The annual growth and the thickness of cork are known to be highly variable between trees located in the same geographical location. Researching how climate variables affect different trees within the same site is a step forward for the management of cork production since current knowledge focusses only on the average tree response. Quantile regression methodology was applied for the first time to a large data set containing measurements of cork growth, sampled in 35 stands across the cork oak distribution area in Portugal. This methodology proved to be useful for testing the hypothesis raised: does climate affect differently the annual cork growth, and ultimately cork thickness of individual trees located in the same stand? Estimating the amount of cork produced by one stand that has the required thickness for the production of natural cork stoppers is essential to support cork oak management. However, no model, before this work, had been developed to provide managers with this information. A downward parabolic relationship between annual cork growth and annual precipitation was determined for all quantiles, with optimum annual average precipitation value ranging from 1103 to 1007 mm. April to August monthly temperatures, spring average temperature or summer average temperature, showed a negative relationship with annual cork growth, in particular for lower quantiles. Maximum annual temperature was shown to negatively affect annual cork thickness, in particular for the trees under the 6th quantile. The ratio between annual precipitation and average temperature, that define the Lang index (LI), showed a downward parabolic relationship with annual cork growth. Best cork growth conditions are found for Lang index values around 60, corresponding for the transition between semi-arid climate and humid climate. The application of the final model developed for estimating cork thickness of an eight years’ cork growth period allowed the prediction and mapping of the percentage of cork suitable for natural cork stopper production. It showed that higher values are expected in the Southern and Central coastal regions and along the Tagus River basin. The Northern coastal and mountain regions, characterised by Lang index values higher to 60 (humid climates), present lower estimated values for the percentage of cork suitable for natural cork stopper production. The estimated values are expected to be reduced under climate change scenarios in the Southern and Central coastal regions.

众所周知，在同一地理位置的树木之间，软木的年生长量和厚度是高度可变的。研究气候变量如何影响同一地点的不同树木，这对软木塞生产的管理迈出了一步，因为目前的知识仅集中在平均树木反应上。分位数回归方法首次应用于包含软木塞生长量度的大型数据集，该数据集在葡萄牙软木橡树分布区域的35个林分中采样。事实证明，这种方法对于检验提出的假设是有用的：气候对软木塞的年增长率有不同的影响，最终使位于同一林分中的单个树木的软木厚度增加？估算一个具有生产天然软木塞所需厚度的架子的软木产量，对于支持软木橡木管理至关重要。但是，在进行这项工作之前，还没有模型可以为管理人员提供这些信息。确定了所有分位数的年软木塞生长与年降水量之间的下降抛物线关系，最佳年平均降水量范围为1103至1007 mm。4月至8月的月度温度，春季平均温度或夏季平均温度，与软木塞的年生长呈负相关，特别是对于较低的分位数。最高年度温度被证明会对年度软木厚度产生负面影响，特别是对于第6分位数的树木。定义朗指数（LI）的年降水量与平均温度之间的比率与软木塞的年增长率呈下降的抛物线关系。找到最佳的软木生长条件，使Lang指数值达到60，这对应于半干旱气候和潮湿气候之间的过渡。最终模型的开发用于估算软木塞在八年的生长期中的厚度，从而可以预测和绘制适合于天然软木塞生产的软木百分比。结果表明，南部和中部沿海地区以及塔霍河沿岸地区有望获得更高的价值。北部沿海和山区，特征在于Lang指数值高达60（潮湿的气候），目前适用于天然软木塞生产的软木百分比估计值较低。预计在南部和中部沿海地区的气候变化情景下，估计值将降低。