@article{fdi:010088872,
  title = {{P}erformance of regional climate model precipitation simulations over the terrain-complex {A}ndes-{A}mazon transition region},
  author = {{G}utierrez, {R}. {A}. and {J}unquas, {C}l{\'e}mentine and {A}rmijos, {E}. and {S}{\¨o}rensson, {A}. {A}. and {E}spinoza, {J}. {C}.},
  editor = {},
  language = {{ENG}},
  abstract = {{R}egional climate models ({RCM}s) are widely used to assess future impacts associated with climate change at regional and local scales. {RCM}s must represent relevant climate variables in the present-day climate to be considered fit-for-purpose for impact assessment. {T}his condition is particularly difficult to meet over complex regions such as the {A}ndes-{A}mazon transition region, where the {A}ndean topography and abundance of tropical rainfall regimes remain a challenge for numerical climate models. {I}n this study, we evaluate the ability of 30 regional climate simulations (6 {RCM}s driven by 10 global climate models) to reproduce historical (1981-2005) rainfall climatology and temporal variability over the {A}ndes-{A}mazon transition region. {W}e assess spatio-temporal features such as spatial distribution of rainfall, focusing on the orographic effects over the {A}ndes-{A}mazon "rainfall hotspots" region, and seasonal and interannual precipitation variability. {T}he {E}ta {RCM} exhibits the highest spatial correlation (up to 0.6) and accurately reproduces mean annual precipitation and orographic precipitation patterns across the region, while some other {RCM}s have good performances at specific locations. {M}ost {RCM}s simulate a wet bias over the highlands, particularly at the eastern {A}ndean summits, as evidenced by the 100%-2,500% overestimations of precipitation in these regions. {A}nnual cycles are well represented by most {RCM}s, but peak seasons are exaggerated, especially at equatorial locations. {N}o {RCM} is particularly skillful in reproducing the interannual variability patterns. {R}esults highlight skills and weaknesses of the different regional climate simulations, and can assist in the selection of regional climate simulations for impact studies in the {A}ndes-{A}mazon transition zone. {R}egional climate models ({RCM}s) are useful numerical tools to investigate future climate change impacts (e.g., future water availability, frequency of floods and droughts, regional warming). {R}egarding regional scale, {RCM}s are expected to perform better than global climate models due to finer spatial resolution. {H}owever, in the {A}ndes-{A}mazon transition region, assessing the performance of {RCM}s is challenging due to complex terrain and scarcity of observations. {T}his region is of critical importance for the water cycle of local and regional ecological systems, but has been often overlooked in {RCM} assessments. {H}ere, we evaluate how 30 regional climate simulations perform in representing precipitation regional contrasts, wet-dry seasons, and year-to-year changes over the {A}ndes-{A}mazon transition region. {W}e find that models perform differently over specific regions, with prominent overestimations at high altitudes by most {RCM}s. {H}owever, {E}ta {RCM} has the best performance regarding regional patterns of precipitation and its wet-dry fluctuations. {B}esides overestimations during austral summer and spring, wet-dry seasonal fluctuations are well simulated by most {RCM}s, but none excels in representing wet-dry yearly fluctuations. {S}trengths and weaknesses of different regional climate simulations are shown, and can help choose the most appropriate simulations for distinct impact studies in this region. {P}recipitation output from 30 regional climate simulations is assessed over the {A}ndes-{A}mazon in terms of climatology and temporal variability{T}he spatio-temporal behavior of seasonality is well reproduced by most simulations, with overestimations during austral summer and spring{W}hile orographic precipitation is a major challenge for most regional climate models, {E}ta satisfactorily reproduces climate patterns in the {A}ndes-{A}mazon region},
  keywords = {rainfall hotspots ; {S}outh {A}merica ; {CORDEX} ; model evaluation ; {AMERIQUE} {DU} {SUD} ; {ANDES} ; {AMZONIE}},
  booktitle = {},
  journal = {{J}ournal of {G}eophysical {R}esearch : {A}tmospheres},
  volume = {129},
  numero = {1},
  pages = {e2023{JD}038618 [23 p.]},
  ISSN = {2169-897{X}},
  year = {2024},
  DOI = {10.1029/2023jd038618},
  URL = {https://www.documentation.ird.fr/hor/fdi:010088872},
}