Abstract:The eastern Helan Mountains are one of the most important plantings and production bases for high-quality grapes in China. In winter, grape root is prone to freezing damage owing to natural phenomenon such as early and late frost, and methods to assess disaster risk at the grape parcel level is one of the current research hotspots. However, there are two problems at present; first is the lack of grape parcel maps and second is that scale of disaster risk assessment is limited to the level of administrative divisions and grid systems. Therefore, according to the perennial characteristics of the grapes and clear texture of the open-air grape parcels based on public data sets (Sentinel-2A/B) and high-quality winery address information, combined with Google Earth engine (GEE) platform, Google Earth, and Ovitalmap, this study designs a collection scheme of open-air grape parcel samples and non-grape parcel samples, analyzes the environmental characteristics of the distribution of high-quality wineries, constructs a classification feature database in three aspects (time, terrain, and vegetation), uses the random forest classification method to classify open-air grape parcels in the eastern Helan Mountains in 2019, and conducts a late frost disaster risk assessment of the open-air grape parcel. The results show that: (1) The high-quality wineries are mainly distributed in the elevation of 1106–1240 m, with the average slope of 0.82°, the organic matter content of 10.79–14.27 g/kg, the P content of 24.87–37.58 mg/kg, the K content of 116–166 mg/kg, the pH 8.38–8.55, N content of 0.65–0.98 g/kg, annual average active accumulated temperature ≥10 °C of 3490–3590 °C, annual average rainfall of 189–201 mm, the annual average sunshine duration of 7.3–7.6 h, annual average air temperature of 8.8–9.4 °C, annual average temperature difference of 12.7–13.2 °C, annual average frost-free period at 263–270 d and annual average relative air Humidity of 50–54%. (2) The overall classification accuracy of open-air grape parcels reached 98.96% and the verification accuracy of ground truth parcel samples reached 87.89%. The open-air grape parcels in the eastern Helan Mountains are mainly distributed in Helan County, Xixia District, Yongning County, and Qingtongxia City. (3) The highest risk of late frost was found in Xixia District and the lowest risk in Yongning County. Helan County was higher than Qingtongxia City in the medium risk of late frost. The typical small area high risk includes Liangtian Town, the experimental field of Ningxia Academy of Agricultural and Forestry Sciences in Zhenbeibao Town, Yuhuang Winery in Ganchengzi Village, and the State-run Warm Spring Farm in Helan County. Thus, the sample collection and augmentation scheme are feasible, the open-air grape classification accuracy is high and its result is consistent with field investigation, which can provide accurate parcel-level disaster risk assessment and provide decision support for agricultural departments and farmers.

NYUS.2: an Automated Machine Learning Prediction Model for the Large-scale Real-time Simulation of Grapevine Freezing Tolerance in North America

Grape Cold Hardiness Prediction via Multi-Task Learning

GrapevineXL reliably predicts multi-annual dynamics of vine water status, berry growth, and sugar accumulation in vineyards

Integrating Hyperspectral, Thermal, and Ground Data with Machine Learning Algorithms Enhances the Prediction of Grapevine Yield and Berry Composition

Ecophysiological Modeling of Grapevine Water Stress in Burgundy Terroirs by a Machine-Learning Approach

Forecasting severe grape downy mildew attacks using machine learning

Grapevine Disease Prediction Using Climate Variables from Multi-Sensor Remote Sensing Imagery via a Transformer Model

Multi-Task Learning for Budbreak Prediction

Predicting Grapevine Physiological Parameters Using Hyperspectral Remote Sensing Integrated with Hybrid Convolutional Neural Network and Ensemble Stacked Regression

Developing a simple and efficient modeling solution for predicting key phenological stages of table grapes in a non-traditional viticulture zone in south Asia

Historical long-term cultivar×climate suitability data to inform viticultural adaptation to climate change

Deep learning based genomic breeding of pest-resistant grapevine

Combination of an Automated 3D Field Phenotyping Workflow and Predictive Modelling for High-Throughput and Non-Invasive Phenotyping of Grape Bunches

A long short‐term memory‐based model for greenhouse climate prediction

Non-destructive estimation for Kyoho grape shelf-life using Vis/NIR hyperspectral imaging and deep learning algorithm

Grapevine Rootstock and Scion Genotypes’ Symbiosis with Soil Microbiome: A Machine Learning Revelation for Climate-Resilient Viticulture

Near-infrared real time non-destructive monitoring of sugar accumulation reveals that single berries ripen two times faster than previously documented on standard asynchronous samples

Prediction of Grape Sap Flow in a Greenhouse Based on Random Forest and Partial Least Squares Models

Hyperspectral imaging and artificial intelligence enhance remote phenotyping of grapevine rootstock influence on whole vine photosynthesis

Using Reflectance and Gray-Level Texture for Water Content Prediction in Grape Vines

Open-air grape classification and its application in parcel-level risk assessment of late frost in the eastern Helan Mountains