrous oxide (N2O) flux responds exponentially to nitrogen fertilizer in irrigated wheat in the Yaqui Valley, Mexico., 2018, 261: 125–132.

[7] Song X T, Liu M, Ju X T, Gao B, Su F, Chen X P, Rees R M. Nitrous oxide emissions increase exponentially when optimum nitrogen fertilizer rates are exceeded in the North China Plain., 2018, 52: 12504–12513.

[8] Duan J Z, Shao Y H, He L, Li X, Hou G G, Li S N, Feng W, Zhu Y J, Wang Y H, Xie Y X. Optimizing nitrogen management to achieve high yield, high nitrogen efficiency and low nitrogen emission in winter wheat., 2019, 697: 12.

[9] Lyu J L, Yin X H, Dorich C, Olave R, Wang X H, Kou C L, Song X. Net field global warming potential and greenhouse gas intensity in typical arid cropping systems of China: a 3-year field measurement from long-term fertilizer experiments., 2021, 212: 105053.

[10] Chen H H, Li X C, Hu F, Shi W. Soil nitrous oxide emissions following crop residue addition: a meta-analysis., 2013, 19: 2956–2964.

[11] Wang J, Zhu B, Zhang J B, Muller C, Cai Z C. Mechanisms of soil N dynamics following long-term application of organic fertilizers to subtropical rain-fed purple soil in China., 2015, 91: 222–231.

[12] Huang T, Gao B, Christie P, Ju X. Net global warming potential and greenhouse gas intensity in a double-cropping cereal rotation as affected by nitrogen and straw management., 2013, 10: 7897–7911.

[13] Ambus P, Jensen E S, Robertson G P. Nitrous oxide and N-leaching losses from agricultural soil: influence of crop residue particle size, quality and placement., 2001, 41: 7–15.

[14] Yang L, Muhammad I, Chi Y X, Wang D, Zhou X B. Straw return and nitrogen fertilization to maize regulate soil properties, microbial community, and enzyme activities under a dual cropping system., 2022, 13: 823963.

[15] Huang T, Yang H, Huang C C, Ju X T. Effect of fertilizer N rates and straw management on yield-scaled nitrous oxide emissions in a maize-wheat double cropping system., 2017, 204: 1–11.

[16] Akhtar K, Wang W Y, Ren G X, Khan A, Enguang N, Khan A, Feng Y Z, Yang G H, Wang H Y. Straw mulching with inorganic nitrogen fertilizer reduces soil CO2and N2O emissions and improves wheat yield., 2020, 741: 140488.

[17] Glenn A J, Moulin A P, Roy A K, Wilson H F. Soil nitrous oxide emissions from no-till canola production under variable rate nitrogen fertilizer management., 2021, 385: 114857.

[18] Cao J, Zhang Z, Tao F, Zhang L, Luo Y, Zhang J, Han J, Xie J. Integrating multi-source data for rice yield prediction across China using machine learning and deep learning approaches., 2021, 297: 108275.

[19] Wen G, Ma B L, Vanasse A, Caldwell C D, Smith D L. Optimizing machine learning-based site-specific nitrogen application recommendations for canola production., 2022, 288: 108707.

[20] Chen S, Huang Y, Zou J. Relationship between nitrous oxide emission and winter wheat production., 2008, 44: 985–989.

[21] Gao X, Lan T, Deng L, Zeng M. Mushroom residue application affects CH4and N2O emissions from fields under rice-wheat rotation., 2017, 63: 748–760.

[22] Guo L, Zhang L, Liu L, Sheng F, Cao C, Li C. Effects of long-term no tillage and straw return on greenhouse gas emissions and crop yields from a rice-wheat system in central China., 2021, 322: 107650.

[23] Guo T, Luan H, Song D, Zhang S, Zhou W, Liang G. Combined fertilization could increase crop productivity and reduce greenhouse gas intensity through carbon sequestration under rice-wheat rotation.(Basel), 2021, 11: 103390.

[24] He H, Li D, Pan F, Wang F, Wu D, Yang S. Effects of nitrogen reduction and optimized fertilization combined with straw return on greenhouse gas emissions and crop yields of a rice-wheat rotation system., 2022, 16: 669–679.

[25] He H, Zhang T, Yao Y, Yang W, Busayo D, Wen X, Chen X, Yang X, Yang S, Ma Y. Tillage methods on greenhouse gas emissions and yields of rice-wheat rotation system in east China polder area., 2021, 15: 485–498.

[26] Hu N, Wang B, Gu Z, Tao B, Zhang Z, Hu S, Zhu L, Meng Y. Effects of different straw returning modes on greenhouse gas emissions and crop yields in a rice-wheat rotation system., 2016, 223: 115–122.

[27] Ji Y, Liu G, Ma J, Xu H, Yagi K. Effect of controlled-release fertilizer on nitrous oxide emission from a winter wheat field., 2012, 94: 111–122.

[28] Li S H, Guo L J, Cao C G, Li C F. Effects of straw returning levels on carbon footprint and net ecosystem economic benefits from rice-wheat rotation in central China., 2021, 28: 5742–5754.

[29] Liu G, Ma J, Yang Y, Yu H, Zhang G, Xu H. Effects of straw incorporation methods on nitrous oxide and methane emissions from a wheat-rice rotation system., 2019, 29: 204–215.

[30] Liu S, Qin Y, Zou J, Liu Q. Effects of water regime during rice-growing season on annual direct N2O emission in a paddy rice-winter wheat rotation system in southeast China., 2010, 408: 906–913.

[31] Ma E, Zhang G, Ma J, Xu H, Cai Z, Yagi K. Effects of rice straw returning methods on N2O emission during wheat-growing season., 2010, 88: 463–469.

[32] Ma Y C, Kong X W, Yang B, Zhang X L, Yan X Y, Yang J C, Xiong Z Q. Net global warming potential and greenhouse gas intensity of annual rice-wheat rotations with integrated soil-crop system management., 2013, 164: 209–219.

[33] Wang H, Shen M, Hui D, Chen J, Sun G, Wang X, Lu C, Sheng J, Chen L, Luo Y, Zheng J, Zhang Y. Straw incorporation influences soil organic carbon sequestration, greenhouse gas emission, and crop yields in a Chinese rice (L.)-wheat (L.) cropping system., 2019, 195: 104377.

[34] Xia L, Wang S, Yan X. Effects of long-term straw incorporation on the net global warming potential and the net economic benefit in a rice-wheat cropping system in China., 2014, 197: 118–127.

[35] Xiang J, Liu D, Ding W, Yuan J, Lin Y. Effects of biochar on nitrous oxide and nitric oxide emissions from paddy field during the wheat growth season., 2015, 104: 52–58.

[36] Yang B, Xiong Z, Wang J, Xu X, Huang Q, Shen Q. Mitigating net global warming potential and greenhouse gas intensities by substituting chemical nitrogen fertilizers with organic fertilization strategies in rice-wheat annual rotation systems in China: a 3-year field experiment., 2015, 81: 289–297.

[37] Yao Z, Zheng X, Wang R, Xie B, Butterbach-Bahl K, Zhu J. Nitrous oxide and methane fluxes from a rice-wheat crop rotation under wheat residue incorporation and no-tillage practices., 2013, 79: 641–649.

[38] Yao Z, Zheng X, Xie B, Mei B, Wang R, Butterbach-Bahl K, Zhu J, Yin R. Tillage and crop residue management significantly affects N-trace gas emissions during the non-rice season of a subtropical rice-wheat rotation., 2009, 41: 2131–2140.

[39] Zhang L, Zheng J, Chen L, Shen M, Zhang X, Zhang M, Bian X, Zhang J, Zhang W. Integrative effects of soil tillage and straw management on crop yields and greenhouse gas emissions in a rice-wheat cropping system., 2015, 63: 47–54.

[40] Zou J, Huang Y, Lu Y, Zheng X, Wang Y. Direct emission factor for N2O from rice-winter wheat rotation systems in southeast China., 2005, 39: 4755–4765.

[41] 江波, 楊書運(yùn), 馬友華, 賀非, 左懷峰, 范東福, 楊小兵. 耕作方式對(duì)圩區(qū)冬小麥溫室氣體排放通量的影響. 安徽農(nóng)業(yè)大學(xué)學(xué)報(bào), 2014, 41: 241–247. Jiang B, Yang S Y, Ma Y H, He F, Zuo H F, Fan D F, Yang X B, Effects on emission of greenhouse gas by different tillage treatments to winter wheat in polder areas., 2014, 41: 241–247 (in Chinese with English abstract).

[42] 靳紅梅, 沈明星, 王海候, 陸長(zhǎng)嬰, 常志州, 郭瑞華. 秸稈還田模式對(duì)稻麥兩熟農(nóng)田麥季CH4和N2O排放特征的影響, 江蘇農(nóng)業(yè)學(xué)報(bào), 2017, 33: 333–339. Jin H M, Shen M X, Wang H H, Lu C Y, Chang Z Z, Guo R H. Influence of straw returning patterns on CH4and N2O emission during wheat-growing season in a rice-wheat double cropping system., 2017, 33: 333–339 (in Chinese with English abstract).

[43] 牛東, 潘慧, 叢美娟, 尹萍, 吳浩, 孫娟, 朱新開, 郭文善. 氮肥運(yùn)籌和秸稈還田對(duì)麥季土壤溫室氣體排放的影響. 麥類作物學(xué)報(bào), 2016, 36: 1667–1673.Niu D, Pan H, Cong M J, Yin P, Wu H, Sun J, Zhu X K, Guo W S. Effect of nitrogen application ratio and straw returning on soil greenhouse gas emission during wheat growing period., 2016, 36: 1667–1673 (in Chinese with English abstract).

[44] 孫國(guó)峰, 鄭建初, 陳留根, 何加駿, 張?jiān)婪? 配施豬糞對(duì)麥季CH4和N2O排放及溫室效應(yīng)的影響. 生態(tài)與農(nóng)村環(huán)境學(xué)報(bào), 2012, 28: 349–354. Sun G F, Zheng J C, Chen L G, He J J, Zhang Y F. Effects of application of pig manure in combination with chemical fertili-zers on CH4and N2O emissions and their greenhouse effects in wheat field., 2012, 28: 349–354 (in Chinese with English abstract).

[45] 孫國(guó)峰, 鄭建初, 陳留根, 何加駿, 張?jiān)婪? 沼液替代化肥對(duì)麥季CH4、N2O排放及溫室效應(yīng)的影響. 農(nóng)業(yè)環(huán)境科學(xué)學(xué)報(bào), 2012, 31: 1654–1661. Sun G F, Zheng J C, Chen L G, He J J, Zhang Y F. Effects of chemical fertilizers substitution by biogas slurry on CH4and N2O emissions and their greenhouse effects in wheat field., 2012, 31: 1654–1661 (in Chinese with English abstract).

[46] 王海云, 邢光熹. 不同施氮水平對(duì)稻麥輪作農(nóng)田氧化亞氮排放的影響. 農(nóng)業(yè)環(huán)境科學(xué)學(xué)報(bào), 2009, 28: 2631–2646. Wang H Y, Xing G X. Effect of nitrogen fertilizer rates on nitrous oxide emission from paddy field under rice-wheat rotation., 2009, 28: 2631–2636 (in Chinese with English abstract).

[47] 張翰林, 呂衛(wèi)光, 鄭憲清, 李雙喜, 王金慶, 張娟琴, 何七勇, 袁大偉, 顧曉君. 不同秸稈還田年限對(duì)稻麥輪作系統(tǒng)溫室氣體排放的影響. 中國(guó)生態(tài)農(nóng)業(yè)學(xué)報(bào), 2015, 23: 302–308. Zhang H L, Lyu W G, Zheng X Q, Li S X, Wang J Q, Zhang J Q, He Q Y, Yuan D W, Gu X J. Effects of years of straw return to soil on greenhouse gas emission in rice/wheat rotation systems., 2015, 23: 302–308 (in Chinese with English abstract).

[48] 張?jiān)婪? 陳留根, 朱普平, 張傳勝, 盛婧, 王子臣, 鄭建初. 秸稈還田對(duì)稻麥兩熟高產(chǎn)農(nóng)田凈增溫潛勢(shì)影響的初步研究. 農(nóng)業(yè)環(huán)境科學(xué)學(xué)報(bào), 2012, 31: 1647–1653. Zhang Y F, Chen L G, Zhu P P, Zhang C S, Sheng J, Wang Z C, Zheng J C. Preliminary study on effect of straw incorporation on net global warming potential in high production rice-wheat double cropping systems., 2012, 31: 1647–1653 (in Chinese with English abstract).

[49] 鄒建文. 稻麥輪作生態(tài)系統(tǒng)溫室氣體(CO2、CH4和N2O)排放研究. 南京農(nóng)業(yè)大學(xué)博士學(xué)位論文, 江蘇南京, 2005. Zou J W. A Study on Greenhouse Gases (CO2, CH4and N2O) Emission from Rice-winter Wheat Rotations in Southeast China. PhD Dissertation of Nanjing Agricultural University, Nanjing, Jiangsu, China, 2005 (in Chinese with English abstract).

[50] Zhang Z S, Guo L J, Liu T Q, Li C F, Cao C G. Effects of tillage practices and straw returning methods on greenhouse gas emissions and net ecosystem economic budget in rice wheat cropping systems in central China., 2015, 122: 636–644.

[51] Li S H, Guo L J, Cao C G, Li C F. Effects of straw returning levels on carbon footprint and net ecosystem economic benefits from rice-wheat rotation in central China., 2021, 28: 5742–5754.

[52] Xia L L, Xia Y Q, Li B L, Wang J Y, Wang S W, Zhou W, Yan X Y. Integrating agronomic practices to reduce greenhouse gas emissions while increasing the economic return in a rice-based cropping system., 2016, 231: 24–33.

[53] Li B, Fan C H, Zhang H, Chen Z Z, Sun L Y, Xiong Z Q. Combined effects of nitrogen fertilization and biochar on the net global warming potential, greenhouse gas intensity and net ecosystem economic budget in intensive vegetable agriculture in southeastern China., 2015, 100: 10–19.

[54] Lu R, Zhang P, Fu Z, Jiang J, Wu J, Cao Q, Tian Y, Zhu Y, Cao W, Liu X. Improving the spatial and temporal estimation of ecosystem respiration using multi-source data and machine learning methods in a rainfed winter wheat cropland., 2023, 871: 161967.

[55] Guo C, Liu X, He X. A global meta-analysis of crop yield and agricultural greenhouse gas emissions under nitrogen fertilizer application., 2022, 831: 154982.

[56] Muhammad I, Wang J, Sainju U M, Zhang S H, Zhao F Z, Khan A. Cover cropping enhances soil microbial biomass and affects microbial community structure: a meta-analysis., 2021, 381: 114696.

[57] Zhang Y Y, Liu J F, Mu Y J, Pei S W, Lun X X, Chai F H. Emissions of nitrous oxide, nitrogen oxides and ammonia from a maize field in the North China Plain., 2011, 45: 2956–2961.

[58] Liu C Y, Wang K, Meng S X, Zheng X H, Zhou Z X, Han S H, Chen D L, Yang Z P. Effects of irrigation, fertilization and crop straw management on nitrous oxide and nitric oxide emissions from a wheat-maize rotation field in northern China., 2011, 140: 226–233.

[59] Yao Z S, Zheng X H, Xie B H, Mei B L, Wang R, Butterbach-Bahl K, Zhu J G, Yin R. Tillage and crop residue management significantly affects N-trace gas emissions during the non- rice season of a subtropical rice-wheat rotation., 2009, 41: 2131–2140.

[60] Garcia-Ruiz R, Gomez-Munoz B, Hatch D J, Bol R, Baggs E M. Soil mineral N retention and N2O emissions following combined application of15N-labelled fertiliser and weed residues., 2012, 26: 2379–2385.

[61] Wang Y Y, Hu Z H, Shang D Y, Xue Y, Islam A, Chen S T. Effects of warming and elevated O3concentrations on N2O emission and soil nitrification and denitrification rates in a wheat- soybean rotation cropland., 2020, 257: 113556.

[62] Bhattacharyya P, Nayak A K, Mohanty S, Tripathi R, Shahid M, Kumar A, Raja R, Panda B B, Roy K S, Neogi S, Dash P K, Shukla A K, Rao K S. Greenhouse gas emission in relation to labile soil C, N pools and functional microbial diversity as influenced by 39 years long-term fertilizer management in tropical rice., 2013, 129: 93–105.

[63] Li Z, Zeng Z, Song Z, Tian D, Huang X, Nie S, Wang J, Jiang L, Luo Y, Cui J, Niu S. Variance and main drivers of field nitrous oxide emissions: a global synthesis., 2022, 353: 131686.

[64] Jiang Z W, Yang S H, Chen X, Pang Q Q, Xu Y, Qi S T, Yu W Q, Dai H D. Controlled release urea improves rice production and reduces environmental pollution: a research based on meta-analysis and machine learning., 2022, 29: 3587–3599.

[65] Villa-Vialaneix N, Follador M, Ratto M, Leip A. A comparison of eight metamodeling techniques for the simulation of N2O fluxes and N leaching from corn crops., 2012, 34: 51–66.

[66] Qiu H H, Wei W L. Crop straw retention influenced crop yield and greenhouse gas emissions under various external conditions., 2021, 28: 42362–42371.

[67] Rasouli S, Whalen J K, Madramootoo C A. Review: reducing residual soil nitrogen losses from agroecosystems for surface water protection in Quebec and Ontario, Canada: best management practices, policies and perspectives., 2014, 94: 109–127.

[68] Qin Z, Myers D B, Ransom C J, Kitchen N R, Liang S Z, Camberato J J, Carter P R, Ferguson R B, Fernandez F G, Franzen D W, Laboski C A M, Malone B D, Nafziger E D, Sawyer J E, Shanahan J F. Application of machine learning methodologies for predicting corn economic optimal nitrogen rate., 2018, 110: 2596–2607.

[69] 劉新偉, 龔德平, 鞏細(xì)民, 王巍, 婁希鳳, 韓玲君, 楊德樺, 趙竹青. 湖北江北農(nóng)場(chǎng)小麥肥效試驗(yàn)與施肥推薦. 麥類作物學(xué)報(bào), 2012, 32: 338–343. Liu X W, Gong D P, Gong X M, Wang W, Lou X F, Han L J, Yang D H, Zhao Z Q. Fertilizer effect on wheat and recommendation offertilizer for wheat production in Jiangbei farm., 2012, 32: 338–343 (in Chinese with English abstract).

[70] 周琦, 李嵐?jié)? 張露露, 苗玉紅, 王宜倫. 氮肥和播種量互作對(duì)冬小麥產(chǎn)量、生長(zhǎng)發(fā)育和生態(tài)場(chǎng)特性的影響. 作物學(xué)報(bào), 2023, 49: 3100–3109. Zhou Q, Li L T, Zhang L L, Miao Y H, Wang Y L. Effects of interaction of nitrogen level and sowing rate on yield, growth, and ecological field characteristics of winter wheat., 2023, 49: 3100–3109 (in Chinese with English abstract).

Research on the effects of nitrogen fertilizer and rice straw return on wheat yield and N2O emission and recommended fertilization under rice-wheat rotation pattern

LU Ru-Hua1, WANG Wen-Xuan2, CAO Qiang1, TIAN Yong-Chao1, ZHU Yan1, CAO Wei-Xing1, and LIU Xiao-Jun1,*

1National Engineering and Technology Center for Information Agriculture, Nanjing Agricultural University / Engineering and Research Center for Smart Agriculture, Ministry of Education / Key Laboratory for Crop System Analysis and Decision Making, Ministry of Agriculture and Rural Affairs / Jiangsu Key Laboratory for Information Agriculture, Nanjing 210095, Jiangsu, China;2College of Humanities & Social Development, Nanjing 210095, Jiangsu, China

The optimization of agricultural practices such as nitrogen and straw input may be an effective option for maintaining environmental sustainability. However, previous studies on the effects of nitrogen and straw inputs on wheat growth and N2O emission reduction were limited. Therefore, the present study was based on the literature published from 2000 to 2022 about wheat yield and N2O emissions under different nitrogen and straw inputs amendment in the middle and lower reaches of the Yangtze River, a random forest (RF) model of wheat yield and N2O emission was constructed. And the influence of nitrogen and straw inputs on wheat yield and N2O emissions was quantified. Based on the developed model, wheat yield and N2O emission simulations at the experimental site were carried out in combination with scenario settings, and the carbon emission intensity (CEE) and net ecosystem economic benefits (NEEB) were evaluated. The results were as follow: On the regional scale, an RF model was established for the response of wheat yield and N2O emission to the application of nitrogen fertilizer and straw returning. The verification results were2of 0.66 and 0.65, and RMSE of 0.70 and 1.11, respectively. Quantifying the importance of independent variables showed that nitrogen application rate and soil organic matter were essential for yield and N2O models. For nitrogen fertilizer and straw management under different targets, the amount of nitrogen fertilizer required to achieve the highest yield was 208-212 kg hm–2, the amount of nitrogen fertilizer required to achieve the minimum CEE was 113-130 kg hm–2, and the amount of nitrogen fertilizer required to achieve the highest NEEB was 202–205 kg hm–2, of which the highest ecological benefit of 13,669.18 CHY could be obtained by applying 202 kg hm–2nitrogen fertilizer under the straw input of 6.75 t hm–2. Our results indicate that optimizing nitrogen fertilizer and straw inputs has the potential to reduce crop carbon emission intensity and maximize net ecological and environmental benefits.

nitrogen application rate; straw inputs; wheat; N2O; emission model; fertilizer recommendation

10.3724/SP.J.1006.2024.31035

本研究由國(guó)家重點(diǎn)研發(fā)計(jì)劃項(xiàng)目(2022YFD2301402), 南京農(nóng)業(yè)大學(xué)三亞研究院(NAUSY-ZD01)和國(guó)家自然科學(xué)基金項(xiàng)目(32071903)資助。

This study was supported by the National Key Research and Development Program of China (2022YFD2301402), the Hainan Institute of Nanjing Agricultural University (NAUSY-ZD01), and the National Natural Science Foundation of China (32071903).

劉小軍, E-mail: liuxj@njau.edu.cn, Tel: 025-84396804

E-mail: 2020101177@njau.edu.cn

2023-06-04;

2023-10-23;

2023-11-30.

URL: https://link.cnki.net/urlid/11.1809.S.20231129.1348.004

This is an open access article under the CC BY-NC-ND license (http://creativecommons.org/licenses/by-nc-nd/4.0/).