Use of artificial neural networks to rescue agrochemical-based health hazards: A resource optimisation method for cleaner crop production
Use of artificial neural networks to rescue agrochemical-based health hazards: A resource optimisation method for cleaner crop production
复制标题
DOI：
10.1016/j.jclepro.2019.117900
复制
发表时间：
2019-11-20
影响因子：
11.1
通讯作者：
Abid, Muhammad
中科院分区：
文献类型：
--
作者：
Elahi, Ehsan;Weijun, Cui;Abid, Muhammad
该研究的主要目的是估计农用化学品使用的目标值及其对作物生产力和人类健康的影响。还评估了农药施用过程中采取的保护措施及其对人类健康的影响。为了实现研究目标，使用标准化问卷，收集了2017年9月至10月巴基斯坦旁遮普邦哈菲扎巴德和谢胡普拉地区480名稻农的横截面数据。使用各种计量经济学方法进行数据分析。人工神经网络（ANN）发现研究区存在滥用农药的现象，建议在水稻产量一定的情况下，纯氮和农药的施用量分别减少43.6%和52.6%，而纯磷、纯钾和农家肥（FYM）的施用量需要分别增加67.6%、15%和21%，才能获得高效生产水平。柯布-道格拉斯 (CD) 生产函数估计了纯磷、纯钾、FYM、教育和耕作经验对水稻产量的积极且显着的影响。研究结果还表明，滥用农药不仅会影响水稻产量，还会影响田间工人的健康。泊松回归发现，眼睛刺激、头晕、咳嗽、恶心的发生与化学品使用显着相关。逻辑回归发现，在采取各种保护措施时，教育、农耕经验和推广服务显着增加。此外，在施用农药期间采取防护措施的人中，初级健康暴露的病例明显减少。毒物回归函数的结果还证实，喷洒农药时使用防护服、护目镜、口罩、手套和靴子可显着减少人体健康暴露。建议使用 ANN 方法建议的营养素和化学物质的用量。使用生物化学品是一种更可持续、更环保的策略。需要向农民提供有关安全使用农药的教育计划和培训，以避免职业健康暴露。此外，政府需要采取行动，例如限制和/或禁止使用剧毒农药以及在农药施用期间强制采取安全措施，以减少农药暴露。 (C) 2019 Elsevier Ltd. 保留所有权利。
The main aim of the study was to estimate the target values of agrochemical use, and its impact on crop productivity, and human health. Adoptions of protective measures and their impacts on human health during pesticide application were also evaluated. To fulfil the study objectives, using a standardised questionnaire, cross-sectional data of 480 rice growers were collected from September to October, 2017 from the Hafizabad and Sheikhupura districts of Punjab, Pakistan. Various econometric methods were used for data analysis. The artificial neural network (ANN) found an indiscriminate use of agrochemicals in the study area and suggested reducing the applied quantity of Pure N and pesticides by 43.6 and 52.6%, respectively, at a given level of rice yield, while the quantity of Pure P, Pure K, and farmyard manure (FYM) need to increase by 67.6, 15, and 21%, respectively, to obtain an efficient production level. The Cobb-Douglas (CD) production function estimated a positive and significant impact of Pure P, Pure K, FYM, education, and farming experience on rice yield. Results also showed that an indiscriminate use of pesticide not only compromised rice efficiency, it impinged on field worker health. Poisson regression found that occurrence of eye irritation, dizziness, cough, and nausea were significantly related with chemical application. Logistic regression found that education, farming experience, and extension services significantly increased in the adoption of various protective measures. Furthermore, cases of primary health exposures were significantly less among those who were used protective measures during pesticide application. Results of Poison regression function also confirmed that use of protective clothing, goggles, mask, gloves and boots during pesticide spraying significantly reduced human health exposures. It is recommended to apply the quantity of nutrients and chemicals suggested by ANN method. Use of biochemicals is a more sustainable and environmentally friendly strategy. Provision of educational programs and trainings to farmers on safe use of pesticides is required to avert occupational health exposures. Furthermore, governmental actions such as restriction and/or interdiction on use of highly toxic pesticides and enforcement to adopt safety measures during pesticide application, are needed to reduce pesticide exposures. (C) 2019 Elsevier Ltd. All rights reserved.