Access to this full-text is provided by Wiley.
Content available from Journal of Geophysical Research: Atmospheres
This content is subject to copyright. Terms and conditions apply.
1. Introduction
Current and future hydroclimatic trajectories pose significant challenges to global socioeconomic and environ-
mental systems, including fragile ecosystems (Pörtner etal.,2022). Hydroclimatic changes in the western Amazon
Abstract Regional climate models (RCMs) are widely used to assess future impacts associated with climate
change at regional and local scales. RCMs must represent relevant climate variables in the present-day climate
to be considered fit-for-purpose for impact assessment. This condition is particularly difficult to meet over
complex regions such as the Andes-Amazon transition region, where the Andean topography and abundance
of tropical rainfall regimes remain a challenge for numerical climate models. In this study, we evaluate the
ability of 30 regional climate simulations (6 RCMs driven by 10 global climate models) to reproduce historical
(1981–2005) rainfall climatology and temporal variability over the Andes-Amazon transition region. We assess
spatio-temporal features such as spatial distribution of rainfall, focusing on the orographic effects over the
Andes-Amazon “rainfall hotspots” region, and seasonal and interannual precipitation variability. The Eta 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 RCMs have good performances at specific
locations. Most RCMs simulate a wet bias over the highlands, particularly at the eastern Andean summits,
as evidenced by the 100%–2,500% overestimations of precipitation in these regions. Annual cycles are well
represented by most RCMs, but peak seasons are exaggerated, especially at equatorial locations. No RCM is
particularly skillful in reproducing the interannual variability patterns. Results 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 Andes-Amazon transition zone.
Plain Language Summary Regional climate models (RCMs) are useful numerical tools to
investigate future climate change impacts (e.g., future water availability, frequency of floods and droughts,
regional warming). Regarding regional scale, RCMs are expected to perform better than global climate models
due to finer spatial resolution. However, in the Andes-Amazon transition region, assessing the performance
of RCMs is challenging due to complex terrain and scarcity of observations. This region is of critical
importance for the water cycle of local and regional ecological systems, but has been often overlooked in RCM
assessments. Here, we evaluate how 30 regional climate simulations perform in representing precipitation
regional contrasts, wet-dry seasons, and year-to-year changes over the Andes-Amazon transition region. We
find that models perform differently over specific regions, with prominent overestimations at high altitudes by
most RCMs. However, Eta RCM has the best performance regarding regional patterns of precipitation and its
wet-dry fluctuations. Besides overestimations during austral summer and spring, wet-dry seasonal fluctuations
are well simulated by most RCMs, but none excels in representing wet-dry yearly fluctuations. Strengths and
weaknesses of different regional climate simulations are shown, and can help choose the most appropriate
simulations for distinct impact studies in this region.
GUTIERREZ ETAL.
© 2024. The Authors.
This is an open access article under
the terms of the Creative Commons
Attribution License, which permits use,
distribution and reproduction in any
medium, provided the original work is
properly cited.
Performance of Regional Climate Model Precipitation
Simulations Over the Terrain-Complex Andes-Amazon
Transition Region
Ricardo A. Gutierrez1,2 , Clémentine Junquas3,4 , Elisa Armijos1,2 , Anna A. Sörensson5,6,7 ,
and Jhan-Carlo Espinoza3,8
1Escuela de Posgrado, Universidad Nacional Agraria La Molina, Lima, Perú, 2Subdirección de Ciencias de la Atmósfera
e Hidrósfera, Instituto Geofísico del Perú, Lima, Perú, 3Univ. Grenoble Alpes, IRD, CNRS, INRAE, Grenoble INP, IGE,
Grenoble, France, 4Servicio Nacional de Meteorología e Hidrología (SENAMHI), Lima, Peru, 5Facultad de Ciencias Exactas
y Naturales, Universidad de Buenos Aires, Buenos Aires, Argentina, 6Centro de Investigaciones del Mar y la Atmósfera
(CIMA), CONICET ‒ Universidad de Buenos Aires, Buenos Aires, Argentina, 7Instituto Franco-Argentino de Estudios sobre
el Clima y sus Impactos (IFAECI - IRL 3351), CNRS-CONICET-IRDUBA, Buenos Aires, Argentina, 8Departamento de
Ciencias, Sección Matemáticas, Pontificia Universidad Católica del Perú, Lima, Perú
Key Points:
• Precipitation output from 30 regional
climate simulations is assessed
over the Andes-Amazon in terms of
climatology and temporal variability
• The spatio-temporal behavior of
seasonality is well reproduced by most
simulations, with overestimations
during austral summer and spring
• While orographic precipitation is a
major challenge for most regional
climate models, Eta satisfactorily
reproduces climate patterns in the
Andes-Amazon region
Supporting Information:
Supporting Information may be found in
the online version of this article.
Correspondence to:
R. A. Gutierrez,
ra.gvillarreal@gmail.com
Citation:
Gutierrez, R. A., Junquas, C., Armijos,
E., Sörensson, A. A., & Espinoza, J.-C.
(2024). Performance of regional climate
model precipitation simulations over the
terrain-complex Andes-Amazon transition
region. Journal of Geophysical Research:
Atmospheres, 129, e2023JD038618.
https://doi.org/10.1029/2023JD038618
Received 2 FEB 2023
Accepted 28 NOV 2023
10.1029/2023JD038618
RESEARCH ARTICLE
1 of 23
Journal of Geophysical Research: Atmospheres
GUTIERREZ ETAL.
10.1029/2023JD038618
2 of 23
(0°–16°S and 80°W–69°W), including the Andes-Amazon transition region, are linked to the intensification
of seasonal floods and droughts resulting from shifts in precipitation intensities (Arias et al.,2021; Espinoza
etal.,2019,2016; Haghtalab etal.,2020; Pabón-Caicedo etal.,2020). Anthropogenic climate and land-surface
alterations also contribute to these changes (Marengo etal.,2018; Nobre et al.,2016). Global warming and
deforestation have already disrupted the hydroclimatic functioning of ecosystems such as glaciers, paramos, rain-
forests, and montane cloud forests, which provide essential ecological services. Continued human-induced altera-
tions are expected to exacerbate these impacts (Adler etal.,2022; Boulton etal.,2022; Ometto etal.,2022; Vuille
etal.,2018; Young etal.,2011). This is of particular concern for the Andes-Amazon transition region, known
as the rainiest and most biodiverse area within the Amazon basin (Espinoza etal., 2015; Hoorn etal.,2010).
This region is a major source of sediment for the Amazon rivers and plays a crucial role in precipitation recy-
cling within South America (Armijos etal.,2020; Staal et al.,2018). Therefore, there is a pressing need for
future precipitation projections much needed to inform the development of adaptation policies and strategies for
addressing climate change in this region.
Global climate models (GCMs) are widely recognized as essential tools for studying future climate change, with
their simulations being coordinated through the Coupled Model Intercomparison Projects (CMIP3, CMIP5, and
CMIP6; Eyring etal.,2016; Meehl etal.,2007; Taylor etal.,2012, respectively). However, GCMs have limi-
tations due to their relatively coarse spatial resolution, which results in an inability to capture many local and
sub-regional processes. This limitation is particularly pronounced in the Andes-Amazon transition region, given
its complex orography and the associated mesoscale circulation processes that cannot be adequately represented
by these coarse grids. Consequently, these models often struggle to accurately depict the precipitation clima-
tology in this region, frequently leading to substantial overestimations across the tropical Andes (Almazroui
etal., 2021; Ortega etal.,2021). In this context, this bias in precipitation is notable, with values consistently
exceeding 200% throughout the year (Ortega etal.,2021).
Regional climate models (RCMs) are frequently employed as dynamic downscaling tools for GCMs, aiming to
provide more detailed climate information by better capturing topographical and land-surface heterogeneities
(Ambrizzi etal.,2019; Giorgi & Gutowski,2015). Notably, the Coordinated Regional Downscaling Experiment
(CORDEX; Giorgi & Gutowski,2015) has established a unified framework for conducting regional climate simu-
lations worldwide, including South America (CORDEX-SAM).
However, it is important to note that many CORDEX simulations exhibit biases similar to those of GCMs, such as
the substantial overestimation of mean summer precipitation along the tropical Andes, often exceeding observed
amounts by a factor of two or more (Chou etal.,2014; Menéndez etal.,2016; Solman & Blázquez,2019). Never-
theless, RCMs of the CORDEX-type hold the potential to enhance the representation of precipitation climato-
logical fields in regions characterized by complex topography (Bozkurt etal.,2019; Prein etal.,2016; Torma
etal.,2015).
Projections of future precipitation changes generated by CORDEX-SAM RCMs generally align with some
aspects of GCM projections in the Amazon and the Andes, including increased summer precipitation in the tropi-
cal Andes and drier conditions in southwestern Amazonia. However, it is important to emphasize that the magni-
tude and directions of these changes vary depending on the specific RCM used (e.g., Blázquez & Solman,2020;
Llopart etal.,2019; Reboita etal.,2022).
The Andes-Amazon transition region features intricate precipitation patterns influenced by the interplay
of large-scale and local circulation patterns and the region's physio-geographical characteristics (Espinoza
etal.,2020 and references therein). Precipitation patterns across the western Amazon basin typically exhibit a
seasonal distribution, although some Peruvian and Ecuadorian Amazon basins display unimodal and bimodal
regimes (Espinoza etal.,2009; Laraque etal.,2007; Segura etal.,2019; J. C. Sulca and Rocha,2021).
Within the equatorial Amazon basin, a notable rainfall peak occurs during March to May (MAM), corresponding
to the southward movement of the Intertropical Convergence Zone (ITCZ). Another peak is observed in October
and November, associated with the westward transport of moisture linked to the initiation of the South American
Monsoon System (SAMS) (Vera etal.,2006).
In the central and southern Peruvian Amazon, a marked rainy season occurs during the austral summer, span-
ning from December to March, and is closely linked to the mature phase of the SAMS. Conversely, the dry
season prevails from May to September. Furthermore, maximum precipitation regions, often referred to as
Journal of Geophysical Research: Atmospheres
GUTIERREZ ETAL.
10.1029/2023JD038618
3 of 23
rainfall hotspots, are concentrated along the eastern slopes of the Andes (Chavez & Takahashi,2017; Espinoza
etal.,2015). These rainfall hotspots result from forced convection caused by to the orographic effect of the Andes
(e.g., Eghdami & Barros,2020; Garreaud,2009; Junquas etal.,2018).
Therefore, the substantial diversity of rainfall regimes, stemming from the complex orographic features, renders
regional climate modeling in the Andes-Amazon transition region a particularly challenging endeavor. However,
it is worth noting that these challenges are compounded with uncertainties associated with observational
data, including both satellite and gauge-based products, as well as the limited availability of observations in
the Andes-Amazon transition (Cazorla etal.,2022; Condom etal.,2020; Falco etal.,2019; Fassoni-Andrade
etal.,2021; Gibson etal.,2019).
Our study aims to validate RCMs in their ability to accurately simulate the spatial distribution, seasonal patterns,
and interannual variability of precipitation within the Andes-Amazon transition region. To this end, we employ
30 GCM-RCM simulations, encompassing models within the CORDEX framework and the Eta RCM. In our
analysis, we aim to elucidate the primary biases and strengths exhibited these models. Additionally, we delve
into a less explored aspect of RCM evaluation, namely the orographic precipitation patterns found within the
rainfall hotspots. Lastly, we assist in the selection of suitable GCM-RCM combinations for the Andes-Amazon
transition region by ranking the performance of each simulation in the reproduction of various precipitation
features.
2. Study Area and Data Sets
2.1. Study Area
Our study centers on the Andean-Amazon transition region, delineated by the coordinates 0°–16°S and 80°–69°W
(Figure1). This region encompasses the Peruvian-Ecuadorian Andean highlands within the Amazon basin, the
eastern slopes of the Andes, and the western Amazon lowlands (altitudes below 500m.a.s.l.).
In this region, two major Andean-Amazonian river basins are present: the Marañón and the Ucayali river basins.
The Marañón and the Ucayali rivers are the main northwestern and southern tributaries, respectively, of the Peru-
vian Amazon River (blue contoured lines in Figure1). These areas are further explored separately due to their
distinct spatio-temporal precipitation patterns, primarily within the Andes-Amazon transition region (Espinoza
etal.,2009; Figueroa etal.,2020; W. Lavado-Casimiro & Espinoza,2014; W. S. Lavado-Casimiro etal.,2013;
Valenzuela etal.,2023).
2.2. Reference Precipitation Data Sets
We chose two precipitation gridded data sets as reference data sets to address observational uncertainties. First,
we utilized “Rain for Peru and Ecuador” (RAIN4PE version 1), which is a reverse hydrological model derived
from multi-source precipitation data sets (Fernández-Palomino etal.,2022). Second, we employed the “Climate
Hazards Group InfraRed Precipitation with Station data” (CHIRPS 2.0), a product that blends satellite and
rain-gauge data (Funk etal.,2015). Monthly timestep data from 1981 to 2005 was used to match “historical”
RCM simulations.
It is worth noting that RAIN4PE demonstrated superior performance compared to several other precipitation
gridded products, including CHIRPS, when compared against rain-gauge data within the Andes-Amazon tran-
sition region (Fernández-Palomino etal.,2022). Furthermore, when employed as a hydrological model input,
RAIN4PE was the only precipitation data set to achieve water budget balance, significantly enhancing the accu-
racy of daily streamflow simulations within our study area (Fernández-Palomino etal.,2022).
Given that the diagnosis of water budget coherence appears as an effective method to assess the quality of precip-
itation gridded data sets (Fassoni-Andrade etal.,2021), we are confident enough to select RAIN4PE as the main
reference data set and CHIRPS as a secondary reference data set.
To evaluate the spatial variability of orographic rainfall across the Andes-Amazon transition region, we employed
the GTOPO30 elevation data set (Earth Resources Observation and Science (EROS) Center,2018). This data set
offers a horizontal grid resolution of 30 arc seconds, which is approximately equivalent to 1km. To match the
precipitation gridded data sets' horizontal grid resolution, a conservative remapping was performed.
Journal of Geophysical Research: Atmospheres
GUTIERREZ ETAL.
10.1029/2023JD038618
4 of 23
2.3. RCM Simulations
We obtained monthly mean daily (mm/day) precipitation data from 30 “historical” RCM simulations for the
South American domain. This data was sourced from the ESGF site (esgf-data.dkrz.de/search/esgf-dkrz/) at
varying horizontal spatial resolutions: 0.5°, 0.44°, 0.22°, and 0.20°, hereinafter referred to as S50, S44, S22, and
S20, respectively. The analysis period spans from 1981 to 2005.
These simulations were conducted by five RCMs within the CORDEX-SAM framework. These models are:
RegCM v4.7, REMO2015, RegCM v4.3, WRF, and RCA4, denoted as RC47, REMO, RC43, WRF, and RCA,
respectively. In addition, we incorporated three complementary simulations that were performed using the Eta
model, as described by Chou etal.(2014). Terrain elevation data from each RCM was also collected. Table1
further summarizes the details of GCM-RCM simulations (single realizations) used in this study. Detailed infor-
mation about the model physical setup for each RCM can be found in Table A1 in "Description and user guide of
the worldwide CORDEX C3S data set assessing potential conflicts due to overlaps” (available at https://conflu-
ence.ecmwf.int/display/CKB/CORDEX%3A+Regional+climate+projections).
3. Methods
The evaluation of RCMs was focused on the precipitation climatology, seasonal cycles, and interannual varia-
bility across several regions and subregions (Figure2). We performed a bilinear interpolation of precipitation to
a 0.25°×0.25° common grid size to compute the performance of the models within the reference products. In
addition, analyses considered the construction of RCM ensembles, which lets us cluster the simulations into their
Figure 1. The Andes-Amazon basin with topography (GTOPO30) depicted in shades. Boxes 1, 2, 3, 4, and 5 represent
Chazuta, upper Napo-Pastaza, Tamshiyacu, Tingo María, and Quincemil regions. Purple hatching represents the transects
followed to explore the precipitation-topography variability over two precipitation hotspots (i.e., Tingo María and Quincemil,
see Section3.3).
Journal of Geophysical Research: Atmospheres
GUTIERREZ ETAL.
10.1029/2023JD038618
5 of 23
Driving GCM
RCM and available resolution
GCM referenceRCA4RegCM4.3 RegCM4.7 WRF341Iv2 REMO2015 Eta v1
CanESM2 Both S44 and S50 – – S44 – S20 von Salzen etal.(2013)
IPSL-CM5A-MR Only S44 – – – – – Dufresne etal.(2013)
CNRM-CM5 Both S44 and S50 – – – – – Voldoire etal.(2013)
CSIRO Mk3.6 Both S44 and S50 – – – – – Rotstayn etal.(2009)
EC-EARTH Both S44 and S50 – – – – – Hazeleger etal.(2010)
HadGEM2-ES Only S50 S44 S22 – S22 S20 Collins etal.(2011)
MIROC5 Both S44 and S50 – – – – S20 Watanabe etal.(2010)
MPI-ESM-LR Both S44 and S50 – S22 – S22 – Zanchettin etal.(2013)
NorESM1-M Both S44 and S50 – S22 – S22 – Bentsen etal.(2013)
GFDL-ESM2M Both S44 and S50 S44 – – – – Dunne etal.(2012)
RCM reference Samuelsson etal.(2011,2015) Giorgi etal.(2012) Giorgi etal.(2012) Skamarock etal.(2008) Jacob etal.(2012) Mesinger etal.(2012)
Note. The dashed lines mean that no GCM-RCM combination is available. Spatial resolutions of the RCM simulations used in this study, namely 0.2°×0.2°, 0.22°×0.22°, 0.44°×0.44°, and 0.5°×0.5°,
are represented by S20, S22, S44, and S50, respectively. The red characters within the GCM and RCM names are used as acronyms throughout the text (e.g., HadG for HadGEM2-ES and RC47 for
RegCM4.7).
Table 1
Summary of the Regional Climate Model (RCM) Simulations Used in This Study, Showing the Available Global Climate Model (GCM)-RCM Combinations Used Throughout This Study, Spatial
Resolutions of RCM Output, and the References of RCMs and Driving GCMs
Journal of Geophysical Research: Atmospheres
GUTIERREZ ETAL.
10.1029/2023JD038618
6 of 23
downscaling RCM and correspondent spatial resolution. No post-processing, such as bias correction or statistical
downscaling, has been applied to the RCM simulations as the purpose of this study is model evaluation.
3.1. Spatial Pattern Assessment
The Taylor diagram (Taylor,2001) was utilized to assess the skill of the spatial pattern of mean annual precipita-
tion (MAP) simulated by the RCMs in comparison to a specific gridded precipitation data set. The visualization
of the Taylor diagram involves three key metrics: Pearson's correlation coefficient (Corr), normalized standard
deviation, and normalized root mean square error (RMSE). Normalization is achieved by dividing the standard
deviation by the reference standard deviation. Normalized RMSE is obtained by applying a law of cosines-like
relationship to the correlation and the normalized standard deviation, as described by Taylor(2001). The reader
is referred to Taylor(2001) for more in-depth information on the formulation of these metrics and the underlying
technique.
Additionally, the assessment derived from the Taylor diagram was further complemented by employing empirical
cumulative density functions (CDFs), which were applied to the mean annual values of each grid cell within the
study area.
3.2. Orographic Precipitation Relations Over the Andes-Amazon Transition Region
We performed two precipitation-topography profile analyses (purple hatching in Figure1) within two distinct
rainfall hotspot regions. For each hotspot, a minimum of eight transects were designed to follow a trajectory
commencing from the windward side of the mountains and concluding at the eastern Andean summits. To ensure
clarity and prevent visual clutter in the figures, we selected the most effective combinations of S20 and S22
RCMs based of their performance as depicted in the Taylor diagrams across both rainfall hotspot regions.
Within these hotspot regions, a linear relationship between elevation and precipitation is not readily discernible;
instead, precipitation tends to peak at altitudes ranging from 1,000 to 1,500m.a.s.l., as observed in prior studies
(Chavez & Takahashi,2017; Espinoza etal., 2015,2009), Consequently, we calculated orographic gradients,
which are defined as the quotient between differences in precipitation rates at two distinct locations and their
Figure 2. Flowchart of the evaluation process performed in this study. Overall precipitation and equivalent elevation input for the assessment process are marked
by blue shapes. Ranked metrics and their visualizations are represented by orange quadrangles. Distillation and summarization of results are represented by green
quadrangles. Red lines (blue) represent processes and visualizations that are performed at original grid-size (interpolated grid-size to 0.25°×0.25°).
Journal of Geophysical Research: Atmospheres
GUTIERREZ ETAL.
10.1029/2023JD038618
7 of 23
corresponding differences in altitude. These calculations were carried out along two altitudinal sections for each
transect, distributed across the two rainfall hotspots under investigation (see Figure1). These sections were split
by the altitude at which the precipitation maximum is located. Thus, the lower section is bounded by the mini-
mum precipitation in lowlands and the precipitation maximum, and the upper section extends from the precipi-
tation maximum to the summits.
3.3. Seasonal and Interannual Timescales Assessment
The analysis of annual cycle regimes was concentrated on five distinct regions within the Andes-Amazon transi-
tion region area. These regions include Chazuta, upper Napo-Pastaza, Tamshiyacu, Tingo María, and Quincemil,
with their boundaries demarcated by the outlined boxes in Figure1.
Following Espinoza etal.(2009), seasonal and interannual rainfall variability over the Andes-Amazon basin were
evaluated using the seasonal coefficient of variation (sVC) and the interannual coefficient of variation (iVC),
respectively. The sVC was computed by determining the coefficient of variation for the monthly mean precipi-
tation values within each grid cell. To calculate the iVC, we initially applied a 12-month moving average with
a sliding temporal window to emphasize the annual variability. Subsequently, we calculated the coefficient of
variation for the smoothed time-series data, resulting in the iVC (1981–2005) values for each grid cell.
Additionally, we computed the monthly mean RMSE for each grid cell's monthly mean precipitation climatology,
employing RAIN4PE as the reference data set.
3.4. Metrics to Rank the RCM Performances
Inspired by Mascaro etal. (2018), we computed an error metric for each of the seven metrics defined above
(Table2). Error metrics were developed for MAP and spatial correlation to evaluate the spatial distribution.
Mean precipitation profile differences (MPD) between RCMs and the reference data set were used to assess the
orographic rainfall within the precipitation hotspots. In addition, error metrics for sVC and RMSE (iVC) were
constructed to evaluate seasonal (interannual) variability.
Subsequently, we assigned a ranking to each model based on the performance according to each error metric, with
lower rankings indicating superior RCM performance. Additional details regarding the formulation of these error
metrics can be found in Supporting InformationS1.
4. Results
4.1. Spatial Pattern and Annual Cycles of Rainfall Simulated by RCMs
We validate RCM simulations based on the patterns represented by RAIN4PE and CHIRPS precipitation data
sets. Both reference data sets exhibit a north-to-south gradient over the Amazonian lowlands, with a maximum
over the Equator above 6mm/d and at the Ecuadorian eastern Andean slopes, with maximums of 10–14mm/d
Property Variable Calculated as Error
Spatial pattern MAP The annual mean of each grid cell |MAPref–MAPrcm|
Corr Spatial correlation coefficient between MAPrcm and MAPref 1–Corrrcm
Orographic rainfall MPD across Quincemil and Tingo María
profiles
The average profile of the precipitation and topography transects
in references and simulations
|Pref–Prcm| across
mean profile
Seasonal variability sVC The coefficient of variation between monthly means (m=1, 2,
…, 12) of each grid cell
|sVCref–sVCrcm|
RMSE The average of monthly mean root mean square error (RMSE)
between references and simulations
(
1
12 )
12
∑
=1,2...
RMSErcm
,
Interannual variability iVC The coefficient of variation between 12-month-windowed moving
averages of monthly precipitation of each grid cell
|iVCref–iVCrcm|
Table 2
Summary of the Statistics Used to Rank Regional Climate Model (RCM) Performances
Journal of Geophysical Research: Atmospheres
GUTIERREZ ETAL.
10.1029/2023JD038618
8 of 23
(first row of Figure3). Additionally, both products depict a continuous rainfall hotspot along the eastern flank of
the Peruvian Andes, featuring two rainfall hotspot maximums above 10mm/d over Quincemil (12.5°S, 70.5°W)
and Tingo María (9°S, 75.5°W). Furthermore, at the Andean highlands (>4,000m.a.sl., see Figure1), the data
sets consistently indicate low precipitation rates, typically below 4mm/d.
The three S20 Eta simulations successfully capture the north-to-south precipitation gradient, and represent a
relatively narrow, continuous rainfall hotspot region, albeit without clear maximum centers (Figure3, second
row). However, Eta simulations exhibit significant spatial variability in the Andean highlands, with certain grid
Figure 3. Daily mean precipitation between 1981 and 2005 for the western Amazon basin for the (first row) precipitation gridded data sets, (second and third row)
S20 and S22 horizontal resolution regional climate model (RCM) output, (fourth and fifth row) S44 horizontal resolution RCM output, (sixth and seventh row) S50
horizontal resolution RCM output, and (eighth row) RCM ensemble means. Same colors on the names represent simulations belonging to the same RCM, and this color
code will be used throughout the paper.
Journal of Geophysical Research: Atmospheres
GUTIERREZ ETAL.
10.1029/2023JD038618
9 of 23
cells displaying dry conditions (below 1mm/d) and other receiving higher annual precipitation (ranging from
6 to 10mm/d). Notably, the overestimations are more pronounced in the southern part of the Andean highlands
(13–15°S), with the most significant biases occurring during the summer months (not shown).
Furthermore, the Taylor diagram in Figure4 reveals that Eta performs most effectively in replicating spatial patterns
across basins and precipitation hotspots, as seen by correlations ranging from 0.3 to 0.6 when compared to RAIN4PE.
S22 REMO simulations tend to exhibit excessive precipitation over the eastern Andean summits. These simu-
lations shift the orographic rainfall maximum westward, moving it upward along the slopes and resulting in
excessive precipitation over the Andean highlands (>4,000m.a.s.l.) south of 10°S. Some grids represent precipi-
tation rates of approximately 4–8mm/d in stark contrast to the observed 1–3.5mm/d, leading to overestimations
ranging from 100% to 800%.
Similarly, the S22 RC47 simulations also manifest an overproduction of precipitation over the summits, particu-
larly in the vicinity of the Quincemil hotspot, where precipitation levels reach 50mm/d. This corresponds to a
Figure 4. Taylor diagram showing the 30 global climate model-regional climate model (GCM-RCM) combinations and 6 RCM ensembles over (a) the entire domain
(see Figure1), (b) Marañón basin, (c) Ucayali basin, (d) Tingo María box, and (e) Quincemil box (see Figure1). Same colors represent simulations belonging to
the same RCM. RAIN4PE was selected as the reference data set. RAIN4PE and GCM-RCM combinations have been interpolated to a common 0.25° spatial grid
resolution. The black star represents the reference values where the spatial correlation and the normalized standard deviation is equal to 1. Dashed black semicircle is
located where normalized standard deviation is equal to 1. The radial distance from the black star quantifies the RMSE normalized by the reference standard deviation.
The radial distance and the azimuthal position from the origin quantify the normalized standard deviation and the spatial correlation, respectively.
Journal of Geophysical Research: Atmospheres
GUTIERREZ ETAL.
10.1029/2023JD038618
10 of 23
25-fold overestimation in certain locations. Additionally, these simulations
depict a drier Amazon region south of 5°S, with precipitation rates ranging
from 1.5 to 2mm/d, in comparison to the observed 2–6mm/d, as well as
a drier eastern slope (1.5–4mm/d compared to the observed 8–13mm/d).
Furthermore, these simulations rank the lowest in performance according
to the Taylor diagram, characterized by the strongest spatial variances (2.2–
4.3) and negative to near-zero anticorrelations (ranging from −0.6 to 0.1),
particularly at the hotspots.
Regarding the S44 and S50 resolutions, RCA simulations severely under-
estimate rainfall over the eastern Andean slopes between 0 and 6°S, with
the Ecuadorian region experiencing particularly significant underestimations
(approximately 0.5–1.5mm/d compared to the observed 6–11 mm/d). In
contrast, the RC43 simulations offer a more accurate representation of this
region.
The RCA simulations, when driven by ECEA, MIRO, Nor, and HadG,
successfully capture the latitudinal gradient of rainfall from the Equator to
16°S, producing a spatial maximum over the equatorial Amazon lowlands.
Moreover, these GCM-RCM combinations partially replicate the orographic
rainfall pattern south of 10°S in the precipitation hotspot region, although
they exhibit some wet biases over the mountain summits.
When dividing the entire domain into the Marañón and Ucayali basins
(Figures4b and4c), it becomes apparent that the S44–S50 RCA simulations
generally perform better in the latter basin, as quantified by the higher correlations (−0.4 to 0 compared to 0 to
0.25). This discrepancy may be attributed to the substantial underestimation of precipitation over the Ecuadorian
eastern slopes by these models (approximately 0.5–1.5mm/d compared to 6–11mm/d).
Overall, across all simulations ranging from S20 to S50, wet biases over the Andes are stronger during the
summer season (not shown), being a recurrent bias in both GCM and RCM simulations at the Andean cordillera
(e.g., Falco etal.,2019; Ortega etal.,2021).
The spatial mean of the RCM ensembles (last row of Figure3) shows that while the highest resolution model (S20
Eta) tends to yield the best results, increasing the spatial resolution of available RCMs from S50 to S20 does not
consistently improve precipitation patterns in the Andes-Amazon transition region. Notably, lower performance
exhibited by the S22 RC47 model across different basins and precipitation hotspots is the strongest argument in
support of this observation. S22 RC47 performance degrades in comparison to S44 RC43 as the former simulates
drier Amazonian lowlands south of 5°S and greatly overestimates precipitation at the summits, with certain grid
points reaching precipitation rates as high as 50mm/d. However, besides spatial resolution, RC47 and RC43
models also differ in their physical setup, particularly regarding their convection and land surface parameter-
ization schemes (see Section2.3). While the S20 Eta model excels in reproducing precipitation pattern in the
hotspots, the S22 REMO model still maintains overestimations over the eastern Andean flank slopes and summits
(around 8–20mm/d; Figure6). The REMO model is the second-best model at the Marañón basin and Quincemil
with the best correlations being around 0.3, but does poorly regarding spatial pattern correlation over the Ucayali
basin and the Tingo María hotspot.
The resulting empirical CDF summarizes the wide ranges of the simulated spatial variabilities by the RCM
ensembles, as seen in Figure5. The overestimation of precipitation by the REMO model is evident along all
percentiles, and its 90th percentile is about 1.4mm/d higher than RAIN4PE and CHIRPS. Similarly, the RC47
model demonstrates overestimation at percentiles higher than 96, reaching its maximum at around 55 mm/d
(not shown in the figure x-axis). This model also underestimates precipitation as shown until percentile 50,
as a result of underestimation on Amazonian lowlands south of 5°S. The RCA model tends to underestimate
precipitation due to strong dry biases over most of the study area along all percentiles. The Eta CDF resembles
more the CHIRPS CDF, although it seems to slightly underestimate 35% of its pixels, and its percentile 90 is
underestimated by 1.6mm/d. RC43 CDF overestimates precipitation up to its eighteenth percentile, which can be
attributed to the absence of annual precipitation rates below 2mm/d in the study area.
Figure 5. Empirical cumulative distribution function of daily mean
precipitation (mm/d) over the entire domain for RAIN4PE, CHIRPS, and RCM
ensembles. A black line is drawn at the non-exceedance probability of 90%.
Journal of Geophysical Research: Atmospheres
GUTIERREZ ETAL.
10.1029/2023JD038618
11 of 23
4.2. Orographic Rainfall
Both RAIN4PE and CHIRPS show heterogeneous representations of rainfall across both hotspots, as shown in
Figure6. Specifically, over Tingo María, RAIN4PE (CHIRPS) locates a precipitation maximum at 1,150 (600) m.a.s.l.
Similarly, in Quincemil, RAIN4PE (CHIRPS) places the maximum at 1,300 (1,000) m.a.s.l. It is noteworthy that
RAIN4PE and CHIRPS also exhibit varying precipitation quantities within the regions spanning 1,500–3,000m.a.s.l.,
with RAIN4PE producing approximately twice the precipitation rates compared to CHIRPS within this altitude range.
Regarding selected S20 and S22 RCM simulations, the Eta RCM closely aligns with the quantities and altitudes
of the precipitation maximum observed in both profiles when compared to gridded data sets. However, it is
important to note that precipitation rates to the west of this maximum decrease rapidly upslope across the Tingo
María and Quincemil profiles.
In Quincemil, RC47 tends to significantly underestimate precipitation rates along the slopes (500–1,500m.a.s.l)
with values around 3.5 mm/d, while the observed maximum (10–12 mm/d) is located between 1,000 and
1,500 m.a.s.l. This model also produces an overestimated maximum (mean of 35 mm/d) at an altitude of
4,100m.a.s.l, whereas the gridded products indicate precipitation rates below 5mm/d at this altitude.
It is worth noting that both the best and the worst performances are found with the highest resolution models
(S20 Eta, S22 REMO, and S22 RC47). Excessive precipitation orographic gradients further illustrate that
S22 spatial resolution RCMs may have stronger biases than the S44 and S50 RCMs across both precipita-
tion hotspots (Figure7). High orographic gradients between minimum in lowlands and precipitation maxi-
mum in the Tingo María hotspot in CHIRPS and RAIN4PE may be a consequence of the relatively low
altitude of the precipitation maximum, thereby diminishing their altitudinal differences. Some simulations'
Figure 6. Mean profiles of topographical height (dashed lines) and daily mean precipitation (solid lines) of CHIRPS,
RAIN4PE, and selected S20 and S22 global climate model-regional climate model (GCM-RCM) combinations through the
transects constructed across (a) Tingo María and (b) Quincemil (over region 4 and 5, respectively, see Figure1). Gray shading
represents maximum and minimum altitude across transects by GTOPO30. Single S20 and S22 GCM-RCM combinations
over Tingo María and Quincemil are selected on the basis of the “best” spatial pattern member within the RCM ensemble, as
seen in Figure3. Transects follow a windward-summit orientation (from right to left on this figure).
Journal of Geophysical Research: Atmospheres
GUTIERREZ ETAL.
10.1029/2023JD038618
12 of 23
Figure 7. Boxplots of orographic gradients (calculated as the quotient between differences in precipitation rates at two locations and their corresponding altitudes)
across transects over both precipitation hotspots (see Figure1). Orographic gradients in (a–c) were calculated in the “lower” section of each transect (i.e., between
minimum at lowlands and the precipitation maximum). Orographic gradients in (b–d) were calculated in the “upper” section of each transect (i.e., between precipitation
maximum and the summits). Numbers in parenthesis in (a–c) represent the median altitude of maximum precipitation (in km) across transects for each simulation/
ensemble.
Journal of Geophysical Research: Atmospheres
GUTIERREZ ETAL.
10.1029/2023JD038618
13 of 23
orographic gradients may appear closer to the observed values due to compensations caused by increases of
altitudinal differences, as a consequence of the overestimation of the altitudes of precipitation maximums
(Figures7a–7c).
Focusing in the highest resolution RCMs, the Eta model simulations are closer in representing the altitude of the
spatial maximum in both hotspots (between 590 and 1,210m.a.s.l.). In contrast, REMO and RC47 tend to simu-
late the spatial maximum at higher altitudes in comparison to the observed data. These model differences have
an important effect masking the results of the orographic gradient, which would otherwise be higher if maximum
precipitation altitudes were closer to the observations. However, in the case of the Quincemil hotspot, RC47
notably stands out with the most substantial orographic gradient until the precipitation maximum as a result of
its prominent wet bias (around 50mm/d) around 3,900m.a.s.l (see Figure6). Furthermore, REMO and RC47
simulations exhibit the most prominent negative orographic gradients between the precipitation maximum and
the summits in Tingo María and Quincemil, respectively. In the Eta RCM simulations, the orographic gradients
between the precipitation maximum and the summits are roughly twice the magnitude of the observed values,
possibly due to strong underestimations occurring after the maximum is reached over both hotspots. However, it
is more consistent with RAIN4PE and CHIRPS over Quincemil.
4.3. Seasonal Variability
RCM simulations across the Andes-Amazon transition region successfully capture the overall seasonal fluctua-
tions. However, some RCMs tend to overestimate precipitation mostly during the rainy seasons, especially in the
equatorial-most boxes (Figures8a–8c). Among the analyzed boxes, certain RCMs can effectively represent the
bimodal cycle observed in these regions (Laraque etal.,2007; Segura etal.,2019).
Figure 8. Annual regimes of precipitation mean between 1981 and 2005 for the five boxes defined in Figure1. The thick
black line and shading around it represents the average annual cycle and spread among RAIN4PE and CHIRPS, respectively.
Spread characterizes the maximum and minimum monthly climatologies between RAIN4PE and CHIRPS.
Journal of Geophysical Research: Atmospheres
GUTIERREZ ETAL.
10.1029/2023JD038618
14 of 23
For the Upper Napo-Pastaza and Tamshiyacu boxes, the Eta ensemble displays a relatively uniform pattern
throughout the year. However, it is worth noting that individual simulations within this ensemble exhibit signif-
icant variability in their seasonal cycles (not shown). In contrast, RC43 and REMO consistently produce over-
estimations of approximately 100% during the peak rainy seasons in February–April and October–November.
In the case of the Upper Napo-Pastaza box (Figure8c), both the RCA ensembles and the WRF simulation exhibit
a dry bias (not lower than 70%) throughout the year. Notably, these models simulate almost no precipitation
during June–August, whereas RAIN4PE and CHIRPS data indicate an average rainfall rate of around 10mm/d
during that period.
In the southernmost boxes of Tingo María and Quincemil (Figures8d and8e), RCMs generally exhibit a higher
level of agreement in terms of the shape and quantities of the seasonal cycle, albeit with some biases present in
specific ensembles. For instance, in Tingo María, the S44 RC43 ensemble and both S44 and S50 RCA ensembles
erroneously simulate a bimodal cycle. Specifically, RC43 overestimates rainfall during April and September–
November compared to gridded data sets. Furthermore, the S22 REMO ensemble overestimates precipitation
during August–November in comparison to RAIN4PE data.
In Quincemil, all simulated and observed seasonal cycles show an unimodal regime. Nevertheless, the S44 RC43
and S22 REMO ensembles overestimate precipitation during October–April and September–April, respectively.
S22 RC47 also overestimates precipitation during August–March, primarily due to excessive quantities simulated
at altitudes above 2,000m.a.s.l. Moreover, the S44 and S50 RCA ensembles tend to underestimate precipitation
throughout the year.
The spatial pattern of seasonal variability is illustrated by sVC maps in Figure9. Notably, certain features, such
as the low seasonality observed over the equatorial western Amazon and its subsequent increase south of 6°S are
most accurately captured by the Eta RCM. Furthermore, the Eta RCM adequately reproduces the sVC pattern
over the Andean highlands.
In contrast, REMO and RC47 models exhibit higher seasonal variability over the equatorial Amazon (0.2–0.4)
while simulating lower seasonal variability in the Andean highlands, resulting in sVC values below 0.5. In
addition, RC43, RCA, and WRF models simulate an excessive degree of seasonal variability over the Amazon
lowlands south of 5°S, with sVC values exceeding 0.7.
4.4. Interannual Variability
Figure10 shows the iVC of precipitation gridded data sets and GCM-RCM combinations for the period between
1981 and 2005. Both gridded products consistently identify the Andean highlands as the region with the high-
est interannual variability. However, there are scattered regions in the Amazonian lowlands north of 7.5°S that
exhibit relatively high interannual variability (iVC>0.15). In addition, RAIN4PE represents lower interannual
variability at the Marañón basin than CHIRPS.
None of the GCM-RCM combinations adequately represent the iVC spatial pattern, with a general tendency to
underestimate it, particularly over the Amazonian lowlands, where most GCM-RCM combinations show iVC values
below 0.1 (Figure10). Regarding S20 and S22 RCMs, the Eta and RC47 RCMs, and notably MIRO and HadG
Eta and Nor RC47, demonstrate a relatively high interannual variability (iVC between 0.2 and 0.5, second row in
Figure10) over the Andean highlands. In contrast, REMO RCM simulates low interannual variability (iVC below
0.1) over the Andean highlands. Almost none of these GCM-RCM combinations simulate significant interannual
variability conditions over the Amazonian lowlands north of 7.5°S, which suggests they are unable to reproduce it.
4.5. Final Rankings and Summary of Results
Figure11 presents the rankings of GCM-RCM combinations based on the assessment of seven spatio-temporal
features outlined in Table2, focusing on precipitation during the 1981–2005 period in the Andes-Amazon transi-
tion region, with RAIN4PE serving as the reference. Within these metrics, the Eta RCM excels in the spatial and
seasonal aspects, as evident in the first six rows of Figure11.
The assessment of interannual variability assessment carries out uncertainties due to large spread of iVC, espe-
cially in the Amazonian lowlands. However, as seen in Figure9, no GCM-RCM combinations exhibit significant
skill in representing the spatial features of interannual variability, besides some S20 and S22 GCM-RCM combi-
nations at the Andean highlands. The actual iVC error values exhibit minimal differences between models, with
Journal of Geophysical Research: Atmospheres
GUTIERREZ ETAL.
10.1029/2023JD038618
15 of 23
the best and least ranked models yielding 0.026 and 0.082, respectively, differing by only 0.056 (Figure S4 in
Supporting InformationS1).
Furthermore, it might be expected that high-resolution simulations (S20 and S22) would outperform low-resolution
simulations (S44 and S50). However, S22 RC47 simulations are surpassed by several lower resolution simula-
tions across multiple metrics. Additional analyses are presented using CHIRPS as a reference (Figure S4 in
Supporting InformationS1). The results prove to be more sensitive over both of the analyzed rainfall hotspots
(i.e., Quincemil and Tingo María profiles). Nevertheless, Eta consistently remains as the top-performing model
under both reference data sets.
Figure12 summarizes the main findings in the Andes-Amazon transition region. The spatio-temporal distribu-
tion of seasonality is well-represented across the study area, with the Eta RCM performing best in simulating
the spatial distribution of seasonality. In the northern Marañón basin, it is crucial to consider the representation
Figure 9. Seasonal coefficient of variation (sVC) calculated from monthly precipitation between 1981 and 2005 for the (first
row) precipitation gridded data sets, (second and third row) S20 and S22 horizontal resolution regional climate model (RCM)
output, (fourth and fifth row) S44 horizontal resolution RCM output, (sixth and seventh row) S50 horizontal resolution RCM
output, and (eighth row) RCM ensembles means.
Journal of Geophysical Research: Atmospheres
GUTIERREZ ETAL.
10.1029/2023JD038618
16 of 23
of rainfall in the equatorial Andes and lowlands. Only the Eta and RC43 models capture the spatial maximum at
the eastern flank of the Ecuadorian Andes, with precipitation rates ranging from 10 to 14mm/d. Some Eta and
REMO simulations, particularly those forced by Can and HadG, and by HadG and Nor, respectively, successfully
reproduce a continuous equatorial maximum in the lowlands. These simulations exhibit similar precipitation
intensity as RAIN4PE and CHIRPS. Moving to the southern Ucayali basin and the rainfall hotspots region, Eta
and some S50 RCA simulations effectively depict the spatial extent of rainfall hotspots. Among these models, Eta
is the most accurate in representing the altitude of the maximum (1,000–1,500m.a.sl.).
Regarding biases, in the lowlands and the Andes-Amazon transition region, models tend to exaggerate the annual
cycle of precipitation, leading to overestimations during peak months, with RC43 and REMO reaching biases of
Figure 10. Interannual coefficient of variation (iVC) calculated from mean annual precipitation timeseries between 1981
and 2005 for the (first row) precipitation gridded data sets, (second and third row) S20 and S22 horizontal resolution regional
climate model (RCM) output, (fourth and fifth row) S44 horizontal resolution RCM output, (sixth and seventh row) S50
horizontal resolution RCM output, and (eighth row) RCM-averaged iVC.
Journal of Geophysical Research: Atmospheres
GUTIERREZ ETAL.
10.1029/2023JD038618
17 of 23
approximately 100% during the rainiest months. On the Andean eastern slopes of Ecuador, S44 WRF and S44
and S50 RCA models underestimate precipitation by approximately 70%–90%. In the southern region, wet biases
related to orographic rainfall in the Andean highlands during summer are common among REMO, RCA43,
RC47, and, to a lesser extent, Eta. Another noteworthy bias observed is the upslope shift of rainfall hotspots in
most models, leading to overestimations over the summits of the eastern Andean flanks. RC47 stands out with
Figure 11. Ranks of the relative error defined in Table2 for the seven spatio-temporal properties over the western Amazon basin with RAIN4PE used as the reference.
Journal of Geophysical Research: Atmospheres
GUTIERREZ ETAL.
10.1029/2023JD038618
18 of 23
the most significant wet biases in this region, with certain grids displaying daily precipitation rates in the order
of 50mm/d, which represents an overestimation of approximately 2,500%.
The two latter biases may arise due to the inherent physics of the models and the insufficient spatial resolutions
of the simulations to, which hinder their ability to accurately represent rainfall in complex terrain (e.g., Chou
etal.,2014; Torma et al.,2015). In addition, another major difference related to the simulation of dynamic
processes in mountainous terrain are the vertical coordinates used in the model structure. The Eta model
employs eta vertical coordinates, which remain approximately horizontal with respect to mountains (Mesinger
etal.,2012). Conversely, the other models utilize either sigma terrain-following vertical coordinates or hybrid
vertical coordinates.
5. Discussion
Uncertainties in observational data can significantly influence what is regarded as “skill,” particularly in our
study area, which is a data-scarce region. We selected RAIN4PE and CHIRPS as reference products based on
their performance in grid-to-point comparisons with rain gauges (Fernández-Palomino etal.,2022). Good perfor-
mances of RAIN4PE as a driver for hydrological modeling in the region further supports its selection as the
primary reference product. However, since RAIN4PE is a hydrological model output, caution should be taken as
it is still subject to uncertainties related to water cycle components and model input, such as the streamflow, evap-
otranspiration, precipitation gridded data sets, and the chosen hydrological model structure (Fernández-Palomino
etal.,2022).
We found similar results with both RAIN4PE and CHIRPS as reference products (Figures S5–S7 in Support-
ing InformationS1), except for the MPD at the two evaluated precipitation hotspots. Such differences should
be expected as the product of observational uncertainties in this region due to scarce monitoring, which limits
further calibration of gridded data sets (Condom etal.,2020). Nevertheless, when assessing based on this metric,
Eta simulations still holds best performances regardless of precipitation gridded data set used, reinforcing our
confidence in its ability in resolving precipitation in this region. Nonetheless, there remains a pressing need
Figure 12. A summary of main findings over the Andes-Amazon transition region. Specific regions are depicted by colors
(i.e., the Ecuadorian Amazon slopes are represented by black, the equatorial region by blue, the hotspots region by orange,
and the Andean highlands by green).
Journal of Geophysical Research: Atmospheres
GUTIERREZ ETAL.
10.1029/2023JD038618
19 of 23
for long-term and comprehensive precipitation monitoring on the eastern slopes of the Andean cordillera. The
complex interactions between terrain and regional atmospheric circulation in this area give rise to a wide range
of hydroclimatic regimes (Cazorla etal.,2022; Condom etal.,2020). For example, Newell etal.(2022) demon-
strated the importance of a network of rain gauges to capture the fine-scale spatiotemporal variability of rainfall in
montane cloud forests in the northeastern Andes of Peru. These observations can serve to validate satellite-based
and merged precipitation products which can then be utilized as references to properly assess RCM's ability in the
reproduction of precipitation patterns over the Amazon's rainiest zone (e.g., Chavez & Takahashi,2017; Espinoza
etal.,2015).
While the highest resolution model (S20 Eta) yields the best results, the impact of increasing RCM resolution
from S44–S50 to S20–S22 does not seem to significantly improve the diagnostics of MAP over the rainfall
hotspots for all RCMs. In fact, S22 models exhibit worse performance than S44 models for certain criteria. There-
fore, aside from horizontal resolution, other model setup characteristics may also affect the model performances
(e.g., parameterization, forcing data, vertical levels treatment). Caution should be taken in high-resolution mode-
ling to adequately set-up the model relatively to the climate characteristics of the region of study.
The impact of increasing RCM resolution from S44–S50 to S20–S22 does not appear to markedly affect the diag-
nostics of MAP over the rainfall hotspots. Previous evaluation studies regarding CORDEX RCMs have suggested
spatial resolutions around approximately 12.5×12.5km (not available in CORDEX-SAM) for improved resolu-
tion of precipitation due to enhanced topography representation (Lucas-Picher etal.,2017; Mascaro etal.,2018;
Prein etal.,2016; Torma etal.,2015). Currently, state-of-art RCMs involve simulations at convection-permitting
scales, characterized by spatial resolutions finer than 4 km, enabling the explicit resolution of convection
processes without the need for a convective parameterization scheme (Kendon etal.,2021).
At higher spatial resolutions, RCMs in the tropical Andean region significatively improve precipitation features
such as the spatial pattern, mesoscale processes linked to the diurnal cycle of convection (Gómez-Ríos etal.,2023;
Junquas etal.,2022, 2018; Rosales etal.,2022; Sierra etal., 2022), and the internal structure of mesoscale
convective systems and hailstorms (Flores-Rojas etal.,2021; Moya-Alvarez etal., 2019). However, as these
spatial resolutions range within the so-called “gray zone” of convection, some local convection processes can
be explicitly resolved, while others still require the use of a convection parameterization (Kendon etal.,2021).
The interplay between tropical atmospheric circulation regimes and local physio-geographical features further
amplifies these uncertainties at very high spatial resolutions (1km), posing a persistent challenge for the numer-
ical modeling community (e.g., Junquas etal.,2022).
Our results indicate uncertainties in RCM configurations, specifically related to the choice of downscaling model
and the physical parameterization schemes. Within set of GCM-RCM combinations considered, these choices
appear to exert a more significant influence on the simulation of precipitation, particularly in rainfall hotspots
regions.
For example, a common bias highlighted in this study is the dry bias over the eastern Ecuadorian Andes slopes
by the RCA model and, specifically, the S44 WRF model. Several WRF-based studies have shown that, under
certain parameterizations and higher spatial resolution, precipitation over this region can be either overestimated
or improved (e.g., Chimborazo & Vuille,2021; Junquas etal.,2022; Ochoa etal.,2016). Consequently, the choice
of an appropriate convection parameterization scheme becomes crucial in enhancing the reliability of RCM simu-
lations due to better representation of rainfall characteristics.
Many GCM-RCM combinations successfully replicate both unimodal and bimodal annual precipitation patterns
in the Andes-Amazon transition region (Espinoza etal.,2009; Segura etal.,2019). However, some combinations
tend to overestimate precipitation during peak months at the equatorial-most locations. Additionally, some of
them do at the Andean highlands, and the overestimation is stronger during summer. These findings align with the
analyses of CMIP5 and CMIP6 GCMs simulations performed by Ortega etal.(2021) and Almazroui etal.(2021),
where they also found precipitation overestimations in the tropical Andes, particularly during the rainy seasons.
The source of excessive convection at the equatorial-most Andes-Amazon transition region during wet months
might be produced mainly by the physics choices of RCMs regarding cumulus convection and land-surface
parameterizations, which requires further investigation (Chou etal.,2014).
None of the models is particularly skillful in capturing the interannual variability of rainfall during the period
from 1981 to 2005. This challenge may be linked to the complexities associated with simulating related to the
Journal of Geophysical Research: Atmospheres
GUTIERREZ ETAL.
10.1029/2023JD038618
20 of 23
simulation of teleconnection patterns by both GCMs and GCM-driven RCMs (e.g., see Sections 10.4.2. and 10.6
in Doblas-Reyes etal.,2021). Further advancements in this area are imperative, given that precipitation patterns
over the Andes-Amazon transition region are significantly influenced by tropical oceans and their interaction with
local physio-geographical features (e.g., Arias etal.,2021; Espinoza etal.,2019; Marengo & Espinoza,2016;
Segura etal.,2019; J. Sulca etal.,2018).
6. Conclusions
We compared and evaluated the performance of 30 GCM-RCM simulations within the framework of CORDEX-
SAM and Eta RCM in representing the precipitation spatio-temporal climatological features and interannual
variability during the “historical” period (1981–2005). These simulations result from combination of 6 RCMs
and 10 GCMs at spatial resolutions ranging from 0.2° to 0.5°. The results reveal a mixed performance, with some
aspects well-reproduced such as the spatial behavior of seasonality. However, most RCM simulations struggled
to accurately replicate the spatial patterns of orographic rainfall, and no model excelled in capturing the spatial
features of interannual rainfall variability observed in the period between 1981 and 2005.
In this set of simulations, the best performance is observed in simulations with the finest spatial resolution
when it comes to reproducing orographic precipitation over the Andes-Amazon transition region (specifically,
the Eta RCM at 0.2°×0.2° resolution). However, it is noteworthy that simulations with a grid size of 0.22° may
underperform in comparison to coarser grid-size (ranging from 0.44° to 0.5°) simulations in simulating vari-
ous orographic precipitation features. For example, excessive overestimations reaching as much as 2,500% are
reached by the S22 RC47 simulations in some Andes-Amazon transition region locations. These wet biases are
more pronounced during the rainiest months; and, in the lowlands, precipitation can be overestimated by as much
as 100%, especially at equatorial-most regions.
Addressing these biases in future GCM dynamical downscaling efforts for this region will require consideration
of convection-permitting scales and the selection of an appropriate, high-resolution adapted physical parameter-
ization setup.Further development in these areas is essential to improve the simulation of intricate interactions
between local terrain and tropical rainfall regimes in this complex region.
Finally, the results of this study offer valuable insights that can enhance the application of this set of regional
climate simulations for both climate and non-climate scientists engaged in vulnerability and impact studies at the
local scale under future climate scenarios (Figures11 and12). Future research should focus on a process-oriented
approach to identify sub-monthly (e.g., synoptic) mechanisms leading to model biases. This, in turn, can lead
to the improvement of bias correction techniques (e.g., Maraun etal.,2021). Such approaches would serve in
order to project future scenarios related to local hydrometeorological risks (e.g., Figueroa etal.,2020; Valenzuela
etal.,2023).
Data Availability Statement
The precipitation gridded datasets used in this paper were acquired from the Climate Hazards Group (CHIRPS,
https://chc.ucsb.edu/data/chirps/) and German Research Center for Geosciences (RAIN4PE, https://datapub.
gfz-potsdam.de/download/10.5880.PIK.2020.010enouiv/). The CORDEX-SAM and Eta RCM's data are publicly
available and can be downloaded from the Earth System Grid Federation (ESGF) portals (https://esgf-data.dkrz.
de/search/esgf-dkrz/). The GTOPO30 dataset is available at the U.S. Geological Survey site (https://www.usgs.
gov/centers/eros/science/usgs-eros-archive-digital-elevation-global-30-arc-second-elevation-gtopo30).
References
Adler, C., Wester, P., Bhatt, I., Huggel, C., Insarov, G. E., Morecroft, M. D., etal. (2022). Cross-chapter paper 5: Mountains. In H.-O. Pörtner,
D. C. Roberts, M. Tignor, E. S. Poloczanska, K. Mintenbeck, A. Alegría, etal. (Eds.), Climate change 2022: Impacts, adaptation and vulner-
ability. Contribution of Working Group II to the sixth assessment report of the intergovernmental panel on climate change(pp.2273–2318).
Cambridge University Press. https://doi.org/10.1017/9781009325844.022
Almazroui, M., Ashfaq, M., Islam, M. N., Rashid, I. U., Kamil, S., Abid, M. A., etal. (2021). Assessment of CMIP6 performance and projected
temperature and precipitation changes over South America. Earth Systems and Environment, 5(2), 155–183. https://doi.org/10.1007/
s41748-021-00233-6
Ambrizzi, T., Reboita, M. S., da Rocha, R. P., & Llopart, M. (2019). The state of the art and fundamental aspects of regional climate modeling in
South America. Annals of the New York Academy of Sciences, 1436(1), 98–120. https://doi.org/10.1111/nyas.13932
Acknowledgments
We thank three anonymous reviewers for
their helpful comments, which helped us
enrich the discussions of our study and
improve the quality and clarity of the
manuscript. This research was funded
by the No. 77-2021-FONDECYT/BM
Project and J.-C.E. received the support
of the French AMANECER-MOPGA
project funded by ANR and IRD
(ref. ANR-18-MPGA-0008). We also
acknowledge the contribution of project
ACE-Amazon funded by the regional
program CLIMAT-AmSud (21-CLIMAT-
01). Thanks to Livia Dutra (Universidade
de São Paulo) and Daniela Carneiro
(Instituto Nacional de Pesquisas Espa-
ciais) for providing the orography fields
of the RegCM4.3 and the Eta models,
respectively.
Journal of Geophysical Research: Atmospheres
GUTIERREZ ETAL.
10.1029/2023JD038618
21 of 23
Arias, P.A., Garreaud, R., Poveda, G., Espinoza, J. C., Molina-Carpio, J., Masiokas, M., etal. (2021). Hydroclimate of the Andes Part II: Hydro-
climate variability and sub-continental patterns. Frontiers in Earth Science, 8, 505467. https://doi.org/10.3389/feart.2020.505467
Armijos, E., Crave, A., Espinoza, J. C., Filizola, N., Espinoza-Villar, R., Ayes, etal. (2020). Rainfall control on Amazon sediment flux: Synthesis
from 20 years of monitoring. Environmental Research Communications, 2(5), 051008. https://doi.org/10.1088/2515-7620/ab9003
Bentsen, M., Bethke, I., Debernard, J. B., Iversen, T., Kirkevåg, A., Seland, Ø., et al. (2013). The Norwegian Earth system model,
NorESM1-M—Part 1: Description and basic evaluation of the physical climate. Geoscientific Model Development, 6(3), 687–720. https://doi.
org/10.5194/gmd-6-687-2013
Blázquez, J., & Solman, S. A. (2020). Multiscale precipitation variability and extremes over South America: Analysis of future changes from a
set of CORDEX regional climate model simulations. Climate Dynamics, 55(7–8), 2089–2106. https://doi.org/10.1007/s00382-020-05370-8
Boulton, C. A., Lenton, T. M., & Boers, N. (2022). Pronounced loss of Amazon rainforest resilience since the early 2000s. Nature Climate
Change, 12(3), 271–278. https://doi.org/10.1038/s41558-022-01287-8
Bozkurt, D., Rojas, M., Boisier, J. P., Rondanelli, R., Garreaud, R., & Gallardo, L. (2019). Dynamical downscaling over the complex terrain
of southwest South America: Present climate conditions and added value analysis. Climate Dynamics, 53(11), 6745–6767. https://doi.
org/10.1007/s00382-019-04959-y
Cazorla, M., Gallardo, L., & Jimenez, R. (2022). The complex Andes region needs improved efforts to face climate extremes. Elementa: Science
of the Anthropocene, 10(1), 00092. https://doi.org/10.1525/elementa.2022.00092
Chavez, S. P., & Takahashi, K. (2017). Orographic rainfall hotspots in the Andes-Amazon transition according to the TRMM precipitation radar
and in situ data. Journal of Geophysical Research: Atmospheres, 122(11), 5870–5882. https://doi.org/10.1002/2016JD026282
Chimborazo, O., & Vuille, M. (2021). Present-day climate and projected future temperature and precipitation changes in Ecuador. Theoretical and
Applied Climatology, 143(3–4), 1581–1597. https://doi.org/10.1007/s00704-020-03483-y
Chou, S. C., Lyra, A., Mourão, C., Dereczynski, C., Pilotto, I., Gomes, J., etal. (2014). Evaluation of the Eta simulations nested in three global
climate models. American Journal of Climate Change, 03(05), 438–454. https://doi.org/10.4236/ajcc.2014.35039
Collins, W. J., Bellouin, N., Doutriaux-Boucher, M., Gedney, N., Halloran, P., Hinton, T., et al. (2011). Development and evaluation of an
Earth-System model—HadGEM2. Geoscientific Model Development, 4(4), 1051–1075. https://doi.org/10.5194/gmd-4-1051-2011
Condom, T., Martínez, R., Pabón, J. D., Costa, F., Pineda, L., Nieto, J. J., etal. (2020). Climatological and hydrological observations for the South
American Andes: In situ stations, satellite, and reanalysis data sets. Frontiers in Earth Science, 8, 92. https://doi.org/10.3389/feart.2020.00092
Doblas-Reyes, F. J., Sörensson, A. A., Almazroui, M., Dosio, A., Gutowski, W. J., Haarsma, R., etal. (2021). Linking global to regional climate
change. In V. Masson-Delmotte, P. Zhai, A. Pirani, S. L. Connors, C. Péan, S. Berger, etal. (Eds.), Climate change 2021: The physical science
basis. Contribution of Working Group I to the sixth assessment report of the Intergovernmental Panel on Climate Change (pp.1363–1512).
Cambridge University Press. https://doi.org/10.1017/9781009157896.012
Dufresne, J.-L., Foujols, M.-A., Denvil, S., Caubel, A., Marti, O., Aumont, O., etal. (2013). Climate change projections using the IPSL-CM5
Earth system model: From CMIP3 to CMIP5. Climate Dynamics, 40(9–10), 2123–2165. https://doi.org/10.1007/s00382-012-1636-1
Dunne, J. P., John, J. G., Adcroft, A. J., Griffies, S. M., Hallberg, R. W., Shevliakova, E., etal. (2012). GFDL’s ESM2 global coupled climate–
carbon Earth system models. Part I: Physical formulation and baseline simulation characteristics. Journal of Climate, 25(19), 6646–6665.
https://doi.org/10.1175/JCLI-D-11-00560.1
Earth Resources Observation and Science Center. (2018). Global 30 arc-second elevation (GTOPO30) [Dataset]. United States Geological Survey,
11, 40. Retrieved from https://www.usgs.gov/centers/eros/science/usgs-eros-archive-digital-elevation-global-30-arc-second-elevation-gtopo30
Eghdami, M., & Barros, A. P. (2020). Deforestation impacts on orographic precipitation in the tropical Andes. Frontiers in Environmental
Science, 8, 580159. https://doi.org/10.3389/fenvs.2020.580159
Espinoza, J. C., Chavez, S., Ronchail, J., Junquas, C., Takahashi, K., & Lavado, W. (2015). Rainfall hotspots over the southern tropical Andes:
Spatial distribution, rainfall intensity, and relations with large-scale atmospheric circulation: Rainfall hotspots over the southern tropical
Andes. Water Resources Research, 51(5), 3459–3475. https://doi.org/10.1002/2014WR016273
Espinoza, J. C., Garreaud, R., Poveda, G., Arias, P.A., Molina-Carpio, J., Masiokas, M., etal. (2020). Hydroclimate of the Andes Part I: Main
climatic features. Frontiers in Earth Science, 8, 64. https://doi.org/10.3389/feart.2020.00064
Espinoza, J. C., Ronchail, J., Guyot, J. L., Cochonneau, G., Naziano, F., Lavado, W., etal. (2009). Spatio-temporal rainfall variability in the
Amazon basin countries (Brazil, Peru, Bolivia, Colombia, and Ecuador). International Journal of Climatology, 29(11), 1574–1594. https://
doi.org/10.1002/joc.1791
Espinoza, J. C., Ronchail, J., Marengo, J. A., & Segura, H. (2019). Contrasting North–South changes in Amazon wet-day and dry-day frequency
and related atmospheric features (1981–2017). Climate Dynamics, 52(9–10), 5413–5430. https://doi.org/10.1007/s00382-018-4462-2
Espinoza, J. C., Segura, H., Ronchail, J., Drapeau, G., & Gutierrez-Cori, O. (2016). Evolution of wet-day and dry-day frequency in the western
Amazon basin: Relationship with atmospheric circulation and impacts on vegetation. Water Resources Research, 52(11), 8546–8560. https://
doi.org/10.1002/2016WR019305
Eyring, V., Bony, S., Meehl, G. A., Senior, C. A., Stevens, B., Stouffer, R. J., & Taylor, K. E. (2016). Overview of the Coupled Model Inter-
comparison Project Phase 6 (CMIP6) experimental design and organization. Geoscientific Model Development, 9(5), 1937–1958. https://doi.
org/10.5194/gmd-9-1937-2016
Falco, M., Carril, A. F., Menéndez, C. G., Zaninelli, P.G., & Li, L. Z. X. (2019). Assessment of CORDEX simulations over South Amer-
ica: Added value on seasonal climatology and resolution considerations. Climate Dynamics, 52(7–8), 4771–4786. https://doi.org/10.1007/
s00382-018-4412-z
Fassoni-Andrade, A. C., Fleischmann, A. S., Papa, F., Paiva, R. C. D. D., Wongchuig, S., Melack, J. M., etal. (2021). Amazon hydrology from
space: Scientific advances and future challenges. Reviews of Geophysics, 59(4), 97. https://doi.org/10.1029/2020RG000728
Fernández-Palomino, C. A., Hattermann, F. F., Krysanova, V., Lobanova, A., Vega-Jácome, F., Lavado, W., etal. (2022). A novel high-resolution
gridded precipitation dataset for Peruvian and Ecuadorian watersheds: Development and hydrological evaluation [Dataset]. Journal of Hydro-
meteorology, 23(3), 309–336. https://doi.org/10.1175/JHM-D-20-0285.1
Figueroa, M., Armijos, E., Espinoza, J. C., Ronchail, J., & Fraizy, P. (2020). On the relationship between reversal of the river stage (repi-
quetes), rainfall and low-level wind regimes over the western Amazon basin. Journal of Hydrology: Regional Studies, 32, 100752. https://doi.
org/10.1016/j.ejrh.2020.100752
Flores-Rojas, J. L., Moya-Álvarez, A. S., Valdivia-Prado, J. M., Piñas-Laura, M., Kumar, S., Karam, H. A., etal. (2021). On the dynamic mech-
anisms of intense rainfall events in the central Andes of Peru, Mantaro valley. Atmospheric Research, 248, 105188. https://doi.org/10.1016/j.
atmosres.2020.105188
Funk, C., Peterson, P., Landsfeld, M., Pedreros, D., Verdin, J., Shukla, S., etal. (2015). The climate hazards infrared precipitation with stations—A
new environmental record for monitoring extremes [Dataset]. Scientific Data, 2(1), 150066. https://doi.org/10.1038/sdata.2015.66
Garreaud, R. D. (2009). The Andes climate and weather. Advances in Geosciences, 22, 3–11. https://doi.org/10.5194/adgeo-22-3-2009
Journal of Geophysical Research: Atmospheres
GUTIERREZ ETAL.
10.1029/2023JD038618
22 of 23
Gibson, P.B., Waliser, D. E., Lee, H., Tian, B., & Massoud, E. (2019). Climate model evaluation in the presence of observational uncer-
tainty: Precipitation indices over the contiguous United States. Journal of Hydrometeorology, 20(7), 1339–1357. https://doi.org/10.1175/
JHM-D-18-0230.1
Giorgi, F., Coppola, E., Solmon, F., Mariotti, L., Sylla, M., Bi, X., etal. (2012). RegCM4: Model description and preliminary tests over multiple
CORDEX domains. Climate Research, 52, 7–29. https://doi.org/10.3354/cr01018
Giorgi, F., & Gutowski, W. J. (2015). Regional dynamical downscaling and the CORDEX initiative. Annual Review of Environment and
Resources, 40(1), 467–490. https://doi.org/10.1146/annurev-environ-102014-021217
Gomez-Rios, S., Zuluaga, M. D., & Hoyos, C. D. (2023). Orographic controls over convection in an Inter-Andean valley in northern South Amer-
ica. Monthly Weather Review, 151(1), 145–162. https://doi.org/10.1175/MWR-D-21-0231.1
Haghtalab, N., Moore, N., Heerspink, B. P., & Hyndman, D. W. (2020). Evaluating spatial patterns in precipitation trends across the Amazon
basin driven by land cover and global scale forcings. Theoretical and Applied Climatology, 140(1–2), 411–427. https://doi.org/10.1007/
s00704-019-03085-3
Hazeleger, W., Severijns, C., Semmler, T., Ştefănescu, S., Yang, S., Wang, X., etal. (2010). EC-Earth: A seamless earth-system prediction
approach in action. Bulletin of the American Meteorological Society, 91(10), 1357–1364. https://doi.org/10.1175/2010BAMS2877.1
Hoorn, C., Wesselingh, F. P., ter Steege, H., Bermudez, M. A., Mora, A., Sevink, J., etal. (2010). Amazonia through time: Andean uplift, climate
change, landscape evolution, and biodiversity. Science, 330(6006), 927–931. https://doi.org/10.1126/science.1194585
Jacob, D., Elizalde, A., Haensler, A., Hagemann, S., Kumar, P., Podzun, R., etal. (2012). Assessing the transferability of the regional climate
model REMO to different COordinated Regional Climate Downscaling EXperiment (CORDEX) regions. Atmosphere, 3(1), 181–199. https://
doi.org/10.3390/atmos3010181
Junquas, C., Heredia, M. B., Condom, T., Ruiz-Hernández, J. C., Campozano, L., Dudhia, J., etal. (2022). Regional climate modeling of the
diurnal cycle of precipitation and associated atmospheric circulation patterns over an Andean glacier region (Antisana, Ecuador). Climate
Dynamics, 58(11–12), 3075–3104. https://doi.org/10.1007/s00382-021-06079-y
Junquas, C., Takahashi, K., Condom, T., Espinoza, J.-C., Chavez, S., Sicart, J.-E., & Lebel, T. (2018). Understanding the influence of orography
on the precipitation diurnal cycle and the associated atmospheric processes in the central Andes. Climate Dynamics, 50(11–12), 3995–4017.
https://doi.org/10.1007/s00382-017-3858-8
Kendon, E. J., Prein, A. F., Senior, C. A., & Stirling, A. (2021). Challenges and outlook for convection-permitting climate modelling. Philosoph-
ical Transactions of the Royal Society A: Mathematical, Physical & Engineering Sciences, 379(2195), 20190547. https://doi.org/10.1098/
rsta.2019.0547
Laraque, A., Ronchail, J., Cochonneau, G., Pombosa, R., & Guyot, J. L. (2007). Heterogeneous distribution of rainfall and discharge regimes in
the Ecuadorian Amazon basin. Journal of Hydrometeorology, 8(6), 1364–1381. https://doi.org/10.1175/2007JHM784.1
Lavado-Casimiro, W., & Espinoza, J. C. (2014). Impactos de El Niño y La Niña en las lluvias del Perú (1965–2007). Revista Brasileira de Mete-
orologia, 29(2), 171–182. https://doi.org/10.1590/S0102-77862014000200003
Lavado-Casimiro, W. S., Felipe, O., Silvestre, E., & Bourrel, L. (2013). ENSO impact on hydrology in Peru. Advances in Geosciences, 33, 33–39.
https://doi.org/10.5194/adgeo-33-33-2013
Llopart, M., Reboita, M. S., & da Rocha, R. P. (2019). Assessment of multi-model climate projections of water resources over South America
CORDEX domain. Climate Dynamics, 54(1–2), 99–116. https://doi.org/10.1007/s00382-019-04990-z
Lucas-Picher, P., Laprise, R., & Winger, K. (2017). Evidence of added value in North American regional climate model hindcast simulations
using ever-increasing horizontal resolutions. Climate Dynamics, 48(7–8), 2611–2633. https://doi.org/10.1007/s00382-016-3227-z
Maraun, D., Truhetz, H., & Schaffer, A. (2021). Regional climate model biases, their dependence on synoptic circulation biases and the potential
for bias adjustment: A process-oriented evaluation of the Austrian regional climate projections. Journal of Geophysical Research: Atmos-
pheres, 126(6), e2020JD032824. https://doi.org/10.1029/2020JD032824
Marengo, J. A., & Espinoza, J. C. (2016). Extreme seasonal droughts and floods in Amazonia: Causes, trends and impacts. International Journal
of Climatology, 36(3), 1033–1050. https://doi.org/10.1002/joc.4420
Marengo, J. A., Souza, C. M., Thonicke, K., Burton, C., Halladay, K., Betts, R. A., etal. (2018). Changes in climate and land use over the Amazon
region: Current and future variability and trends. Frontiers in Earth Science, 6, 228. https://doi.org/10.3389/feart.2018.00228
Mascaro, G., Viola, F., & Deidda, R. (2018). Evaluation of precipitation from EURO-CORDEX regional climate simulations in a small-scale
Mediterranean site. Journal of Geophysical Research: Atmospheres, 123(3), 1604–1625. https://doi.org/10.1002/2017JD027463
Meehl, G. A., Covey, C., Delworth, T., Latif, M., McAvaney, B., Mitchell, J. F. B., etal. (2007). THE WCRP CMIP3 multimodel dataset: A new
era in climate change research. Bulletin of the American Meteorological Society, 88(9), 1383–1394. https://doi.org/10.1175/BAMS-88-9-1383
Menéndez, C., Zaninelli, P., Carril, A., & Sánchez, E. (2016). Hydrological cycle, temperature, and land surface-atmosphere interaction in the La
Plata basin during summer: Response to climate change. Climate Research, 68(2–3), 231–241. https://doi.org/10.3354/cr01373
Mesinger, F., Chou, S. C., Gomes, J. L., Jovic, D., Bastos, P., Bustamante, J. F., etal. (2012). An upgraded version of the Eta model. Meteorology
and Atmospheric Physics, 116(3–4), 63–79. https://doi.org/10.1007/s00703-012-0182-z
Moya-Álvarez, A. S., Martínez-Castro, D., Kumar, S., Estevan, R., & Silva, Y. (2019). Response of the WRF model to different resolutions in
the rainfall forecast over the complex Peruvian orography. Theoretical and Applied Climatology, 137(3), 2993–3007. https://doi.org/10.1007/
s00704-019-02782-3
Newell, F. L., Ausprey, I. J., & Robinson, S. K. (2022). Spatiotemporal climate variability in the Andes of northern Peru: Evaluation of
gridded datasets to describe cloud forest microclimate and local rainfall. International Journal of Climatology, 42(11), 5892–5915.
https://doi.org/10.1002/joc.7567
Nobre, C. A., Sampaio, G., Borma, L. S., Castilla-Rubio, J. C., Silva, J. S., & Cardoso, M. (2016). Land-use and climate change risks in the
Amazon and the need of a novel sustainable development paradigm. Proceedings of the National Academy of Sciences, 113(39), 10759–10768.
https://doi.org/10.1073/pnas.1605516113
Ochoa, A., Campozano, L., Sánchez, E., Gualán, R., & Samaniego, E. (2016). Evaluation of downscaled estimates of monthly temperature and
precipitation for a Southern Ecuador case study. International Journal of Climatology, 36(3), 1244–1255. https://doi.org/10.1002/joc.4418
Ometto, J. P., Kalaba, K., Anshari, G. Z., Chacón, N., Farrell, A., Halim, S. A., etal. (2022). Cross-chapter paper 7: Tropical forests. In H.-O.
Pörtner, D. C. Roberts, M. Tignor, E. S. Poloczanska, K. Mintenbeck, A. Alegría, etal. (Eds.), Climate change 2022: Impacts, adaptation and
vulnerability. Contribution of Working Group II to the sixth assessment report of the Intergovernmental Panel on Climate Change (pp.2369–
2410). Cambridge University Press. https://doi.org/10.1017/9781009325844.024
Ortega, G., Arias, P.A., Villegas, J. C., Marquet, P.A., & Nobre, P. (2021). Present-day and future climate over central and South America accord-
ing to CMIP5/CMIP6 models. International Journal of Climatology, 41(15), 6713–6735. https://doi.org/10.1002/joc.7221
Pabón-Caicedo, J. D., Arias, P.A., Carril, A. F., Espinoza, J. C., Borrel, L. F., Goubanova, K., etal. (2020). Observed and projected hydroclimate
changes in the Andes. Frontiers in Earth Science, 8, 61. https://doi.org/10.3389/feart.2020.00061
Journal of Geophysical Research: Atmospheres
GUTIERREZ ETAL.
10.1029/2023JD038618
23 of 23
Pörtner, H.-O., Roberts, D. C., Adams, H., Adelekan, I., Adler, C., Adrian, R., etal. (2022). Technical Summary. In H.-O. Pörtner, D. C. Roberts,
M. Tignor, E. S. Poloczanska, K. Mintenbeck, M. Tignor, etal. (Eds.), Climate change 2022: Impacts, adaptation and vulnerability. Contribu-
tion of Working Group II to the sixth assessment report of the Intergovernmental Panel on Climate Change (pp.37–118). Cambridge University
Press. https://doi.org/10.1017/9781009325844.002
Prein, A. F., Gobiet, A., Truhetz, H., Keuler, K., Goergen, K., Teichmann, C., etal. (2016). Precipitation in the EURO-CORDEX 0.11° and 0.44°
simulations: High resolution, high benefits? Climate Dynamics, 46(1–2), 383–412. https://doi.org/10.1007/s00382-015-2589-y
Reboita, M. S., da Rocha, R. P., de Souza, C. A., Baldoni, T. C., da Silva, P.L. L. S., & Ferreira, G. W. S. (2022). Future projections of extreme
precipitation climate indices over South America based on CORDEX-CORE multimodel ensemble. Atmosphere, 13(9), 1463. https://doi.
org/10.3390/atmos13091463
Rosales, A. G., Junquas, C., Da Rocha, R. P., Condom, T., & Espinoza, J.-C. (2022). Valley–mountain circulation associated with the diurnal cycle
of precipitation in the tropical Andes (Santa River Basin, Peru). Atmosphere, 13(2), 344. https://doi.org/10.3390/atmos13020344
Rotstayn, L. D., Collier, M. A., Dix, M. R., Feng, Y., Gordon, H. B., Ofarrell, S. P., etal. (2009). Improved simulation of Australian climate and
ENSO-related rainfall variability in a global climate model with an interactive aerosol treatment. International Journal of Climatology, 30(7),
1067–1088. https://doi.org/10.1002/joc.1952
Samuelsson, P., Gollvik, S., Jansson, C., Kupiainen, M., Kourzeneva, E., & van de Berg, W. M. (2015). The sur face processes of the Rossby
Centre regional atmospheric climate model (RCA4). SMHI. Retrieved from http://urn.kb.se/resolve?urn=urn:nbn:se:smhi:diva-2840
Samuelsson, P., Jones, C. G., Willén, U., Ullerstig, A., Gollvik, S., Hansson, U., etal. (2011). The Rossby Centre regional climate model RCA3: Model
description and performance. Tellus A: Dynamic Meteorology and Oceanography, 63(1), 4. https://doi.org/10.1111/j.1600-0870.2010.00478.x
Segura, H., Junquas, C., Espinoza, J. C., Vuille, M., Jauregui, Y. R., Rabatel, A., etal. (2019). New insights into the rainfall variability in the
tropical Andes on seasonal and interannual time scales. Climate Dynamics, 53(1–2), 405–426. https://doi.org/10.1007/s00382-018-4590-8
Sierra, J. P., Junquas, C., Espinoza, J. C., Segura, H., Condom, T., Andrade, M., etal. (2022). Deforestation impacts on Amazon-Andes hydrocli-
matic connectivity. Climate Dynamics, 58(9–10), 2609–2636. https://doi.org/10.1007/s00382-021-06025-y
Skamarock, W. C., Klemp, J. B., Dudhia, J., Gill, D. O., Barker, D. M., Duda, M. G., etal. (2008). A description of the advanced research WRF
version 3, note NCAR/TN-475+ STR. NCAR Tech Colorado. https://doi.org/10.5065/D68S4MVH
Solman, S. A., & Blázquez, J. (2019). Multiscale precipitation variability over South America: Analysis of the added value of CORDEX RCM
simulations. Climate Dynamics, 53(3–4), 1547–1565. https://doi.org/10.1007/s00382-019-04689-1
Staal, A., Tuinenburg, O. A., Bosmans, J. H. C., Holmgren, M., van Nes, E. H., Scheffer, M., etal. (2018). Forest-rainfall cascades buffer against
drought across the Amazon. Nature Climate Change, 8(6), 539–543. https://doi.org/10.1038/s41558-018-0177-y
Sulca, J., Takahashi, K., Espinoza, J.-C., Vuille, M., & Lavado-Casimiro, W. (2018). Impacts of different ENSO flavors and tropical Pacific
convection variability (ITCZ, SPCZ) on austral summer rainfall in South America, with a focus on Peru. International Journal of Climatology,
38(1), 420–435. https://doi.org/10.1002/joc.5185
Sulca, J. C., & da Rocha, R. P. (2021). Influence of the coupling South Atlantic Convergence Zone-El Niño-Southern Oscillation (SACZ-ENSO)
on the projected precipitation changes over the central Andes. Climate, 9(5), 77. https://doi.org/10.3390/cli9050077
Taylor, K. E. (2001). Summarizing multiple aspects of model performance in a single diagram. Journal of Geophysical Research, 106(D7),
7183–7192. https://doi.org/10.1029/2000JD900719
Taylor, K. E., Stouffer, R. J., & Meehl, G. A. (2012). An overview of CMIP5 and the experiment design. Bulletin of the American Meteorological
Society, 93(4), 485–498. https://doi.org/10.1175/BAMS-D-11-00094.1
Torma, C., Giorgi, F., & Coppola, E. (2015). Added value of regional climate modeling over areas characterized by complex terrain-precipitation
over the Alps. Journal of Geophysical Research: Atmospheres, 120(9), 3957–3972. https://doi.org/10.1002/2014JD022781
Valenzuela, J., Figueroa, M., Armijos, E., Espinoza, J.-C., Wongchuig, S., & Ramirez-Avila, J. J. (2023). Flooding risk of cropland areas by repi-
quetes in the western Amazon basin: A case study of Peruvian Tamshiyacu City. Journal of Hydrology: Regional Studies, 47, 101428. https://
doi.org/10.1016/j.ejrh.2023.101428
Vera, C., Higgins, W., Amador, J., Ambrizzi, T., Garreaud, R., Gochis, D., etal. (2006). Toward a unified view of the American Monsoon systems.
Journal of Climate, 19(20), 4977–5000. https://doi.org/10.1175/JCLI3896.1
Voldoire, A., Sanchez-Gomez, E., Salas y Mélia, D., Decharme, B., Cassou, C., Sénési, S., etal. (2013). The CNRM-CM5.1 global climate
model: Description and basic evaluation. Climate Dynamics, 40(9–10), 2091–2121. https://doi.org/10.1007/s00382-011-1259-y
von Salzen, K., Scinocca, J. F., McFarlane, N. A., Li, J., Cole, J. N. S., Plummer, D., etal. (2013). The Canadian fourth generation atmospheric
global climate model (CanAM4). Part I: Representation of physical processes. Atmosphere-Ocean, 51(1), 104–125. https://doi.org/10.1080/0
7055900.2012.755610
Vuille, M., Carey, M., Huggel, C., Buytaert, W., Rabatel, A., Jacobsen, D., etal. (2018). Rapid decline of snow and ice in the tropical Andes—
Impacts, uncertainties and challenges ahead. Earth-Science Reviews, 176, 195–213. https://doi.org/10.1016/j.earscirev.2017.09.019
Watanabe, M., Suzuki, T., Oishi, R., Komuro, Y., Watanabe, S., Emori, S., etal. (2010). Improved climate simulation by MIROC5: Mean states,
variability, and climate sensitivity. Journal of Climate, 23(23), 6312–6335. https://doi.org/10.1175/2010JCLI3679.1
Young, B., Young, K. R., & Josse, C. (2011). Vulnerability of tropical Andean ecosystems to climate change. In S. K. Herzog, R. Martínez, P.M.
Jorgensen, & H. Tiessen (Eds.), Climate change and biodiversity in the tropical Andes (pp.170–181). Inter American Institute for Global
Change Research (IAI)—Scientific Committee on Problems of the Environment (SCOPE).
Zanchettin, D., Rubino, A., Matei, D., Bothe, O., & Jungclaus, J. H. (2013). Multidecadal-to-centennial SST variability in the MPI-ESM simula-
tion ensemble for the last millennium. Climate Dynamics, 40(5–6), 1301–1318. https://doi.org/10.1007/s00382-012-1361-9
