Spatiotemporal analysis of meteorological drought across China based on the high-spatial-resolution multiscale SPI generated by machine learning

Weather and Climate Extremes, Journal Year: 2023, Volume and Issue: 40, P. 100567 - 100567

Published: May 6, 2023

An accurate gridded Standardized Precipitation Index (SPI) at high spatial resolution is important for meteorological drought monitoring and assessment. However, the estimation of SPI solely derived from precipitation products or interpolation station-based data subject to low accuracy large uncertainty. Here, we developed a Gaussian process regression-based method by blending multiple (GPR-BMP) with information stations topographical characteristics improve SPI. GPR-BMP was applied over China in this study, using five widely used station data, as well elevation, slope, longitude latitude. The multiscale (i.e., 1-month, 3-month, 6-month, 9-month 12-month) datasets 1984–2020 (GPR-BMP-SPI) were generated. Results showed that GPR-BMI-SPIs have higher than SPIs single all time scales. Additionally, GPR-BMP-SPIs can relatively accurately identify historical events. By analysing spatiotemporal variation during 1984–2020, it found annual frequency, severity duration southeast 1984–2020. trends significant decrease on Qinghai-Tibet Plateau. Winter exhibited most distinctive features between east west. area percentage an overall decreasing trend whole region both seasonal It also frequency contributed more duration. GPR-BMP-SPI generated study serve fundamental support future studies. proposed model be other regions globally mitigation.

Language: Английский

---

Exploring dynamic response of agrometeorological droughts towards winter wheat yield loss risk using machine learning approach at a regional scale in Pakistan

Field Crops Research, Journal Year: 2023, Volume and Issue: 302, P. 109057 - 109057

Published: July 25, 2023

Language: Английский

---

Short-term drought Index forecasting for hot and semi-humid climate Regions: A novel empirical Fourier decomposition-based ensemble Deep-Random vector functional link strategy

Computers and Electronics in Agriculture, Journal Year: 2024, Volume and Issue: 217, P. 108609 - 108609

Published: Jan. 11, 2024

Language: Английский

---

Machine Learning Application in Horticulture and Prospects for Predicting Fresh Produce Losses and Waste: A Review

Plants, Journal Year: 2024, Volume and Issue: 13(9), P. 1200 - 1200

Published: April 25, 2024

The current review examines the state of knowledge and research on machine learning (ML) applications in horticultural production potential for predicting fresh produce losses waste. Recently, ML has been increasingly applied horticulture efficient accurate operations. Given health benefits need food nutrition security, postharvest management are important. This aims to assess application preharvest reducing waste by their magnitude, which is crucial practices policymaking loss reduction. starts assessing horticulture. It then presents handling processing, lastly, prospects its quantification. findings revealed that several algorithms perform satisfactorily classification prediction tasks. Based that, there a further investigate suitability more models or combination with higher prediction. Overall, suggested possible future directions related

Language: Английский

---

Impacts of climate change on spatial wheat yield and nutritional values using hybrid machine learning

Environmental Research Letters, Journal Year: 2024, Volume and Issue: 19(10), P. 104049 - 104049

Published: Aug. 30, 2024

Abstract Wheat’s nutritional value is critical for human nutrition and food security. However, more attention needed, particularly regarding the content concentration of iron (Fe) zinc (Zn), especially in context climate change (CC) impacts. To address this, various controlled field experiments were conducted, involving cultivation three wheat cultivars over growing seasons at multiple locations with different soil conditions under varying Fe Zn treatments. The yield attributes, including values such as nitrogen (N), Zn, from these integrated national statistics other to train test machine learning (ML) algorithms. Automated ML leveraging a large number models, outperformed traditional enabling training testing numerous achieving robust predictions grain (GY) ( R 2 > 0.78), N 0.75), 0.71) through stacked ensemble all models. model predicted GY, N, Fe, spatial explicit mid-century (2020–2050) using Global Circulation Models (GCMs): GFDL-ESM4, HadGEM3-GC31-MM, MRI-ESM2-0 two shared socioeconomic pathways (SSPs) specifically SSP2-45 SSP5-85, downscaled NEX-GDDP-CMIP6. Averaged across GCMs SSPs, CC projected increase by 4.5%, protein 0.8% high variability. it expected decrease 5.5%, 4.5% relative historical period (1980–2010). Positive impacts on encountered negative concentrations, further exacerbating challenges related security nutrition.

Language: Английский

---

Analysis of Wheat-Yield Prediction Using Machine Learning Models under Climate Change Scenarios

Sustainability, Journal Year: 2024, Volume and Issue: 16(16), P. 6976 - 6976

Published: Aug. 14, 2024

Climate change has emerged as one of the most significant challenges in modern agriculture, with potential implications for global food security. The impact changing climatic conditions on crop yield, particularly staple crops like wheat, raised concerns about future production. By integrating historical climate data, GCM (CMIP3) projections, and wheat-yield records, our analysis aims to provide insights into how may affect wheat output. This research uses advanced machine learning models explore intricate relationship between prediction. Machine used include multiple linear regression (MLR), boosted tree, random forest, ensemble models, several types ANNs: ANN (multi-layer perceptron), (probabilistic neural network), (generalized feed-forward), (linear regression). model was evaluated validated against yield weather data from three Punjab, Pakistan, regions (1991–2021). calibrated response downscaled (GCM) outputs SRA2, B1, A1B average collective CO2 emissions scenarios anticipate changes through 2052. Results showed that maximum temperature (R = 0.116) primary factor affecting preceding Tmin 0.114), while rainfall had a negligible 0.000). 0.988, nRMSE= 8.0%, MAE 0.090) demonstrated outstanding performance, outperforming Random Forest Regression 0.909, nRMSE 18%, 0.182), ANN(MLP) 0.902, 0.238, 17.0%), boosting tree 20%, 0.198). ANN(PNN) performed inadequately. RF better results R2 0.953, 0.791. expected is 5.5% lower than greatest reported at site study predicts site-specific output will experience loss due change. decrease, which anticipated be highest ever recorded, points might worsen insecurity. Additionally, findings highlighted approaches leveraging strengths could offer more accurate reliable predictions under varying scenarios. suggests developing climate-resilient agricultural practices, paving way sustainable security solutions.

Language: Английский

---

Dynamics of Tree outside Forest Land Cover Development and Ecosystem Carbon Storage Change in Eastern Coastal Zone, Bangladesh

Land, Journal Year: 2022, Volume and Issue: 11(1), P. 76 - 76

Published: Jan. 4, 2022

Tree outside forest (TOF) has immense potential in economic and environmental development by increasing the amount of tree vegetation around rural settlements. It is an important source carbon stocks a critical option for climate change regulation, especially land-scarce, densely populated developing countries such as Bangladesh. Spatio-temporal changes TOF eastern coastal zone Bangladesh were analyzed mapped over 1988–2018, using Landsat land use cover (LULC) maps associated ecosystem storage linking InVEST model. TM OLI-TIRS data classified through Maximum Likelihood Classifier (MLC) algorithm Semi-Automated Classification (SAC). In model, aboveground, belowground, dead organic matter, soil densities different LULC types used. The findings revealed that studied landscapes have differential features changing trends where TOF, mangrove forest, built-up land, salt-aquaculture increased due to loss agricultural mudflats, water bodies, hill vegetation. Among biomes, experienced largest increase (1453.9 km2), it also 9.01 Tg C. However, decreased rapidly 1285.8 km2 365.7 reduced 3.09 C 4.89 C, respectively. total regional 1.27 during 1988–2018. addition anthropogenic drivers, erosion accretion observed significantly alter storage, necessitating effective river channel embankment management minimize food security tradeoff landscape.

Language: Английский

---

Development of a new hybrid ensemble method for accurate characterization of future drought using multiple global climate models

Stochastic Environmental Research and Risk Assessment, Journal Year: 2023, Volume and Issue: 37(12), P. 4567 - 4587

Published: July 27, 2023

Language: Английский

---

The effects of iron-based nanomaterials (Fe NMs) on plants under stressful environments: Machine learning-assisted meta-analysis

Journal of Environmental Management, Journal Year: 2024, Volume and Issue: 354, P. 120406 - 120406

Published: Feb. 18, 2024

Language: Английский

---

Performances of different yield-detrending methods in assessing the impacts of agricultural drought and flooding: A case study in the middle-and-lower reach of the Yangtze River, China

Agricultural Water Management, Journal Year: 2024, Volume and Issue: 296, P. 108812 - 108812

Published: April 4, 2024

In evaluating the impacts of drought and flooding disasters on crop yields, accurately calculating meteorological yield (i.e., detrended yield) is an important procedure. The present work aimed to compare various yield-detrending methods in terms characterizing regression relationships between intensities flooding. Taking middle-and-lower reach Yangtze River (MLRYR) as study region, during growing seasons four crops (cotton, oilseed rape, wheat, maize) were quantified using standardized precipitation evapotranspiration index. Nine popular first-difference method employed determine yields. results indicated that examined consistently identified cases with very significant yield-reducing 58 relationships, outperforming (identifying 39–44 relationships). 20-yr moving average linear fitting performed best at provincial level, while cubic smoothing spline quadratic polynomial district level. Based these best-performing methods, was proposed exhibited wider applicability better performance than individual methods. losses, cotton most affected (35% all districts experienced severe loss) Anhui region (with loss 14.06%). impact 41 districts; comparison, only 7 districts. Additionally, rape by These can provide guidance for assessing agricultural under climate change.

Language: Английский

---