Statistical modeling for the monthly estimate of chilling hours and units in Plateau regions of Rio Grande do Sul and Santa Catarina

ABSTRACT The objective of this study was to obtain and evaluate models for the monthly estimation of chilling hours/units based on air temperature data in the Plateau regions of Rio Grande do Sul and Santa Catarina. Data of air temperature were used from May to August 2012 to 2017 of seven meteorological stations. For each month the chilling hours/unit were calculated by the Utah method and the sum of temperatures lower or equal to 7.2 °C that were correlated with the minimum and average monthly temperature with the use of the Regression Analysis. Six statistical methods were used to evaluate the monthly estimate of chilling hours/units obtained by previously obtained models. The minimum temperature and the monthly average explained between 65% and 96% of the chilling hours by the temperature threshold ≤ 7.2 °C and the chilling units by the Utah method. The minimum temperature and the monthly air average are recommended for the estimation of chilling hours and chilling units, respectively. The climatic variability of the winter in the study region results in different performances of the models of estimates of chilling hours/unit.摘要（ABSTRACT） 本研究旨在获取并评估基于南里奥格兰德（Rio Grande do Sul）与圣卡塔琳娜（Santa Catarina）高原地区气温数据的低温时长（chilling hours）/低温单位（chilling units）月度估算模型。研究采用了2012至2017年每年5月至8月间7个气象站点的气温观测数据。针对每个月份，分别通过犹他法（Utah method）计算低温单位，并统计温度≤7.2℃的累积时长；随后借助回归分析将上述指标与月度最低气温、平均气温进行相关性分析。本研究采用6种统计方法，对此前构建的低温时长/单位月度估算模型开展评估。结果表明，月度最低气温可解释温度阈值≤7.2℃对应的低温时长65%至96%的变异，月度平均气温则可解释犹他法下低温单位65%至96%的变异。研究建议分别采用月度最低气温与平均气温，来估算低温时长与低温单位。研究区域冬季的气候变异性会导致低温时长/单位估算模型的表现存在差异。