ABSTRACT
Several recent studies have evidenced the relevance of machine-learning for soil salinity mapping using Sentinel-2 reflectance as input data and field soil salinity measurement (i.e., Electrical Conductivity-EC) as the target. As soil EC monitoring is costly and time consuming, most learning databases used for training/validation rely on a limited number of soil samples, which can affect the model consistency. Based on the low soil salinity variation at the Sentinel-2 pixel resolution, this study proposes to increase the learning database's number of observations by assigning the EC value obtained on the sampled pixel to the eight neighboring pixels. The method allowed extending the original learning database made up of 97 field EC measurements (OD) to an enhanced learning database made up of 691 observations (ED). Two classification machine-learning models (i.e., Random Forest-RF and Support Vector Machine-SVM) were trained with both OD and ED to assess the efficiency of the proposed method by comparing the models' outcomes with EC observations not used in the models´ training. The use of ED led to a significant increase in both models' consistency with the overall accuracy of the RF (SVM) model increasing from 0.25 (0.26) when using the OD to 0.77 (0.55) when using ED. This corresponds to an improvement of approximately 208% and 111%, respectively. Besides the improved accuracy reached with the ED database, the results showed that the RF model provided better soil salinity estimations than the SVM model and that feature selection (i.e., Variance Inflation Factor-VIF and/or Genetic Algorithm-GA) increase both models´ reliability, with GA being the most efficient. This study highlights the potential of machine-learning and Sentinel-2 image combination for soil salinity monitoring in a data-scarce context, and shows the importance of both model and features selection for an optimum machine-learning set-up.
ABSTRACT
During the austral winter (June-August) of 2021, the meteorological services of Brazil, Argentina, Peru, Paraguay, Bolivia, and Chile all issued forecasts for unusually cold conditions. Record-low minimum temperatures and cold spells were documented, including one strong cold wave episode that affected 5 countries. In this study, we define a cold wave as a period in which daily maximum and minimum air temperatures are below the corresponding climatological 10th percentile for three or more consecutive days. The intense cold wave event in the last week of June, 2021, resulted in record-breaking minimum daily temperatures in several places in central South America and Chile. Several locations had temperatures about 10 °C below average, central South America had freezing conditions, and southern Brazil even saw snow. The cold air surge was characterized by an intense upper-air trough located close to 35° S and 70° W. The southerly flow to the west of this trough brought very cold air northward into subtropical and tropical South America. A northward flow between the lower-level cyclonic and anticyclonic perturbations caused the intense southerly flow between the upper-level ridge and trough. This condition facilitated the inflow of near-surface cold air from southern Argentina into southeastern Brazil and tropical South America east of the Andes. In the city of São Paulo, the cold wave caused the death of 13 homeless people from hypothermia. Frost and snow across southern and southeastern Brazil caused significant damage to coffee, sugarcane, oranges, grapes, and other fruit and vegetable crops. Wine and coffee production fell, the latter by 30%, and prices of food and commodities in the region rose. Supplementary Information: The online version contains supplementary material available at 10.1007/s00382-023-06701-1.
