The Peruvian upwelling system (PUS) is the most productive Eastern Boundary Upwelling System (EBUS) of the world ocean. Contrarily to higher latitude EBUSs, there is no consensus yet on the response of upwelling-favorable winds to regional climate change in this region. Global climate models are not able to reproduce the nearshore surface winds, and only a few downscaling studies have been performed by using relatively coarse-grid atmospheric models forced by idealized climate change scenarios. In the present study, the impact of climate change on the PUS upwelling-favorable winds was assessed using a high resolution regional atmospheric model to dynamically downscale the multi-model mean projection of an ensemble of 31 CMIP5 global models under the RCP8.5 worst-case climate scenario. We performed a 10-year retrospective simulation (1994–2003) forced by NCEP2 reanalysis data and a 10-year climate change simulation forced by a climate change forcing (i.e. differences between monthly-mean climatologies for 2080–2100 and 1989–2009) from CMIP5 ensemble added to NCEP2 data. We found that changes in the mean upwelling-favorable winds are weak (less than 0.2 m s⁻¹). Seasonally, summer winds weakly decrease (by 0–5%) whereas winter winds weakly increase (by 0–10%), thus slightly reinforcing the seasonal cycle. A momentum balance shows that the wind changes are mainly driven by the alongshore pressure gradient, except in a local area north of the Paracas peninsula, downstream the main upwelling center, where wind increase in winter is driven by the shoreward advection of offshore momentum. Sensitivity experiments show that the north–south sea surface temperature gradient plays an important role in the wind response along the north and central coasts, superimposed onto the South Pacific Anticyclone large-scale forcing. A reduction (increase) of the gradient induces a wind weakening (strengthening) up to 15% (25%) off the northern coast during summer. This local mechanism is not well represented in global climate models projections, which underlines the strong need for dynamical downscaling of coastal wind in order to study the impact of climate change on the Peruvian upwelling ecosystem.

秘鲁上升流系统（PUS）是世界海洋上生产力最高的东部边界上升流系统（EBUS）。与纬度较高的EBUS相反，尚无共识认为该地区上升的有利风对区域气候变化的响应。全球气候模型无法再现近海地表风，仅通过使用理想化的气候变化情景所强迫的相​​对粗网格大气模型，进行了一些规模缩小研究。在本研究中，使用高分辨率区域大气模型评估了气候变化对PUS上升有利风的影响，以在RCP8.5最差条件下动态缩小31个CMIP5全球模型集合的多模型平均投影。案例气候方案。我们根据NCEP2再分析数据进行了为期10年的回顾性模拟（1994-2003年），并根据气候变化强迫（即2080-2100年和1989-2009年的月平均气候差异）进行了10年的气候变化模拟。 CMIP5合奏已添加到NCEP2数据中。我们发现平均上升气流的变化很小（小于0.2 m s^-1）。季节性上，夏季风微弱地减少（0-5％），而冬季风微弱地增加（0-10％），从而略微增强了季节周期。动量平衡表明，风的变化主要由沿岸的压力梯度驱动，除了在帕拉卡斯半岛以北的局部区域，在主要上升流中心的下游，冬季的风量增加是由近海动量的沿岸平流驱动的。敏感性实验表明，南北海表温度梯度在沿北海岸和中海岸的风响应中起着重要作用，并叠加在南太平洋反气旋的大尺度强迫作用上。夏季，梯度的减小（增加）会导致北海岸风向减弱（增强）达15％（25％）。