Pub Date : 2024-08-29DOI: 10.1016/j.eja.2024.127329
T. McCarthy , D.P. Wall , P.J. Forrestal , I.A. Casey , J. Humphreys
A proportion of potassium (K) exits grassland-based dairy farms in tradeable products. Potassium imports are typically needed to offset depletion of soil reserves. The objectives of this study were to (i) quantify K entering and exiting a grassland-based dairy farm including K lost to water, (ii) to relate the balance between K entering and existing the farm to soil K fertility status in order to (iii) design a better K fertilisation strategy for grassland under temperate climatic conditions. The quantities of K entering and exiting a grassland-based dairy farm (Solohead Research Farm; 52⁰51’N, 08⁰21’W) were determined each year between 2005 and 2022. Potassium losses to groundwater were measured during the winters of 2020/21, 2021/22 and 2022/23. Averaged over 18 years, K entering (kg ha−1 ± standard error) was 82 ± 11 and exiting was 41 ± 4. The annual average farm K balance was 41 ± 12 kg ha−1 and ranged between −36 and 136 kg ha−1. Annual K loss to groundwater (mean ± SE kg ha−1) ranged between 6.9 ± 6.13 and 59 ± 7.4. Annual average soil test K (STK; following extraction using Morgan's solution (Na acetate + acetic acid, pH 4.8)) concentrations in paddocks across the farm ranged from 85 to 253 mg L−1. The yearly change in average STK concentrations correlated with annual farm K balance in the preceding year (R2=0.59; P<0.001). Annual farm-scale K budgets were useful in quantifying K flows in products and losses. Potassium leaching to groundwater represented the majority (55 %) of K exiting the farm; exceeding export of K in milk and other products. Maintaining overall farm STK status required annual fertiliser K inputs of 22.5 kg ha−1 between 2016 and 2022. This study elucidates the challenges in managing soil K fertility on grassland based dairy farms.
{"title":"Circularity of potassium in a grassland-based dairy farm on a clay loam soil","authors":"T. McCarthy , D.P. Wall , P.J. Forrestal , I.A. Casey , J. Humphreys","doi":"10.1016/j.eja.2024.127329","DOIUrl":"10.1016/j.eja.2024.127329","url":null,"abstract":"<div><p>A proportion of potassium (K) exits grassland-based dairy farms in tradeable products. Potassium imports are typically needed to offset depletion of soil reserves. The objectives of this study were to (i) quantify K entering and exiting a grassland-based dairy farm including K lost to water, (ii) to relate the balance between K entering and existing the farm to soil K fertility status in order to (iii) design a better K fertilisation strategy for grassland under temperate climatic conditions. The quantities of K entering and exiting a grassland-based dairy farm (Solohead Research Farm; 52⁰51’N, 08⁰21’W) were determined each year between 2005 and 2022. Potassium losses to groundwater were measured during the winters of 2020/21, 2021/22 and 2022/23. Averaged over 18 years, K entering (kg ha<sup>−1</sup> ± standard error) was 82 ± 11 and exiting was 41 ± 4. The annual average farm K balance was 41 ± 12 kg ha<sup>−1</sup> and ranged between −36 and 136 kg ha<sup>−1</sup>. Annual K loss to groundwater (mean ± SE kg ha<sup>−1</sup>) ranged between 6.9 ± 6.13 and 59 ± 7.4. Annual average soil test K (STK; following extraction using Morgan's solution (Na acetate + acetic acid, pH 4.8)) concentrations in paddocks across the farm ranged from 85 to 253 mg L<sup>−1</sup>. The yearly change in average STK concentrations correlated with annual farm K balance in the preceding year (R<sup>2</sup>=0.59; P<0.001). Annual farm-scale K budgets were useful in quantifying K flows in products and losses. Potassium leaching to groundwater represented the majority (55 %) of K exiting the farm; exceeding export of K in milk and other products. Maintaining overall farm STK status required annual fertiliser K inputs of 22.5 kg ha<sup>−1</sup> between 2016 and 2022. This study elucidates the challenges in managing soil K fertility on grassland based dairy farms.</p></div>","PeriodicalId":51045,"journal":{"name":"European Journal of Agronomy","volume":"160 ","pages":"Article 127329"},"PeriodicalIF":4.5,"publicationDate":"2024-08-29","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"https://www.sciencedirect.com/science/article/pii/S1161030124002508/pdfft?md5=83a005e3eb8cc7a179232ec895be8a0b&pid=1-s2.0-S1161030124002508-main.pdf","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"142097928","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":1,"RegionCategory":"农林科学","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":"OA","EPubDate":null,"PubModel":null,"JCR":null,"JCRName":null,"Score":null,"Total":0}
Pub Date : 2024-08-27DOI: 10.1016/j.eja.2024.127319
Haidee Tang , Xiaojun Zhai , Xiangming Xu
The timing of the transition between endodormancy and ecodormancy remains uncertain. However, with advancements in phenology modelling, we can now fit models which allow for variable transitions between chilling and forcing models. Previous studies have primarily focused on single-cultivar parameterisation, and few have explored multi-cultivar comparative modelling. In this paper, we address this gap by evaluating three parameterisation approaches based on the recently developed PhenoFlex framework using a large flowering time dataset of twenty-six apple cultivars collected at the same location in England. The three parameterisation approaches were: cultivar-specific, group-specific with the groups derived using the K-means algorithm on mean bloom and variation of bloom dates, and a common model (for all twenty-six cultivars). The three PhenoFlex models fitted to each of three groups of cultivars based on their flowering time and the common model fitted to all cultivars achieved similar predictive performance, better than predictions using the average bloom date of each cultivar. The best approach to apply would depend on the amount of data present. The common model works best with large number of cultivars with small datasets (∼10 years), the mean flowering date grouped works best with medium numbers of datasets (∼20 years) and the cultivar-specific model should only be used when each cultivar has at least 30 years of data, however, it is more biased, so it is likely to predict bloom dates later than the observed bloom dates. Finally, the PhenoFlex model was shown to perform better than the StepChill model, where no overlapping is allowed between chilling and heat models. The result of this study indicates that the PhenoFlex model can be used to determine apple flowering time at the species level.
{"title":"Evaluating the performance of models predicting the flowering times of twenty-six apple cultivars in England","authors":"Haidee Tang , Xiaojun Zhai , Xiangming Xu","doi":"10.1016/j.eja.2024.127319","DOIUrl":"10.1016/j.eja.2024.127319","url":null,"abstract":"<div><p>The timing of the transition between endodormancy and ecodormancy remains uncertain. However, with advancements in phenology modelling, we can now fit models which allow for variable transitions between chilling and forcing models. Previous studies have primarily focused on single-cultivar parameterisation, and few have explored multi-cultivar comparative modelling. In this paper, we address this gap by evaluating three parameterisation approaches based on the recently developed PhenoFlex framework using a large flowering time dataset of twenty-six apple cultivars collected at the same location in England. The three parameterisation approaches were: cultivar-specific, group-specific with the groups derived using the K-means algorithm on mean bloom and variation of bloom dates, and a common model (for all twenty-six cultivars). The three PhenoFlex models fitted to each of three groups of cultivars based on their flowering time and the common model fitted to all cultivars achieved similar predictive performance, better than predictions using the average bloom date of each cultivar. The best approach to apply would depend on the amount of data present. The common model works best with large number of cultivars with small datasets (∼10 years), the mean flowering date grouped works best with medium numbers of datasets (∼20 years) and the cultivar-specific model should only be used when each cultivar has at least 30 years of data, however, it is more biased, so it is likely to predict bloom dates later than the observed bloom dates. Finally, the PhenoFlex model was shown to perform better than the StepChill model, where no overlapping is allowed between chilling and heat models. The result of this study indicates that the PhenoFlex model can be used to determine apple flowering time at the species level.</p></div>","PeriodicalId":51045,"journal":{"name":"European Journal of Agronomy","volume":"160 ","pages":"Article 127319"},"PeriodicalIF":4.5,"publicationDate":"2024-08-27","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"https://www.sciencedirect.com/science/article/pii/S1161030124002405/pdfft?md5=6b729707294ffdd8ea41b4aa05cc7469&pid=1-s2.0-S1161030124002405-main.pdf","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"142088590","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":1,"RegionCategory":"农林科学","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":"OA","EPubDate":null,"PubModel":null,"JCR":null,"JCRName":null,"Score":null,"Total":0}
Pub Date : 2024-08-27DOI: 10.1016/j.eja.2024.127323
Naijie Chang , Di Chen
Soil organic matter (SOM) is crucial for karst ecosystems, affecting cropland health, climate change mitigation, and rocky desertification control. However, there are limited research on cropland SOM prediction in karst areas with complex topography and diverse microclimates. Here, we compared the performance of four machine learning algorithms—random forest (RF), support vector regression (SVR), multilayer perceptron regression (MLP), and gradient boosting regression trees (GBRT)—for predicting cropland SOM in a typical karst landform area in 2019. Our results indicated that the GBRT model achieved the highest prediction accuracy with an R² of 0.69, MAE of 2.19 g/kg, RMSE of 3.37 g/kg, and LCCC of 0.82. Using the GBRT model and spatial data on climate, topography, and remote sensing, we predicted SOM for each 30 m × 30 m grid cell. The analysis revealed higher SOM content in the northeastern and southwestern regions and lower content in the central area, ranging from 13.95 to 47.81 g/kg, with an average of 27.16 g/kg. Lime soil had the highest SOM content, while purple soil had the lowest. Paddy fields showed significantly higher SOM than dry land. Over the past 40 years, SOM content has slightly increased, while its spatial distribution has remained stable.
土壤有机质(SOM)对岩溶生态系统至关重要,影响着耕地健康、气候变化减缓和石漠化控制。然而,在地形复杂、小气候多样的喀斯特地区,有关耕地土壤有机质预测的研究十分有限。在此,我们比较了四种机器学习算法--随机森林(RF)、支持向量回归(SVR)、多层感知器回归(MLP)和梯度提升回归树(GBRT)--在2019年典型喀斯特地貌地区预测耕地SOM的性能。结果表明,GBRT 模型的预测精度最高,R² 为 0.69,MAE 为 2.19 g/kg,RMSE 为 3.37 g/kg,LCCC 为 0.82。利用 GBRT 模型以及气候、地形和遥感空间数据,我们预测了每个 30 m × 30 m 网格单元的 SOM。分析表明,东北部和西南部地区的 SOM 含量较高,中部地区较低,从 13.95 克/千克到 47.81 克/千克不等,平均为 27.16 克/千克。石灰土的 SOM 含量最高,紫色土最低。水田的 SOM 含量明显高于旱地。在过去 40 年中,SOM 含量略有增加,但其空间分布保持稳定。
{"title":"Prediction of soil organic matter using Landsat 8 data and machine learning algorithms in typical karst cropland in China","authors":"Naijie Chang , Di Chen","doi":"10.1016/j.eja.2024.127323","DOIUrl":"10.1016/j.eja.2024.127323","url":null,"abstract":"<div><p>Soil organic matter (SOM) is crucial for karst ecosystems, affecting cropland health, climate change mitigation, and rocky desertification control. However, there are limited research on cropland SOM prediction in karst areas with complex topography and diverse microclimates. Here, we compared the performance of four machine learning algorithms—random forest (RF), support vector regression (SVR), multilayer perceptron regression (MLP), and gradient boosting regression trees (GBRT)—for predicting cropland SOM in a typical karst landform area in 2019. Our results indicated that the GBRT model achieved the highest prediction accuracy with an R² of 0.69, MAE of 2.19 g/kg, RMSE of 3.37 g/kg, and LCCC of 0.82. Using the GBRT model and spatial data on climate, topography, and remote sensing, we predicted SOM for each 30 m × 30 m grid cell. The analysis revealed higher SOM content in the northeastern and southwestern regions and lower content in the central area, ranging from 13.95 to 47.81 g/kg, with an average of 27.16 g/kg. Lime soil had the highest SOM content, while purple soil had the lowest. Paddy fields showed significantly higher SOM than dry land. Over the past 40 years, SOM content has slightly increased, while its spatial distribution has remained stable.</p></div>","PeriodicalId":51045,"journal":{"name":"European Journal of Agronomy","volume":"160 ","pages":"Article 127323"},"PeriodicalIF":4.5,"publicationDate":"2024-08-27","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"142088591","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":1,"RegionCategory":"农林科学","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":"","EPubDate":null,"PubModel":null,"JCR":null,"JCRName":null,"Score":null,"Total":0}
Pub Date : 2024-08-27DOI: 10.1016/j.eja.2024.127327
Yanxi Zhao, Zhihao Zhang, Yining Tang, Caili Guo, Xia Yao, Tao Cheng, Yan Zhu, Weixing Cao, Yongchao Tian
Accurate estimation of wheat maturity date (MD) is helpful to make reasonable harvest planning and guarantee crop yield and quality. In this study, wheat phenology extracted from satellite images was assimilated into WheatGrow model to develop wheat maturity date estimation model. Theoretical uncertainty was introduced into assimilation system as the error covariance matrix of remote sensing observations, which improved the performance of maturity date estimation model. Compared with the simulated maturity date of crop growth model and assimilation system combined with the constant uncertainty (Assimilation1), the accuracy of assimilation system combined with the theoretical uncertainty (Assimilation2) was higher (r = 0.81, RMSE = 4.5 d). Assimilation2 has better performance and robustness in different years and different subregions. The mean relative errors between the estimated values of Assimilation2 and the observations were generally small and concentrated in the range of −5 % to 5 %. The estimated maturity date showed latitude variation in spatial distribution in the Huang-Huai-Hai Plain (HHHP). In addition, the trend of wheat maturity date from 2001 to 2020 in the central region of HHHP was significant (p < 0.05), and the mean change rate of maturity date reached 3–6 d/10a. However, the overall change trend of maturity date in the HHHP was not significant. Temperature was main driver affecting the spatiotemporal variation of wheat maturity date. The regional wheat maturity date estimation model can provide technical support for wheat maturity date estimation at regional scale.
{"title":"Improving the estimation accuracy of wheat maturity date by coupling WheatGrow with satellite images","authors":"Yanxi Zhao, Zhihao Zhang, Yining Tang, Caili Guo, Xia Yao, Tao Cheng, Yan Zhu, Weixing Cao, Yongchao Tian","doi":"10.1016/j.eja.2024.127327","DOIUrl":"10.1016/j.eja.2024.127327","url":null,"abstract":"<div><p>Accurate estimation of wheat maturity date (MD) is helpful to make reasonable harvest planning and guarantee crop yield and quality. In this study, wheat phenology extracted from satellite images was assimilated into WheatGrow model to develop wheat maturity date estimation model. Theoretical uncertainty was introduced into assimilation system as the error covariance matrix of remote sensing observations, which improved the performance of maturity date estimation model. Compared with the simulated maturity date of crop growth model and assimilation system combined with the constant uncertainty (Assimilation1), the accuracy of assimilation system combined with the theoretical uncertainty (Assimilation2) was higher (r = 0.81, RMSE = 4.5 d). Assimilation2 has better performance and robustness in different years and different subregions. The mean relative errors between the estimated values of Assimilation2 and the observations were generally small and concentrated in the range of −5 % to 5 %. The estimated maturity date showed latitude variation in spatial distribution in the Huang-Huai-Hai Plain (HHHP). In addition, the trend of wheat maturity date from 2001 to 2020 in the central region of HHHP was significant (p < 0.05), and the mean change rate of maturity date reached 3–6 d/10a. However, the overall change trend of maturity date in the HHHP was not significant. Temperature was main driver affecting the spatiotemporal variation of wheat maturity date. The regional wheat maturity date estimation model can provide technical support for wheat maturity date estimation at regional scale.</p></div>","PeriodicalId":51045,"journal":{"name":"European Journal of Agronomy","volume":"160 ","pages":"Article 127327"},"PeriodicalIF":4.5,"publicationDate":"2024-08-27","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"142083932","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":1,"RegionCategory":"农林科学","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":"","EPubDate":null,"PubModel":null,"JCR":null,"JCRName":null,"Score":null,"Total":0}
Pub Date : 2024-08-21DOI: 10.1016/j.eja.2024.127292
Shuaibin Lai, Zhe Wu, Yang Liu, Fujiang Hou
Grazing is a major driving force of biodiversity, functions and stability of alpine meadows. In selenium-deficient alpine meadows, moderate selenium supplementation can promote plant growth and increase selenium content within the food web. However, the combined effect of grazing and selenium (Se) addition on the stability of alpine meadow communities within selenium-deficient soil is still unclear. Therefore, we conducted a three-year experiment in an alpine meadow of Qinghai-Tibetan Plateau (QTP), with two stocking rates (0 and 6 sheep months ha−1) and six levels of Se addition (0, 5, 10, 20, 40 and 80 g ha−1) to explore how grazing and Se addition affect community stability and its relationship with species richness, species asynchrony and functional group stability. Results showed that the community stability, species richness and biomass response ratio of the community increased gradually with the increase in the selenium application under grazing and enclosure treatments, reaching the maximum values at 20 g ha−1. However, when the selenium addition exceeded 20 g ha−1, the above mentioned indexes were decreased gradually, especially under grazing. Structural equation model showed that grazing and selenium addition indirectly affected the temporal stability of sedges and forbs, thus influencing the temporal stability of the community. Results of this study indicated that the alpine meadow can maintain high species diversity and community temporal stability under moderate grazing combined with selenium addition, providing a scientific basis for selenium-enriched grazing management in alpine meadows.
放牧是高山草甸生物多样性、功能和稳定性的主要驱动力。在缺硒的高山草甸上,适度补硒可以促进植物生长,增加食物网中的硒含量。然而,放牧和硒(Se)添加对缺硒土壤中高山草甸群落稳定性的综合影响尚不清楚。因此,我们在青藏高原高寒草甸进行了为期三年的试验,采用两种放牧率(0 和 6 个羊月/公顷)和六种 Se 添加水平(0、5、10、20、40 和 80 克/公顷),探讨放牧和 Se 添加如何影响群落稳定性及其与物种丰富度、物种异步性和功能群稳定性的关系。结果表明,在放牧和圈养处理下,群落稳定性、物种丰富度和生物量响应比随着施硒量的增加而逐渐增加,在20 g ha-1时达到最大值。然而,当施硒量超过 20 g ha-1 时,上述指标逐渐下降,尤其是在放牧条件下。结构方程模型表明,放牧和加硒间接影响了莎草和牧草的时间稳定性,从而影响了群落的时间稳定性。研究结果表明,高寒草甸在适度放牧与补硒相结合的条件下,能保持较高的物种多样性和群落时间稳定性,为高寒草甸的富硒放牧管理提供了科学依据。
{"title":"Grazing and selenium influence community stability by increasing asynchrony and sedge, forbs stability in alpine meadow of Qinghai-Tibet Plateau","authors":"Shuaibin Lai, Zhe Wu, Yang Liu, Fujiang Hou","doi":"10.1016/j.eja.2024.127292","DOIUrl":"10.1016/j.eja.2024.127292","url":null,"abstract":"<div><p>Grazing is a major driving force of biodiversity, functions and stability of alpine meadows. In selenium-deficient alpine meadows, moderate selenium supplementation can promote plant growth and increase selenium content within the food web. However, the combined effect of grazing and selenium (Se) addition on the stability of alpine meadow communities within selenium-deficient soil is still unclear. Therefore, we conducted a three-year experiment in an alpine meadow of Qinghai-Tibetan Plateau (QTP), with two stocking rates (0 and 6 sheep months ha<sup>−1</sup>) and six levels of Se addition (0, 5, 10, 20, 40 and 80 g ha<sup>−1</sup>) to explore how grazing and Se addition affect community stability and its relationship with species richness, species asynchrony and functional group stability. Results showed that the community stability, species richness and biomass response ratio of the community increased gradually with the increase in the selenium application under grazing and enclosure treatments, reaching the maximum values at 20 g ha<sup>−1</sup>. However, when the selenium addition exceeded 20 g ha<sup>−1</sup>, the above mentioned indexes were decreased gradually, especially under grazing. Structural equation model showed that grazing and selenium addition indirectly affected the temporal stability of sedges and forbs, thus influencing the temporal stability of the community. Results of this study indicated that the alpine meadow can maintain high species diversity and community temporal stability under moderate grazing combined with selenium addition, providing a scientific basis for selenium-enriched grazing management in alpine meadows.</p></div>","PeriodicalId":51045,"journal":{"name":"European Journal of Agronomy","volume":"160 ","pages":"Article 127292"},"PeriodicalIF":4.5,"publicationDate":"2024-08-21","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"142040737","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":1,"RegionCategory":"农林科学","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":"","EPubDate":null,"PubModel":null,"JCR":null,"JCRName":null,"Score":null,"Total":0}
Cover crops (CCs) are recognised as valuable for weed management, while fallow soil between cash crop seasons likely increases weed presence. Weeds may offer similar ecosystem services as CCs, although they pose a risk of seedbank buildup. This study evaluated the impact of two winter CC systems (3-year triticale cultivation, TRIT; and a 3-year succession of rye, clover, and mustard, RCM) compared to weedy fallow (WF) on weed seedbank size and composition in a 3-year ‘maize (Zea mays L.)–maize–soybean (Glycine max (L.) Merr)’ crop succession. After 3 years, seed density of spring/summer weeds reduced in all treatments, potentially stemming from herbicide use during cash crop seasons and tillage operations. Triticale had the lowest seedbank density (9,487 seeds m−²) and higher diversity (Shannon Index 6.9) compared to WF (28,543 seeds m-² and 4.1, respectively). Furthermore, stochastic analysis revealed a lower risk of enlarging weed seedbanks in TRIT compared to WF (for seed densities above 900 seeds m−2). Moreover, management practices (CCs, cash crop sowing, termination/harvest) synchronised with weed seed production and germination likely contributed to the decreasing seed density of species including Portulaca oleracea and Chenopodium album, which were reduced by 90 and 80 %, respectively, by the study’s end. Over three years, autumn/winter and indifferent weed seed densities increased 4.2 times more in WF and RCM (22,638 seeds m−²) than in TRIT. This may be due to the varying growth rates among CC species in RCM, whereas TRIT consistently established rapidly, potentially outcompeting weeds until termination. Fallow periods between cash crops may increase weed species linked to that season and future crop–weed interference in varied crop rotations. Introducing CCs can mitigate this effect, although the choice of CC species may influence the extent of the impact.
{"title":"Unveiling the impact of winter cover crops and weedy fallow on the soil seedbank","authors":"Giorgia Raimondi , Donato Loddo , Vittoria Giannini , Maurizio Borin","doi":"10.1016/j.eja.2024.127309","DOIUrl":"10.1016/j.eja.2024.127309","url":null,"abstract":"<div><p>Cover crops (CCs) are recognised as valuable for weed management, while fallow soil between cash crop seasons likely increases weed presence. Weeds may offer similar ecosystem services as CCs, although they pose a risk of seedbank buildup. This study evaluated the impact of two winter CC systems (3-year triticale cultivation, TRIT; and a 3-year succession of rye, clover, and mustard, RCM) compared to weedy fallow (WF) on weed seedbank size and composition in a 3-year ‘maize (<em>Zea mays</em> L.)–maize–soybean (<em>Glycine max</em> (L.) Merr)’ crop succession. After 3 years, seed density of spring/summer weeds reduced in all treatments, potentially stemming from herbicide use during cash crop seasons and tillage operations. Triticale had the lowest seedbank density (9,487 seeds m<sup>−</sup>²) and higher diversity (Shannon Index 6.9) compared to WF (28,543 seeds m<sup>-</sup>² and 4.1, respectively). Furthermore, stochastic analysis revealed a lower risk of enlarging weed seedbanks in TRIT compared to WF (for seed densities above 900 seeds m<sup>−2</sup>). Moreover, management practices (CCs, cash crop sowing, termination/harvest) synchronised with weed seed production and germination likely contributed to the decreasing seed density of species including <em>Portulaca oleracea</em> and <em>Chenopodium album</em>, which were reduced by 90 and 80 %, respectively, by the study’s end. Over three years, autumn/winter and indifferent weed seed densities increased 4.2 times more in WF and RCM (22,638 seeds m<sup>−</sup>²) than in TRIT. This may be due to the varying growth rates among CC species in RCM, whereas TRIT consistently established rapidly, potentially outcompeting weeds until termination. Fallow periods between cash crops may increase weed species linked to that season and future crop–weed interference in varied crop rotations. Introducing CCs can mitigate this effect, although the choice of CC species may influence the extent of the impact.</p></div>","PeriodicalId":51045,"journal":{"name":"European Journal of Agronomy","volume":"160 ","pages":"Article 127309"},"PeriodicalIF":4.5,"publicationDate":"2024-08-20","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"https://www.sciencedirect.com/science/article/pii/S1161030124002302/pdfft?md5=9fbad3f1cf86359a22a2ee6ff5654a10&pid=1-s2.0-S1161030124002302-main.pdf","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"142012361","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":1,"RegionCategory":"农林科学","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":"OA","EPubDate":null,"PubModel":null,"JCR":null,"JCRName":null,"Score":null,"Total":0}
Pub Date : 2024-08-17DOI: 10.1016/j.eja.2024.127318
Wenjia Yang , Jianli Zhou , Shiwen Wang , Lina Yin
Supplemental irrigation (SI) at critical growth stages is a practical way to mitigate water shortage of winter wheat in dryland. However, the yield performance after SI is unstable in different precipitation years. To obtain a better understanding about the applicability of SI in dryland, the utilization efficiency of precipitation and irrigation water (IUE) was divided into 5 sequential ratios: water storage ratio, water consumption ratio, water transpired ratio, transpiration efficiency and harvest index. In this study, those ratios were investigated under four SI methods: no irrigation (W0), SI once at jointing (Wj), SI once at booting (Wb) and SI twice at jointing and booting (Wj+b) throughout four years. Our results showed that water storage ratio was increased under all SI treatments, but water transpired ratio was only significantly increased under Wj and Wj+b, due to their greater development of plant population compared to W0. Thus, wheat yield was greatly improved by 7–12 % under Wj and Wj+b than that under W0 in dry and normal years. However, IUE was significantly decreased under Wj+b. Compared with W0, Wj+b had higher evaporation during the fallow period and lower water consumption ratio. Furthermore, 100–300 cm subsoil water utilization was decreased under Wj+b from jointing to harvest time due to the low root length density (RDL) in subsoil. Under Wj, evaporation, subsoil water utilization and RDL were not negatively affected and water use efficiency was increased compared to W0. Thus, SI once at jointing stage is a more suitable practice in dryland wheat farmland when considering the dual goal of effectively using water while increasing yields under worse precipitation year.
{"title":"Different supplemental irrigation methods result in discrepant water use efficiency and yield by changing steps of water use process in dryland wheat","authors":"Wenjia Yang , Jianli Zhou , Shiwen Wang , Lina Yin","doi":"10.1016/j.eja.2024.127318","DOIUrl":"10.1016/j.eja.2024.127318","url":null,"abstract":"<div><p>Supplemental irrigation (SI) at critical growth stages is a practical way to mitigate water shortage of winter wheat in dryland. However, the yield performance after SI is unstable in different precipitation years. To obtain a better understanding about the applicability of SI in dryland, the utilization efficiency of precipitation and irrigation water (IUE) was divided into 5 sequential ratios: water storage ratio, water consumption ratio, water transpired ratio, transpiration efficiency and harvest index. In this study, those ratios were investigated under four SI methods: no irrigation (W<sub>0</sub>), SI once at jointing (W<sub>j</sub>), SI once at booting (W<sub>b</sub>) and SI twice at jointing and booting (W<sub>j+b</sub>) throughout four years. Our results showed that water storage ratio was increased under all SI treatments, but water transpired ratio was only significantly increased under W<sub>j</sub> and W<sub>j+b</sub>, due to their greater development of plant population compared to W<sub>0</sub>. Thus, wheat yield was greatly improved by 7–12 % under W<sub>j</sub> and W<sub>j+b</sub> than that under W<sub>0</sub> in dry and normal years. However, IUE was significantly decreased under W<sub>j+b</sub>. Compared with W<sub>0</sub>, W<sub>j+b</sub> had higher evaporation during the fallow period and lower water consumption ratio. Furthermore, 100–300 cm subsoil water utilization was decreased under W<sub>j+b</sub> from jointing to harvest time due to the low root length density (RD<sub>L</sub>) in subsoil. Under W<sub>j</sub>, evaporation, subsoil water utilization and RD<sub>L</sub> were not negatively affected and water use efficiency was increased compared to W<sub>0</sub>. Thus, SI once at jointing stage is a more suitable practice in dryland wheat farmland when considering the dual goal of effectively using water while increasing yields under worse precipitation year.</p></div>","PeriodicalId":51045,"journal":{"name":"European Journal of Agronomy","volume":"160 ","pages":"Article 127318"},"PeriodicalIF":4.5,"publicationDate":"2024-08-17","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"141998418","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":1,"RegionCategory":"农林科学","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":"","EPubDate":null,"PubModel":null,"JCR":null,"JCRName":null,"Score":null,"Total":0}
Pub Date : 2024-08-15DOI: 10.1016/j.eja.2024.127317
Gang Zhao , Quanying Zhao , Heidi Webber , Andreas Johnen , Vittorio Rossi , Antonio Fernandes Nogueira Junior
Crop diseases are increasingly causing devastating yield losses each year, posing a significant threat to global food security. Currently, fungicide treatment is one of the most effective measures to mitigate these disease-induced yield losses. Accurately predicting disease occurrence, onset timing, and development is critical for optimizing fungicide application schedules to achieve high efficacy. In this study, we compiled weekly disease observations, crop management, and weather data from 207 locations across China, India, and Japan. We compared six machine learning (ML) models for their performance in simulating disease severity. By mimicking disease development curves and trends at early stages, we developed a novel change detection (CD) model, named rolling linear regression (RLR), and combined it with the best-performing ML model to predict the occurrence, severity, and onset dates of four major rice diseases: leaf blast (Magnaporthe oryzae), panicle-neck blast (Magnaporthe oryzae), sheath blight (Rhizoctonia solani), and false smut (Ustilaginoidea virens). Our findings showed that the RandomForestRegressor demonstrated the highest performance in simulating disease severity, with mean absolute errors (MAE) below 0.5 % and root mean square errors (RMSE) below 2.5 %. The RLR model showed a distinct advantage over the other four widely used CD models based on five evaluation metrics. Additionally, the paired RandomForestRegressor+RLR model achieved the highest F1-score, ranging from 0.7 to 0.8, in predicting disease occurrence, outperforming 29 other ML+CD model pairs. Furthermore, the RandomForestRegressor+RLR model predicted onset dates with fewer than 6 error days and accuracies ranging from 74 % to 87 %. Combining ML with CD models not only shows robust generalization across diverse environmental conditions but also proves highly effective for large-scale disease risk forecasting in rice farming regions. The adaptability of ML techniques, when sufficient training data are available, particularly enhances decision support systems aimed at optimizing rice disease management practices for growers across various regions. Our hybrid model thus presents a compelling advancement in the precision agriculture domain, with significant implications for improving disease management strategies for crops beyond rice through timely intervention. This approach can contribute to safeguarding global food security by reducing crop losses due to disease damage.
农作物病害每年都会造成越来越多的毁灭性减产,对全球粮食安全构成重大威胁。目前,杀菌剂处理是减轻这些病害造成的产量损失的最有效措施之一。准确预测病害的发生、发病时间和发展情况,对于优化杀菌剂施用计划以实现高效杀菌至关重要。在这项研究中,我们汇编了来自中国、印度和日本 207 个地点的每周病害观测数据、作物管理数据和天气数据。我们比较了六种机器学习(ML)模型在模拟病害严重程度方面的性能。通过模仿病害早期的发展曲线和趋势,我们开发了一种名为滚动线性回归(RLR)的新型变化检测(CD)模型,并将其与表现最佳的 ML 模型相结合,预测了四种主要水稻病害的发生、严重程度和发病日期:叶瘟(Magnaporthe oryzae)、穗颈瘟(Magnaporthe oryzae)、鞘枯病(Rhizoctonia solani)和假烟病(Ustilaginoidea virens)。我们的研究结果表明,随机森林回归模型(RandomForestRegressor)在模拟病害严重程度方面表现最佳,平均绝对误差(MAE)低于 0.5%,均方根误差(RMSE)低于 2.5%。根据五项评估指标,RLR 模型与其他四种广泛使用的 CD 模型相比具有明显优势。此外,RandomForestRegressor+RLR 配对模型在预测疾病发生率方面取得了最高的 F1 分数(从 0.7 到 0.8 不等),优于其他 29 个 ML+CD 配对模型。此外,RandomForestRegressor+RLR 模型预测发病日期的误差天数少于 6 天,准确率在 74% 到 87% 之间。将 ML 与 CD 模型相结合,不仅能在不同的环境条件下显示出强大的泛化能力,而且在水稻种植区的大规模病害风险预测中也被证明是非常有效的。当有足够的训练数据时,ML 技术的适应性尤其能增强决策支持系统,从而优化不同地区种植者的水稻病害管理实践。因此,我们的混合模型在精准农业领域取得了令人瞩目的进步,对通过及时干预改进水稻以外作物的病害管理策略具有重要意义。这种方法可以减少病害造成的作物损失,从而为保障全球粮食安全做出贡献。
{"title":"Integrating machine learning and change detection for enhanced crop disease forecasting in rice farming: A multi-regional study","authors":"Gang Zhao , Quanying Zhao , Heidi Webber , Andreas Johnen , Vittorio Rossi , Antonio Fernandes Nogueira Junior","doi":"10.1016/j.eja.2024.127317","DOIUrl":"10.1016/j.eja.2024.127317","url":null,"abstract":"<div><p>Crop diseases are increasingly causing devastating yield losses each year, posing a significant threat to global food security. Currently, fungicide treatment is one of the most effective measures to mitigate these disease-induced yield losses. Accurately predicting disease occurrence, onset timing, and development is critical for optimizing fungicide application schedules to achieve high efficacy. In this study, we compiled weekly disease observations, crop management, and weather data from 207 locations across China, India, and Japan. We compared six machine learning (ML) models for their performance in simulating disease severity. By mimicking disease development curves and trends at early stages, we developed a novel change detection (CD) model, named rolling linear regression (RLR), and combined it with the best-performing ML model to predict the occurrence, severity, and onset dates of four major rice diseases: leaf blast (<em>Magnaporthe oryzae</em>), panicle-neck blast (<em>Magnaporthe oryzae</em>), sheath blight (<em>Rhizoctonia solani</em>), and false smut (<em>Ustilaginoidea virens</em>). Our findings showed that the RandomForestRegressor demonstrated the highest performance in simulating disease severity, with mean absolute errors (<em>MAE</em>) below 0.5 % and root mean square errors (<em>RMSE</em>) below 2.5 %. The RLR model showed a distinct advantage over the other four widely used CD models based on five evaluation metrics. Additionally, the paired RandomForestRegressor+RLR model achieved the highest F1-score, ranging from 0.7 to 0.8, in predicting disease occurrence, outperforming 29 other ML+CD model pairs. Furthermore, the RandomForestRegressor+RLR model predicted onset dates with fewer than 6 error days and accuracies ranging from 74 % to 87 %. Combining ML with CD models not only shows robust generalization across diverse environmental conditions but also proves highly effective for large-scale disease risk forecasting in rice farming regions. The adaptability of ML techniques, when sufficient training data are available, particularly enhances decision support systems aimed at optimizing rice disease management practices for growers across various regions. Our hybrid model thus presents a compelling advancement in the precision agriculture domain, with significant implications for improving disease management strategies for crops beyond rice through timely intervention. This approach can contribute to safeguarding global food security by reducing crop losses due to disease damage.</p></div>","PeriodicalId":51045,"journal":{"name":"European Journal of Agronomy","volume":"160 ","pages":"Article 127317"},"PeriodicalIF":4.5,"publicationDate":"2024-08-15","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"141990797","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":1,"RegionCategory":"农林科学","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":"","EPubDate":null,"PubModel":null,"JCR":null,"JCRName":null,"Score":null,"Total":0}
Pub Date : 2024-08-12DOI: 10.1016/j.eja.2024.127316
Inés Gómez-Ramos , Manuel Caro , Juan A. López , David Ruiz , Jose A. Egea
One of the most important agroclimatic variables for stone fruit production is winter chill accumulation. To estimate chill accumulation in locations where climatic data is not recorded, spatial interpolation is necessary. In this study, we compare different interpolation methods for mean and Safe Winter Chill (SWC) in a Mediterranean stone fruit production area (Region of Murcia, SE Spain) using data from 49 climatic stations. To choose the most accurate interpolation method, as its choice may substantially influence the prediction accuracy, the predictive capability of several interpolation methods with different parameterizations (for a total number of 32 instances) was compared through out-of-bag bootstrap cross-validation, concluding that the best ones were Radial Basis Functions applied on the altitude-dependent regression residuals for mean winter chill and the altitude+latitude linear regression for SWC. The incorporation of altitude in the interpolation increased greatly the accuracy of the estimation. In fact, most of the chill accumulation spatial dependency was explained through altitude. The accuracy of the interpolation was not homogeneous across the study area. Chill accumulation in warmer coastal localities was overestimated by all the methods, possibly due to the proximity to the sea, highlighting the importance of microclimatic variables at higher-resolution spatial interpolations. Differences between methods were more notable in higher locations, where distance-only based methods underestimated chill accumulation and methods that consider altitude slightly overestimated it. This study demonstrates the importance of comparing the performance of multiple spatial interpolation methods before applying any for chill accumulation data.
{"title":"A comparison of interpolation methods to predict chill accumulation in a Mediterranean stone fruit production area (Región de Murcia, SE Spain)","authors":"Inés Gómez-Ramos , Manuel Caro , Juan A. López , David Ruiz , Jose A. Egea","doi":"10.1016/j.eja.2024.127316","DOIUrl":"10.1016/j.eja.2024.127316","url":null,"abstract":"<div><p>One of the most important agroclimatic variables for stone fruit production is winter chill accumulation. To estimate chill accumulation in locations where climatic data is not recorded, spatial interpolation is necessary. In this study, we compare different interpolation methods for mean and Safe Winter Chill (SWC) in a Mediterranean stone fruit production area (Region of Murcia, SE Spain) using data from 49 climatic stations. To choose the most accurate interpolation method, as its choice may substantially influence the prediction accuracy, the predictive capability of several interpolation methods with different parameterizations (for a total number of 32 instances) was compared through out-of-bag bootstrap cross-validation, concluding that the best ones were Radial Basis Functions applied on the altitude-dependent regression residuals for mean winter chill and the altitude+latitude linear regression for SWC. The incorporation of altitude in the interpolation increased greatly the accuracy of the estimation. In fact, most of the chill accumulation spatial dependency was explained through altitude. The accuracy of the interpolation was not homogeneous across the study area. Chill accumulation in warmer coastal localities was overestimated by all the methods, possibly due to the proximity to the sea, highlighting the importance of microclimatic variables at higher-resolution spatial interpolations. Differences between methods were more notable in higher locations, where distance-only based methods underestimated chill accumulation and methods that consider altitude slightly overestimated it. This study demonstrates the importance of comparing the performance of multiple spatial interpolation methods before applying any for chill accumulation data.</p></div>","PeriodicalId":51045,"journal":{"name":"European Journal of Agronomy","volume":"160 ","pages":"Article 127316"},"PeriodicalIF":4.5,"publicationDate":"2024-08-12","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"https://www.sciencedirect.com/science/article/pii/S1161030124002375/pdfft?md5=f74c3124f933d73bcd0da2f19386fbb0&pid=1-s2.0-S1161030124002375-main.pdf","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"141964431","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":1,"RegionCategory":"农林科学","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":"OA","EPubDate":null,"PubModel":null,"JCR":null,"JCRName":null,"Score":null,"Total":0}
Pub Date : 2024-08-10DOI: 10.1016/j.eja.2024.127297
Ali Raza, Yongguang Hu, Yongzong Lu
Tea plant (Camellia sinensis) is a major global crop consumed as a drink after water. Quantifying carbon flux, specifically the net ecosystem exchange (NEE), in tea plantations is essential for determining carbon sequestration and ecosystem carbon balance. The Eddy covariance (EC) system is widely used for continuous monitoring of carbon flux but high costs associated with installation and maintenance limit its widespread adoption. In addition, EC flux data is often discarded due to malfunction of instruments caused by adverse weather conditions. Therefore, additional approaches for estimating NEE are necessary to overcome these challenges and ensure accurate NEE measurement. For this purpose, three standalone tree-based machine learning (ML) models were used for NEE estimation using EC flux data collected from tea ecosystem located in subtropical region (Danyang county of Zhenjiang city) of China. To address the accuracy limitations inherent in standalone ML models, the ensemble mechanism based on voting regressor method was proposed. In addition, k-fold cross-validation based on early stopping process was also used to enhance the performance of standalone ML models. Based on visual plots (scatter diagram, heatMap, Taylor diagram) and performing indices (root-mean-square error (RMSE), determination coefficient (R2), mean absolute error (MAE), mean absolute percentage error (MAPE), Nash–Sutcliffe efficiency (NSE), correlation coefficient (r), Kling Gupta Efficiency (KGE) and index of agreement (d)), the findings indicated that non-linear ensemble-generalized regression neural network (NLE-GRNN) significantly improved standalone ML model's results. In current study, the highest NSE, r and d in case of standalone ML model (DT) achieved 0.49, 0.73 and 0.75 respectively while our proposed NLE-GRNN model improved 48 % in NSE value (NSE = 0.97), 25 % in r value (r = 0.98) and 24 % in d value (d = 0.99). Likewise, NLE-GRNN significantly reduce errors (MAE, MAPE and RMSE) and provides NEE estimate closet to the observed value. The impact of climatic variables on NEE using shapley additive explanations (SHAP) analysis revealed that Rg (solar radiation) and Tair (air temperature) were the prime factors controlling NEE variation in the tea ecosystem. Considering the high accuracy and stability of the studied ML models, it is recommended to apply developed ensemble ML model (NLE-GRNN) for significant improvement of NEE estimate in the tea biomes or other ecosystems.
茶树(Camellia sinensis)是全球主要的农作物,是仅次于水的饮料。量化茶园中的碳通量,特别是生态系统净交换量(NEE),对于确定碳固存和生态系统碳平衡至关重要。涡度协方差(EC)系统被广泛用于连续监测碳通量,但其安装和维护成本高昂,限制了其广泛应用。此外,由于恶劣天气条件导致仪器故障,EC 通量数据经常被丢弃。因此,有必要采用其他方法估算净能效,以克服这些挑战并确保准确测量净能效。为此,研究人员利用从中国亚热带地区(镇江市丹阳县)茶叶生态系统收集到的欧共体通量数据,使用了三种基于树的独立机器学习(ML)模型来估算净能效。针对独立 ML 模型固有的精度限制,提出了基于投票回归方法的集合机制。此外,还使用了基于早期停止过程的 k 倍交叉验证来提高独立 ML 模型的性能。根据直观图(散点图、热图、泰勒图)和性能指标(均方根误差(RMSE)、判定系数(R2)、平均绝对误差(MAE)、平均绝对百分比误差(MAPE)、纳什-苏特克利夫效率(NSE)、研究结果表明,非线性集合-广义回归神经网络(NLE-GRNN)显著改善了独立 ML 模型的结果。在当前研究中,独立 ML 模型(DT)的最高 NSE、r 和 d 分别为 0.49、0.73 和 0.75,而我们提出的 NLE-GRNN 模型的 NSE 值(NSE = 0.97)提高了 48%,r 值(r = 0.98)提高了 25%,d 值(d = 0.99)提高了 24%。同样,NLE-GRNN 显著减少了误差(MAE、MAPE 和 RMSE),并提供了接近观测值的 NEE 估计值。使用夏普利加法解释(SHAP)分析气候变量对 NEE 的影响,发现 Rg(太阳辐射)和 Tair(气温)是控制茶叶生态系统 NEE 变化的主要因素。考虑到所研究的 ML 模型的高精度和稳定性,建议应用所开发的集合 ML 模型(NLE-GRNN),以显著改善茶叶生物群落或其他生态系统中的 NEE 估算。
{"title":"Improving carbon flux estimation in tea plantation ecosystems: A machine learning ensemble approach","authors":"Ali Raza, Yongguang Hu, Yongzong Lu","doi":"10.1016/j.eja.2024.127297","DOIUrl":"10.1016/j.eja.2024.127297","url":null,"abstract":"<div><p>Tea plant (<em>Camellia sinensis</em>) is a major global crop consumed as a drink after water. Quantifying carbon flux, specifically the net ecosystem exchange (NEE), in tea plantations is essential for determining carbon sequestration and ecosystem carbon balance. The Eddy covariance (EC) system is widely used for continuous monitoring of carbon flux but high costs associated with installation and maintenance limit its widespread adoption. In addition, EC flux data is often discarded due to malfunction of instruments caused by adverse weather conditions. Therefore, additional approaches for estimating NEE are necessary to overcome these challenges and ensure accurate NEE measurement. For this purpose, three standalone tree-based machine learning (ML) models were used for NEE estimation using EC flux data collected from tea ecosystem located in subtropical region (Danyang county of Zhenjiang city) of China. To address the accuracy limitations inherent in standalone ML models, the ensemble mechanism based on voting regressor method was proposed. In addition, k-fold cross-validation based on early stopping process was also used to enhance the performance of standalone ML models. Based on visual plots (scatter diagram, heatMap, Taylor diagram) and performing indices (root-mean-square error (RMSE), determination coefficient (R<sup>2</sup>), mean absolute error (MAE), mean absolute percentage error (MAPE), Nash–Sutcliffe efficiency (NSE), correlation coefficient (r), Kling Gupta Efficiency (KGE) and index of agreement (d)), the findings indicated that non-linear ensemble-generalized regression neural network (NLE-GRNN) significantly improved standalone ML model's results. In current study, the highest NSE, r and d in case of standalone ML model (DT) achieved 0.49, 0.73 and 0.75 respectively while our proposed NLE-GRNN model improved 48 % in NSE value (NSE = 0.97), 25 % in r value (r = 0.98) and 24 % in d value (d = 0.99). Likewise, NLE-GRNN significantly reduce errors (MAE, MAPE and RMSE) and provides NEE estimate closet to the observed value. The impact of climatic variables on NEE using shapley additive explanations (SHAP) analysis revealed that Rg (solar radiation) and Tair (air temperature) were the prime factors controlling NEE variation in the tea ecosystem. Considering the high accuracy and stability of the studied ML models, it is recommended to apply developed ensemble ML model (NLE-GRNN) for significant improvement of NEE estimate in the tea biomes or other ecosystems.</p></div>","PeriodicalId":51045,"journal":{"name":"European Journal of Agronomy","volume":"160 ","pages":"Article 127297"},"PeriodicalIF":4.5,"publicationDate":"2024-08-10","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"141953738","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":1,"RegionCategory":"农林科学","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":"","EPubDate":null,"PubModel":null,"JCR":null,"JCRName":null,"Score":null,"Total":0}