Hydrological performance of gridded meteorological products in Peruvian Altiplano basins

Desempenho hidrológico de produtos meteorológicos em grade nas bacias do Altiplano peruano

ABSTRACT

Gridded meteorological datasets represent a valuable alternative for hydrometeorological applications, especially in areas with limited ground observations. However, it is important to evaluate these datasets to quantify their accuracy, error, and bias in estimates. The objective of this research was to assess the hydrological performance of gridded meteorological products in Peruvian Altiplano basins. Three evaluation approaches were employed: 1) pixel-to-point, 2) basin-averaged values, and 3) as forcings in hydrological modeling. Using precipitation (P) and potential evapotranspiration (PET) data from AgERA5, ERA5-Land, MERRA-2, and PERSIANN-CDR, comparisons were made with 33 local meteorological stations using statistical metrics such as correlation coefficient (CC), root mean square error (RMSE), and percentage bias (PBIAS). Additionally, the GR4J hydrological model was employed in four watersheds to assess model performance with different data combinations. Results showed that AgERA5 and ERA5-Land excelled in daily precipitation estimation, while MERRA-2 had the lowest PET bias. Using reference data for P and PET as model forcings yielded consistent results across basins. However, the performance declined when using gridded products for both P and PET, revealing significant limitations in replicating hydrological processes and emphasizing the need to enhance these products for areas with limited observational data.

Keywords:
AgERA5; Peruvian plateau; ERA5-Land; GR4J hydrological model; MERRA-2; PERSIANN-CDR; Titicaca hydrographic region

RESUMO

Os conjuntos de dados meteorológicos em grade representam uma alternativa valiosa para aplicações hidrometeorológicas, especialmente em áreas com observações terrestres limitadas. No entanto, é importante avaliar esses conjuntos de dados para quantificar sua precisão, erro e viés nas estimativas. O objetivo desta pesquisa foi avaliar o desempenho hidrológico de produtos meteorológicos em grade em bacias do altiplano peruano. Três abordagens de avaliação foram empregadas: 1) pixel-para-ponto, 2) valores médios sobre a bacia e 3) como forçantes no modelamento hidrológico. Utilizando dados de precipitação (P) e evapotranspiração potencial (PET) de AgERA5, ERA5-Land, MERRA-2 e PERSIANN-CDR, comparações foram feitas com 33 estações meteorológicas locais usando métricas estatísticas como coeficiente de correlação (CC), erro quadrático médio (RMSE) e viés percentual (PBIAS). Além disso, o modelo hidrológico GR4J foi aplicado em quatro bacias hidrográficas para avaliar o desempenho do modelo com diferentes combinações de dados. Os resultados mostraram que AgERA5 e ERA5-Land se destacaram na estimativa de precipitação diária, enquanto MERRA-2 teve o menor viés de PET. O uso de dados de referência para P e PET como forçantes do modelo gerou resultados consistentes nas bacias. No entanto, o desempenho diminuiu ao utilizar produtos quadriculados tanto para P quanto para PET, revelando limitações significativas na replicação dos processos hidrológicos e enfatizando a necessidade de aprimorar esses produtos para áreas com dados observacionais limitados.

Palavras-chave:
AgERA5; Planalto Peruano; ERA5-Land; Modelo hidrológico GR4J; MERRA-2; PERSIANN-CDR; Região hidrográfica do Titicaca

INTRODUCTION

Hydrological modeling requires accurate meteorological forcing data (Probst & Mauser, 2022). Variations in potential evapotranspiration (PET) and precipitation (P) influence the surface water balance and play a key role in defining moisture or dryness conditions in a region (Xu et al., 2024). Meteorological stations provide the best estimation of ground-based weather data at local and regional scales (Tarek et al., 2020) and are a very appropriate source of meteorological forcing data. However, due to complex topography in high-altitude areas and the uneven distribution of measurement sites, accurate meteorological data estimates are still lacking. Even in developed countries, well-equipped meteorological stations that record information with reliable instruments are scarce (Pelosi & Chirico, 2021). On the positive side, gridded products are essential tools for monitoring climate change, validating models, and detecting extreme climate events (Ge et al., 2023), but they require validation (Kumar et al., 2021).

In recent decades, there has been significant progress in the development of various global and quasi-global gridded meteorological products. This advancement has been achieved through the increasing use of remote sensing techniques and advanced computing capabilities (Rahmati Ziveh et al., 2022), including reanalysis products (Dee et al., 2011; Gelaro et al., 2017; Hersbach et al., 2020; Kobayashi et al., 2015; Muñoz-Sabater et al., 2021), satellite-based products, multi-source products, and interpolated products (Kumar et al., 2021; Rahmati Ziveh et al., 2022; Xu et al., 2022). There has also been a renewed interest in researching gridded meteorological products as an alternative source for various global hydrometeorological applications (Arshad et al., 2021; Cantoni et al., 2022; Gomis-Cebolla et al., 2023; Gruber et al., 2022; Pelosi & Chirico, 2021; Rahmati Ziveh et al., 2022; Tarek et al., 2020; X. Xu et al., 2022; Xue et al., 2019; Araghi et al., 2022; Kumar et al., 2021; Roffe & van der Walt, 2023; Wang et al., 2021; Hersbach et al., 2020). Although there are studies in the Peruvian Altiplano (PA) basins aimed at solving hydrological problems using gridded products (Asurza et al., 2018; Lujano et al., 2015; Zubieta et al., 2018), this study represents the first attempt to evaluate gridded P and PET products from AgERA5, ERA5-Land, MERRA-2, and PERSIANN-CDR as meteorological forcings to drive the GR4J hydrological model in PA basins.

The objective of this study was to assess the hydrological performance of gridded meteorological products in Peruvian highland basins. Two key research questions were addressed: 1) To what extent can gridded meteorological products capture daily patterns of precipitation and potential evapotranspiration compared to reference data from meteorological stations and basin-averaged values? and 2) How do gridded meteorological products perform in terms of river flows after being propagated through the GR4J hydrological model to simulate daily flows? The results of this research will significantly contribute to enriching the hydrometeorological database, especially in areas with a scarcity of ground-based observation data.

MATERIAL AND METHODS

Study area

The study area is located in the PA, in southern Peru (Figure 1). Four basins were selected based on data availability: the Ramis River Basin (RRB), the Ilave River Basin (IRB), the Coata River Basin (CRB), and the Huancane River Basin (HRB), with elevations ranging from 3,821 to 5,781 meters above sea level (m.a.s.l.).

Figure 1
Location of the study area.

For the period 2003-2016, the mean annual precipitation was approximately 673.5 mm in the RRB, 632.9 mm in the IRB, 706.5 mm in the CRB, and 647.5 mm in the HRB. On average, the highest precipitation accumulations for the four basins occur in the austral summer (59.7% of the annual precipitation), and the lowest occur in winter (2.0% of the annual precipitation). The transition period from wet to dry accounts for 22.0%, while the transition period from dry to wet accounts for 16.3% of the total annual precipitation. The mean annual potential evapotranspiration, obtained using the Hargreaves-Samani (HS) method (Hargreaves & Samani, 1985), was 1216.7 mm for the RRB, 1235.6 mm for the IRB, 1216.3 mm for the CRB, and 1187.8 mm for the HRB.

Data

Cartographic information

The Digital Elevation Model (DEM) was generated by the NASA Shuttle Radar Topography Mission (SRTM) at a spatial resolution of ~90 meters and was obtained from the Google Earth Engine (GEE) platform, image ID CGIAR/SRTM90_V4. GEE is a web-based platform for processing large satellite image data (Ghosh et al., 2022). For more details on the DEM, please refer to Jarvis et al. (2008).

Reference hydrometeorological data

Daily data on precipitation, air temperature, and flow were obtained from the Servicio Nacional de Meteorología e Hidrología (SENAMHI) Peru (Servicio Nacional de Meteorología e Hidrología, 2023) and the Autoridad Nacional del Agua (ANA) Peru (Autoridad Nacional del Agua, 2023). The period used spans from January 1, 2003, to December 31, 2016, for 33 meteorological stations (Table 1 and Figure 1) and 4 hydrometric stations (Table 2 and Figure 1).

Table 1
Characteristics of the meteorological stations and annual mean values of meteorological variables.

N°	Station	Lat.	Lon.	Alt.	P	%DR	Tmax	%DR	Tmin	%DR	Tmean	PET
1	Ananea	-14.7	-69.5	4653	646	100.0	10.4	100.0	-1.9	95.2	4.3	921.9
2	Arapa	-15.1	-70.1	3892	617	100.0	17.2	100.0	1.4	100.0	9.3	1269
3	Ayaviri	-14.9	-70.6	3921	695	99.4	17.7	99.1	-0.2	99.1	8.8	1321
4	Azángaro	-14.9	-70.2	3840	605	100.0	17.3	99.9	1	99.9	9.2	1282
5	Cabanillas	-15.6	-70.4	3877	706	100.0	17.2	100.0	2.4	100.0	9.8	1252
6	Capachica	-15.6	-69.8	3827	722	100.0	15.1	100.0	1.4	100.0	8.3	1128
7	Capazo	-17.2	-69.7	4470	437	100.0	13.8	98.5	-6.4	99.9	3.7	1133
8	Chuquibambilla	-14.8	-70.7	3899	775	100.0	16.4	99.9	-2.3	99.9	7.1	1248
9	Cojata	-15	-69.4	4325	668	100.0	12.6	99.5	-3.6	100.0	4.5	1038
10	Crucero	-14.4	-70	4131	517	100.0	14.8	100.0	-2.5	100.0	6.2	1158
11	Desaguadero	-16.6	-69	3813	681	100.0	15.2	100.0	1.9	100.0	8.6	1122
12	Huancane	-15.2	-69.8	3860	633	100.0	15.8	100.0	0.1	100.0	8	1191
13	Huaraya Moho	-15.4	-69.5	3818	806	98.8	15.4	98.8	1.9	98.8	8.7	1146
14	Ichuña	-16.1	-70.6	3890	547	99.4	20.3	99.9	2.1	98.1	11.2	1464
15	Ilave	-16.1	-69.7	3837	670	99.1	15.6	97.9	1.2	97.9	8.4	1163
16	Juli	-16.2	-69.5	3814	918	100.0	14.1	100.0	3	100.0	8.6	1032
17	Juliaca	-15.5	-70.2	3828	620	100.0	17.9	100.0	-0.6	100.0	8.7	1327
18	Lampa	-15.4	-70.4	3929	726	100.0	17.2	100.0	-0.5	100.0	8.4	1287
19	Laraqueri	-16.2	-70.1	3952	707	99.4	16.8	99.4	-2	99.4	7.4	1270
20	Llalli	-14.9	-70.9	4211	832	100.0	16.7	100.0	-0.1	100.0	8.3	1259
21	Macusani	-14.1	-70.4	4350	656	100.0	12.1	98.7	-2.5	100.0	4.8	1006
22	Mañazo	-15.8	-70.3	3910	640	100.0	16.9	100.0	1.1	100.0	9	1257
23	Mazocruz	-16.7	-69.7	3963	520	100.0	17.5	100.0	-6	100.0	5.8	1322
24	Muñani	-14.8	-70	3923	631	100.0	16.9	100.0	1.8	100.0	9.4	1249
25	Pampahuta	-15.5	-70.7	4292	778	100.0	13	99.9	-4.7	100.0	4.2	1082
26	Pizacoma	-16.9	-69.4	3913	568	100.0	17.3	99.7	-1.2	99.9	8.1	1303
27	Progreso	-14.7	-70.4	3952	625	100.0	16.7	100.0	0.8	100.0	8.8	1243
28	Pucara	-15	-70.4	3869	693	100.0	17.4	100.0	0	100.0	8.7	1289
29	Puno	-15.8	-70	3831	711	100.0	16.2	99.9	3.4	100.0	9.8	1165
30	Putina	-14.9	-69.9	3853	610	100.0	17.4	96.3	0	100.0	8.7	1297
31	Santa Rosa	-14.6	-70.8	3944	774	100.0	16.6	98.8	-0.1	98.8	8.3	1236
32	Taraco	-15.3	-70	3817	551	100.0	16.2	99.7	-0.4	100.0	7.9	1205
33	Yunguyo	-16.3	-69.1	3864	740	100.0	15	100.0	1.2	100.0	8.1	1141

• Note: Lat: latitude (º); Lon: longitude (º); Alt: altitude (m.a.s.l.); P: precipitation (mm); Tmax: maximum air temperature (ºC); Tmin: minimun air temperature (ºC); Tmean: mean air temperature (ºC); PET: potential evapotranspiration estimated by the Hargreaves-Samani method (mm); %DR: Percentage of data recorded.

Table 2
Characteristics of the hydrometric stations and annual mean flow values.

N°	Hydrometric station	Lat.	Long.	Alt.	Q	%DR	River	Basin	Area
1	Ramis bridge (RAM-B)	-15.3	-69.9	3821	71	99.8	Ramis	RRB	14,676.70
2	Ilave bridge (ILA-B)	-16.1	-69.6	3850	26	99.9	Ilave	IRB	7,771.50
3	Unocolla bridge (UNC-B)	-15.5	-70.2	3835	34	99.4	Coata	CRB	4,448.50
4	Huancane bridge (HNE-B)	-15.2	-69.8	3830	19	99.6	Huancane	HRB	3,530.90

• Note: Lat: latitude (º); Long: longitude (º); Alt: altitude (m.a.s.l.); Q: flow (m³/s); Area (km²); RRB: Ramis river basin; IRB: Ilave river basin; CRB: Coata river basin; HRB: Huancane river basin; %DR: Percentage of data recorded.

Gridded meteorological datasets

AgERA5

Daily estimates of P and PET were obtained from the dataset "Agrometeorological indicators from 1979 to present derived from reanalysis" (AgERA5) available on the Climate Engine research platform (Climate Engine, 2023). AgERA5 (Boogaard et al., 2020) is a gridded dataset providing surface meteorological information from 1979 to the present, designed as a fundamental resource for agricultural and agroecological research, e.g., (Araghi et al., 2022). It has a spatial resolution of 0.1° × 0.1° with a wide range of meteorological variables. AgERA5 was derived from the ERA5 global climate reanalysis (Hersbach et al., 2020) from the European Centre for Medium-Range Weather Forecasts (ECMWF). Reanalysis datasets are produced based on three main approaches (data assimilation method, forecasting model, and input datasets) and provide a comprehensive and consistent series of global atmospheric variables in time and space (Dee et al., 2011).

ERA5-Land

ERA5-Land is an enhanced global dataset for the land component of ERA5, produced by ECMWF as part of the European Commission's Copernicus Climate Change Service (C3S). ERA5-Land data is accessible through the Climate Engine research platform (Climate Engine, 2023). This dataset describes the evolution of water and energy cycles on land consistently from 1950 to the present, with continuous updates. ERA5-Land is characterized by its high spatial resolution of 9 km and hourly temporal resolution. It aims to provide a detailed description of the land's hydrological and energy cycle, making it valuable for hydrological studies, climate modeling, water resource management, and land and environmental applications (Muñoz-Sabater et al., 2021).

MERRA-2

MERRA-2 is NASA's modern-era atmospheric reanalysis, incorporating observations not available in its predecessor, MERRA, along with model updates and analysis schemes to provide a continuous climate analysis. MERRA-2 is characterized by a spatial resolution of ~50 km (0.5º × 0.625º) and hourly temporal resolution. Aiming to address previous limitations and advance towards an integrated Earth system analysis (IESA), MERRA-2 includes significant improvements such as aerosol observation assimilation, enhancements in the representation of the stratosphere and cryospheric processes, and reductions in spurious trends and biases in aspects of the water cycle (Gelaro et al., 2017). P and PET estimates from MERRA-2 are accessible on the Climate Engine research platform (Climate Engine, 2023).

PERSIANN-CDR

Precipitation Estimation from Remotely Sensed Information using Artificial Neural Networks-Climate Data Record (PERSIANN-CDR) provides daily precipitation estimates at a spatial resolution of 0.25° and aims to address the need for a consistent, long-term global precipitation dataset for studying changes and trends in precipitation, especially extreme precipitation events, due to climate change and natural variability. It is generated from the PERSIANN algorithm using historical IR data from Gridded Satellite B1 (GridSat-B1) and adjusted using the Global Precipitation Climatology Project (GPCP) monthly product (Ashouri et al., 2015; Nguyen et al., 2019; Sadeghi et al., 2021). Precipitation estimates from PERSIANN-CDR are accessible via the Center for Hydrometeorology and Remote Sensing (CHRS) data portal (Center for Hydrometeorology and Remote Sensing, 2023).

Hydrological model

The model selected for this research is GR4J (Perrin et al., 2003), a lumped precipitation-runoff model with a daily time step. This model requires the calibration of four parameters (Table 3). A detailed description of the equations and processes is explained in García-Hernández et al. (2020). Figure 2 presents the general scheme of the GR4J hydrological model.

Table 3
Parameters of the GR4J hydrological model.

Model parameter	Description	Unit
X1	Production store capacity	m
X2	Intercatchment exchange coefficient	m/d
X3	Routing store capacity	m
X4	Unit hydrograph time constant	d

Figure 2
Schematic of the GR4J hydrological model (from García-Hernández et al., 2020)

The only inputs required by the GR4J model are PET and P. The calibration and validation of the GR4J model parameters were conducted using routing system (RS) MINERVE v2.8.9 software (García-Hernández et al., 2020). The GR4J model was chosen due to its good performance in highland basins in Peru, as noted by Lujano et al. (2020).

Method

Evaluation of gridded meteorological products

A quality control process for the ground-based hydrometeorological data was conducted, which included verifying physical limits established for Peru, internal consistency, and spatial consistency, as indicated by Vera et al. (2021). Additionally, data homogeneity was verified through visual inspection and the non-parametric CUSUM test. For an unknown change point, CUSUM checks whether the means in two parts of a time series are different (Chiew & Siriwardena, 2005). Homogeneity was verified with monthly data after filling in missing data (Tomas-Burguera et al., 2019; Woldesenbet et al., 2017). Homogeneity tests are generally more robust when used with monthly data (Tomas-Burguera et al., 2019).

Missing daily data were filled using the random forests machine learning algorithm embedded in the MICE (Multivariate Imputation by Chained Equations) package for the R project (van Buuren & Groothuis-Oudshoorn, 2011). Daily data were aggregated to monthly, and data homogeneity was verified using the non-parametric CUSUM test with a 95% confidence level and a 5% significance level. The TREND program was used for this procedure, designed to facilitate statistical tests for analyzing trends, changes, and randomness in hydrometeorological data (Chiew & Siriwardena, 2005). All data series analyzed were found to be homogeneous for the period 2003-2016.

To evaluate the gridded meteorological products of P and PET, two approaches were used. The first was the pixel-to-point approach, where P and PET values were extracted from the pixel grid using the points of the meteorological stations. The second approach involved using mean values over the basin. For this latter approach, the basins were delineated using a Geographic Information System (GIS), with the hydrometric station as the point of interest and the DEM as the topographic information. Subsequently, mean values over the basin were obtained from ground observations using the IDW interpolation method. For the gridded meteorological products of P and PET, the mean values over the basins were obtained from the average of the pixels on the Climate Engine platform. It is important to note that PET at each meteorological station was calculated using the Hargreaves-Samani (HS) method (Hargreaves & Samani, 1985) from observed air temperature data. Similarly, the gridded meteorological products of PET used the same method for calculation.

$PET=0.0023\times \left(Tmed+17.78\right)\times Ro\times {\left(Tmax-Tmin\right)}^{0.5}$ (1)

where Ro is the extraterrestrial solar radiation in mm/day.

Subsequently, the results of the two approaches were compared with ground observations using performance statistics metrics that evaluated accuracy (correlation coefficient, r), error (root mean square error, RMSE), and overestimation/underestimation (percent bias, PBIAS) (Table 4). This process was carried out using the hydroGOF package (Zambrano-Bigiarini, 2020) available for the R programming environment.

Table 4
Statistical metrics for evaluating the performance of gridded meteorological products.

Metric	Ecuation	Unit	Optimal value
Correlation coefficient (r)	$r=\frac{{{\displaystyle \sum }}_{i=1}^n\left(O_i-\overline{O}\right)\left(S_i-\overline{S}\right)}{\sqrt{{{\displaystyle \sum }}_{i=1}^n{\left(O_i-\overline{O}\right)}^2}\sqrt{{{\displaystyle \sum }}_{i=1}^n{\left(S_i-\overline{S}\right)}^2}}$	-	1
Root mean square error (RMSE)	$RMSE=\sqrt{\frac{1}{n}{\displaystyle \sum }_{i=0}^n{\left(S_i-O_i\right)}^2}$	mm and m3/s	0
Percent bias (PBIAS)	$PBIAS=\frac{{{\displaystyle \sum }}_{i=1}^n\left(S_i-O_i\right)}{{{\displaystyle \sum }}_{i=1}^n{O}_i}\times 100$	%	0
Nash-Sutcliffe efficiency (NSE)	$\text{NSE}=1-\frac{{{\displaystyle \sum }}_{i=1}^n{\left(O_i-S_i\right)}^2}{{{\displaystyle \sum }}_{i=1}^n{\left(O_i-\overline{O}\right)}^2}$	-	1
Kling-Gupta Efficiency 2012 (KGE)	$KGE=1-\sqrt{{\left(r-1\right)}^2+{\left(\beta -1\right)}^2+{\left(\gamma -1\right)}^2}$ $\beta =\frac{\overline{S}}{\overline{O}};\gamma =\frac{\raisebox{1ex}{${\sigma }_S$}\!\left/ \!\raisebox{-1ex}{$\overline{S}$}\right.}{\raisebox{1ex}{${\sigma }_O$}\!\left/ \!\raisebox{-1ex}{$\overline{O}$}\right.}$	-	1

• Note: $S$ is the simulated or estimated value; $O$ is the observed value or reference; $\overline{S}$ and $\overline{O}$ denote the mean values of $S$ and $O$, respectively; $n$ indicates the number of data pairs; $r$ is the correlation coefficient between $S$ y $O$ (dimensionless); σ standard deviation; $\beta$ is the bias ratio (dimensionless); $\gamma$ is the variability ratio (dimensionless).

Evaluation of gridded meteorological products for hydrological modeling

Hydrological modeling was performed with GR4J using various combinations of P and PET data, including both ground observations and gridded meteorological products (Table 5), with measured flows at the hydrometric stations serving as a reference. The input data set was then divided into 70% for the calibration phase (january 1, 2003 to december 31, 2011) and 30% for the validation phase (january 1, 2012 to december 31, 2016).

Table 5
Combinations of P and PET input data.

Experiment	P	PET
1	Reference	Reference
2	AgERA5	AgERA5
3	ERA5-Land	ERA5-Land
4	MERRA-2	MERRA-2
5	PERSIANN-CDR	Reference
6	AgERA5	Reference
7	ERA5-Land	Reference
8	MERRA-2	Reference
9	Reference	AgERA5
10	Reference	ERA5-Land
11	Reference	MERRA-2

In hydrological model calibration, objective functions such as the NSE Efficiency (Nash & Sutcliffe, 1970) or the KGE Efficiency score (Gupta et al., 2009; Kling et al., 2012) are commonly used. In this study, NSE for peak flows and logarithmic Nash-Sutcliffe efficiency (NSE-ln) for low flows were used as the primary objective function for calibrating the GR4J model, along with three additional performance metrics: CC, KGE, and PBIAS. It is important to note that no data were considered for the warm-up period. The criteria for interpreting the GR4J model outputs based on NSE were classified into four categories: "Very Good" when NSE > 0.80, "Good" (0.70 < NSE ≤ 0.80), "Satisfactory" (0.50 < NSE ≤ 0.70), and "Unsatisfactory" (NSE ≤ 0.50). For PBIAS, the categories were "Very Good" when PBIAS < ±5%, "Good" (±5% < PBIAS ≤ ±10%), "Satisfactory" (±10% < PBIAS ≤ ±15%), and "Unsatisfactory" (PBIAS ≥ ±15%) (Moriasi et al., 2015). The hydrological performance can be classified using KGE as follows: good (KGE ≥0.75), intermediate (0.75 > KGE ≥ 0.5), poor (0.5 > KGE > 0.0) and very poor (KGE≤ 0.0) (Yuan et al., 2020). Figure 3 shows the methodological flowchart used to evaluate gridded meteorological products for hydrological modeling in the PA basins.

Figure 3
Methodological flow diagram for evaluating gridded meteorological products for hydrological modeling in PA basins.

RESULTS AND DISCUSSION

Evaluation of gridded meteorological products

Evaluation by station points

This research evaluates the ability of gridded precipitation products to detect daily precipitation patterns in comparison to reference station data. The spatial maps display the performance metrics, including CC, RMSE, and PBIAS, for four products (AgERA5, ERA5-Land, MERRA-2, and PERSIANN-CDR) (Figure 4). Overall, all four products exhibit relatively low CC values, with AgERA5 showing the highest values, followed by ERA5-Land, PERSIANN-CDR, and MERRA-2 in descending order of performance. These findings reveal a discrepancy between gridded precipitation estimates and daily reference data, highlighting the need to improve the accuracy of gridded products.

Figure 4
Spatial distribution of performance metrics for precipitation estimation by gridded meteorological products.

Regarding RMSE, MERRA-2 and AgERA5 show lower values, indicating less dispersion between the estimated and reference data, whereas PERSIANN-CDR and ERA5-Land exhibit higher variability, particularly in the northern part of the study area. These variations might be related to climatic characteristics and complex topography (Arshad et al., 2021).

Concerning PBIAS, most products tend to overestimate precipitation, with the exception of MERRA-2, which shows a tendency towards underestimation. A notable finding to consider is that daily scale evaluation could lead to less accurate results for precipitation products, as noted in previous research (Rahmati Ziveh et al., 2022). This consideration is crucial for future hydrological research, as it allows for the appropriate selection of the optimal temporal scale for analysis, which can improve result accuracy and its applicability in water resource management.

In this section, the spatial distribution of statistical performance metrics for PET estimation by three gridded meteorological products (AgERA5, ERA5-Land, and MERRA-2) is presented (Figure 5). The results reveal that, in terms of CC, AgERA5 exhibits the highest values, followed by ERA5-Land and MERRA-2 in descending order of performance. This suggests that AgERA5 has a better ability to reproduce reference PET compared to the other two products. Regarding the RMSE, MERRA-2 shows the lowest values, indicating a smaller discrepancy between reference and estimated data. Conversely, ERA5-Land exhibits the highest RMSE values, suggesting greater variability between reference and estimated data. Concerning PBIAS, MERRA-2 stands out with values close to zero, indicating a tendency to avoid bias in PET estimates. However, all three products show negative PBIAS values, suggesting an underestimation in PET estimates compared to reference data.

Figure 5
Spatial distribution of performance metrics for PET estimation by gridded meteorological products.

Evaluation by basin-averaged values

Figure 6 presents a series of scatter plots showing the performance metrics of gridded meteorological products (AgERA5, ERA5-Land, MERRA-2, and PERSIANN-CDR) across four river basins (RRB, IRB, CRB, and HRB). Overall, AgERA5 shows the strongest correlation with reference data, followed by ERA5-Land, MERRA-2, and PERSIANN-CDR, indicating higher accuracy in daily precipitation estimation for these basins. It is observed that, in most cases, both AgERA5 and ERA5-Land exhibit lower RMSE values compared to MERRA-2 and PERSIANN-CDR, suggesting less variation between estimated and reference data. This supports the notion that AgERA5 and ERA5-Land perform better in estimating daily precipitation in these basins compared to MERRA-2 and PERSIANN-CDR. While most products (AgERA5, ERA5-Land, and PERSIANN-CDR) tend to overestimate precipitation (positive PBIAS), MERRA-2 shows a negative PBIAS, indicating a tendency to underestimate precipitation estimates.

Figure 6
Scatter plots of gridded meteorological products versus average precipitation across various river basins.

Figure 7 shows a comparison between PET estimates obtained from three gridded meteorological products (AgERA5, ERA5-Land, and MERRA-2) and reference PET values for four river basins (RRB, IRB, CRB, and HRB). Overall, all products exhibit high correlation (CC close to 1), with AgERA5 being the most accurate, followed closely by ERA5-Land, while MERRA-2 shows the lowest RMSE values. Although all products tend to underestimate PET (negative PBIAS), MERRA-2 presents a PBIAS closer to zero, indicating a lower bias in the estimates. Collectively, these results suggest that MERRA-2 offers the best accuracy in estimating PET across the different river basins considered. It is important to note that a high CC score does not always guarantee high performance (Rahmati Ziveh et al., 2022), and bias correction could significantly reduce uncertainty in PET product estimates.

Figure 7
Scatter plots of gridded meteorological products versus mean PET across various river basins.

Evaluation of gridded meteorological products for hydrological modeling

This analysis evaluates the performance of gridded meteorological products in hydrological modeling in the study region. Our findings partially align with the results of Satgé et al. (2019), particularly regarding the variability observed in the PERSIANN-CDR product. However, contrary to the proposal by Satgé et al. (2019), who suggest using a 10-day timescale, our choice is based on daily data and a longer time series. Additionally, the studies by Asurza et al. (2018), Lujano et al. (2015), and Zubieta et al. (2018) provide valuable context on the applicability of other products in the PA.

Figure 8 presents the results of hydrological modeling with GR4J for four river basins during the calibration and validation periods. Several performance metrics are shown, including CC, NSE, KGE, and PBIAS, evaluated in different experiments that incorporate various input data combinations. The results indicate significant variability in model performance between experiments and stations, with metric values showing both good and poor capabilities of the model to reproduce reference flows.

Figure 8
Comparison of the GR4J hydrological model performance in four basins: Calibration and validation results with different input data combinations.

For instance, in Experiment 1 (reference forcing), performance ranged from good to very good during the calibration period (NSE between 0.73 and 0.84), and from satisfactory to very good during the validation period (NSE between 0.69 and 0.83). In Experiments 2 through 8 (gridded P + reference/gridded PET), the performance was unsatisfactory in the calibration period, while in the validation period, most experiments were also unsatisfactory, except for Experiments 4 and 8, which showed satisfactory performance in specific hydrometric stations.

Simulations from Experiments 9, 10, and 11(reference P + gridded PET) resulted in satisfactory to very good performance in both calibration and validation periods, except for Experiment 9 and 10 at ILA-B, which were unsatisfactory during validation. These ratings align with the criteria established by Moriasi et al. (2015). Similarly, the KGE performance was consistent for most simulations, both during calibration and validation. However, the best results were obtained using input forcings from Experiment 1, followed by Experiment 11 for the four basins, suggesting that the PET from MERRA-2 is a reliable forcing, while the experiments using PET from AgERA5 and ERA5-Land showed variability depending on the basin.

PBIAS results show a variety of values, some positive and some negative, indicating some bias in flow simulations for certain experiments. However, in the calibration period, CC values ranged between 0.69 and 0.92 for most experiments, except for Experiments 4 and 8 (CC varied between 0.32 and 0.56). During the validation period, most experiments showed CC values between 0.71 and 0.94, except for Experiments 4, 5, and 8 (CC ranged from 0.27 to 0.79). This could be attributed to the marked seasonality of flow, a common characteristic in semi-arid basins where water flows vary significantly throughout the year (Cantoni et al., 2022). These findings highlight the importance of considering different configurations and input data to evaluate the accuracy of hydrological modeling.

The uncertainty in the ILA-B, UNC-B, and HNE-B flows can be attributed to the lack of meteorological stations at high altitudes. To reduce this uncertainty, Wang et al. (2021) recommend the implementation of more stations in these areas.

These results demonstrate that precipitation data, in particular, is more important than PET. Satisfactory simulations could even be obtained using reference precipitation data, and PET from gridded meteorological products can be a reliable alternative, provided that flow data are available to calibrate the hydrological model and transfer information to poorly or unmonitored basins.

Figures 9 and 10 show the hydrographs of daily flow and flow duration curves for four hydrometric stations (RAM-B, ILA-B, UNC-B, and HNE-B) corresponding to the calibration and validation periods, respectively. In general, when reference data of P and PET were used to calibrate the model, the simulated flows reasonably reproduced the reference flows in several sections for the four basins, both during the calibration and validation periods. However, when gridded precipitation data were used, significant discrepancies were observed in the simulated flow, suggesting that gridded data significantly impact the model's performance, especially when the same parameters derived from the calibration with reference data are used. These differences highlight the model's sensitivity to input data sources, particularly in flow estimation.

Figure 9
Hydrographs of daily flows and flow duration curves for the four hydrometric stations under different experimental conditions during the calibration period.

Figure 10
Hydrographs of daily flows and flow duration curves for the four hydrometric stations under different experimental conditions during the validation period.

CONCLUSIONS

This study aimed to evaluate the hydrological performance of gridded meteorological products in basins of PA using the GR4J hydrological model. The conclusions are as follows:

1. The evaluation showed better results when comparing mean values over the basin. In point evaluations of meteorological stations, AgERA5 showed the highest CC for daily precipitation, followed by ERA5-Land, PERSIANN-CDR, and MERRA-2, indicating variations in spatial accuracy. MERRA-2 exhibited the lowest RMSE, indicating less discrepancy between reference and estimated data, though with a tendency to underestimate. For PET, AgERA5 and ERA5-Land were notable for their good reproduction capacity, while MERRA-2 showed the lowest RMSE, reflecting less variability in estimates. When evaluating with mean values over the basins, AgERA5 showed the highest correlations in the RRB, IRB, CRB, and HRB basins for precipitation, though with a significant bias. For PET, AgERA5 was the most accurate, although all products underestimated this variable, with MERRA-2 having a bias closer to zero.

2. The GR4J hydrological model was calibrated and validated using reference data for P and PET. The best results were obtained with both P and PET from reference data. Additionally, combinations where P was from reference data and PET came from AgERA5, ERA5-Land, or MERRA-2 also showed good performance, with the combination of reference P and MERRA-2 PET being the most effective. However, performance was poor when gridded products were used for both P and PET, or when the model was forced with gridded P and reference PET. Since the best simulations were achieved using reference P and PET from MERRA-2, AgERA5, or ERA5-Land, this combination could be used when reference data is not available.

These findings provide a better understanding of the performance of P and PET products from AgERA5, ERA5-Land, MERRA-2, and PERSIANN-CDR in the PA basins. The research includes a comprehensive assessment of the accuracy of gridded meteorological products, both in terms of P and PET, including their application as forcings in hydrological modeling in four river basins. However, significant limitations were identified, such as bias in P and PET estimates and spatial variability in product performance. For future research, it is recommended to explore advanced bias correction methods for gridded meteorological products, as well as to improve their spatial resolution. Additionally, integrating complementary data, such as satellite, reanalysis, multi-source, and observational data, is suggested to enhance the accuracy and applicability of these products in various hydrological and climatic contexts.

ACKNOWLEDGEMENTS

The authors wish to thank the Servicio Nacional de Meteorología e Hidrología of Peru and the Autoridad Nacional del Agua of Peru for providing the hydrometeorological data from the meteorological stations located in the PA region. We also thank AgERA5, ERA5-Land, MERRA-2, and PERSIANN-CDR for making precipitation and evapotranspiration data available for this study.

REFERENCES

• Araghi, A., Martinez, C. J., Olesen, J. E., & Hoogenboom, G. (2022). Assessment of nine gridded temperature data for modeling of wheat production systems. Computers and Electronics in Agriculture, 199, 107189. http://doi.org/10.1016/j.compag.2022.107189
  » http://doi.org/10.1016/j.compag.2022.107189
• Arshad, M., Ma, X., Yin, J., Ullah, W., Liu, M., & Ullah, I. (2021). Performance evaluation of ERA-5, JRA-55, MERRA-2, and CFS-2 reanalysis datasets, over diverse climate regions of Pakistan. Weather and Climate Extremes, 33, 100373. http://doi.org/10.1016/j.wace.2021.100373
  » http://doi.org/10.1016/j.wace.2021.100373
• Ashouri, H., Hsu, K. L., Sorooshian, S., Braithwaite, D. K., Knapp, K. R., Cecil, L. D., & Prat, O. P. (2015). PERSIANN-CDR: daily precipitation climate data record from multisatellite observations for hydrological and climate studies. Bulletin of the American Meteorological Society, 96(1), 69-83. http://doi.org/10.1175/BAMS-D-13-00068.1
  » http://doi.org/10.1175/BAMS-D-13-00068.1
• Asurza, F., Ramos, C., & Lavado, W. (2018). Assessment of Tropical Rainfall Measuring Mission (TRMM) and Global Precipitation Measurement (GPM) products in hydrological modeling of the Huancane river basin, Peru. Scientia Agropecuaria, 9(1), 53-62. http://doi.org/10.17268/sci.agropecu.2018.01.06
  » http://doi.org/10.17268/sci.agropecu.2018.01.06
• Autoridad Nacional del Agua – ANA. (2023). Retrieved in 2023, June 10, from https://snirh.ana.gob.pe/observatorioSNIRH/
  » https://snirh.ana.gob.pe/observatorioSNIRH/
• Boogaard, H., Schubert, J., De Wit, A., Lazebnik, J., Hutjes, R., & Van der Grijn, G. (2020). Agrometeorological indicators from 1979 to present derived from reanalysis, version 1.0 Copernicus Climate Change Service (C3S) Climate Data Store (CDS). http://doi.org/10.24381/cds.6c68c9bb
  » http://doi.org/10.24381/cds.6c68c9bb
• Cantoni, E., Tramblay, Y., Grimaldi, S., Salamon, P., Dakhlaoui, H., Dezetter, A., & Thiemig, V. (2022). Hydrological performance of the ERA5 reanalysis for flood modeling in Tunisia with the LISFLOOD and GR4J models. Journal of Hydrology. Regional Studies, 42, 101169. http://doi.org/10.1016/j.ejrh.2022.101169
  » http://doi.org/10.1016/j.ejrh.2022.101169
• Center for Hydrometeorology and Remote Sensing – CHRS. (2023). Retrieved in 2023, May 17, from https://chrsdata.eng.uci.edu/
  » https://chrsdata.eng.uci.edu/
• Chiew, F., & Siriwardena, L. (2005). TREND trend/change detection software: user guide Melbourne, Australia: CRC for Catchment Hydrology.
• Climate Engine. (2023). Retrieved in 2023, June 26, from https://app.climateengine.com/climateEngine
  » https://app.climateengine.com/climateEngine
• Dee, D. P., Uppala, S. M., Simmons, A. J., Berrisford, P., Poli, P., Kobayashi, S., Andrae, U., Balmaseda, M. A., Balsamo, G., Bauer, P., Bechtold, P., Beljaars, A. C. M., van de Berg, L., Bidlot, J., Bormann, N., Delsol, C., Dragani, R., Fuentes, M., Geer, A. J., Haimberger, L., Healy, S. B., Hersbach, H., Hólm, E. V., Isaksen, L., Kållberg, P., Köhler, M., Matricardi, M., McNally, A. P., Monge-Sanz, B. M., Morcrette, J.-J., Park, B.-K., Peubey, C., de Rosnay, P., Tavolato, C., Thépaut, J.-N., & Vitart, F. (2011). The ERA-Interim reanalysis: configuration and performance of the data assimilation system. Quarterly Journal of the Royal Meteorological Society, 137(656), 553-597. http://doi.org/10.1002/qj.828
  » http://doi.org/10.1002/qj.828
• García-Hernández, J., Foehn, A., Fluixá-Sanmartín, J., Roquier, B., Brauchli, T., Paredes-Arquiola, J., & De Cesare, G. (2020). RS MINERVE – Technical manual, v2.25. Switzerland: Ed. CREALP.
• Ge, S., Jiang, C., Wang, J., & Liu, S. (2023). Analyzing temperature and precipitation extremes in China using multiple gridded datasets: A comparative evaluation. Weather and Climate Extremes, 42, 100614. http://doi.org/10.1016/j.wace.2023.100614
  » http://doi.org/10.1016/j.wace.2023.100614
• Gelaro, R., McCarty, W., Suárez, M. J., Todling, R., Molod, A., Takacs, L., Randles, C. A., Darmenov, A., Bosilovich, M. G., Reichle, R., Wargan, K., Coy, L., Cullather, R., Draper, C., Akella, S., Buchard, V., Conaty, A., da Silva, A. M., Gu, W., Kim, G.-K., Koster, R., Lucchesi, R., Merkova, D., Nielsen, J. E., Partyka, G., Pawson, S., Putman, W., Rienecker, M., Schubert, S. D., Sienkiewicz, M., & Zhao, B. (2017). The Modern-Era Retrospective Analysis for Research and Applications, Version 2 (MERRA-2). Journal of Climate, 30(14), 5419-5454. http://doi.org/10.1175/JCLI-D-16-0758.1
  » http://doi.org/10.1175/JCLI-D-16-0758.1
• Ghosh, S., Kumar, D., & Kumari, R. (2022). Google earth engine based computational system for the earth and environment monitoring applications during the COVID-19 pandemic using thresholding technique on SAR datasets. Physics and Chemistry of the Earth Parts A/B/C, 127, 103163. http://doi.org/10.1016/j.pce.2022.103163
  » http://doi.org/10.1016/j.pce.2022.103163
• Gomis-Cebolla, J., Rattayova, V., Salazar-Galán, S., & Francés, F. (2023). Evaluation of ERA5 and ERA5-Land reanalysis precipitation datasets over Spain (1951-2020). Atmospheric Research, 284, 106606. http://doi.org/10.1016/j.atmosres.2023.106606
  » http://doi.org/10.1016/j.atmosres.2023.106606
• Gruber, K., Regner, P., Wehrle, S., Zeyringer, M., & Schmidt, J. (2022). Towards global validation of wind power simulations: A multi-country assessment of wind power simulation from MERRA-2 and ERA-5 reanalyses bias-corrected with the global wind atlas. Energy, 238, 121520. http://doi.org/10.1016/j.energy.2021.121520
  » http://doi.org/10.1016/j.energy.2021.121520
• Gupta, H. V., Kling, H., Yilmaz, K. K., & Martinez, G. F. (2009). Decomposition of the mean squared error and NSE performance criteria: implications for improving hydrological modelling. Journal of Hydrology (Amsterdam), 377(1–2), 80-91. http://doi.org/10.1016/j.jhydrol.2009.08.003
  » http://doi.org/10.1016/j.jhydrol.2009.08.003
• Hargreaves, G. H., & Samani, Z. A. (1985). Reference crop evapotranspiration from temperature. Applied Engineering in Agriculture, 1(2), 96-99. http://doi.org/10.13031/2013.26773
  » http://doi.org/10.13031/2013.26773
• Hersbach, H., Bell, B., Berrisford, P., Hirahara, S., Horányi, A., Muñoz-Sabater, J., Nicolas, J., Peubey, C., Radu, R., Schepers, D., Simmons, A., Soci, C., Abdalla, S., Abellan, X., Balsamo, G., Bechtold, P., Biavati, G., Bidlot, J., Bonavita, M., De Chiara, G., Dahlgren, P., Dee, D., Diamantakis, M., Dragani, R., Flemming, J., Forbes, R., Fuentes, M., Geer, A., Haimberger, L., Healy, S., Hogan, R. J., Hólm, E., Janisková, M., Keeley, S., Laloyaux, P., Lopez, P., Lupu, C., Radnoti, G., de Rosnay, P., Rozum, I., Vamborg, F., Villaume, S., & Thépaut, J.-N. (2020). The ERA5 global reanalysis. Quarterly Journal of the Royal Meteorological Society, 146(730), 1999-2049. http://doi.org/10.1002/qj.3803
  » http://doi.org/10.1002/qj.3803
• Jarvis, A., Reuter, H.I., Nelson, A., & Guevara, E. (2008). Hole-Filled SRTM for the Globe Version 4. CGIAR-CSI SRTM 90m Database Retrieved in 2023, June 26, from http://srtm.csi.cgiar.org
  » http://srtm.csi.cgiar.org
• Kling, H., Fuchs, M., & Paulin, M. (2012). Runoff conditions in the upper Danube basin under an ensemble of climate change scenarios. Journal of Hydrology (Amsterdam), 424–425, 264-277. http://doi.org/10.1016/j.jhydrol.2012.01.011
  » http://doi.org/10.1016/j.jhydrol.2012.01.011
• Kobayashi, S., Ota, Y., Harada, Y., Ebita, A., Moriya, M., Onoda, H., Onogi, K., Kamahori, H., Kobayashi, C., Endo, H., Miyaoka, K., & Kiyotoshi, T. (2015). The JRA-55 reanalysis: general specifications and basic characteristics. Journal of the Meteorological Society of Japan. Ser. II, 93(1), 5-48. http://dx.doi.org/10.2151/JMSJ.2015-001
  » http://dx.doi.org/10.2151/JMSJ.2015-001
• Kumar, M., Hodnebrog, Ø., Sophie Daloz, A., Sen, S., Badiger, S., & Krishnaswamy, J. (2021). Measuring precipitation in Eastern Himalaya: ground validation of eleven satellite, model and gauge interpolated gridded products. Journal of Hydrology (Amsterdam), 599, 126252. http://doi.org/10.1016/j.jhydrol.2021.126252
  » http://doi.org/10.1016/j.jhydrol.2021.126252
• Lujano, E., Felipe, O. G., Lujano, A., & Quispe, J. P. (2015). Validación de la precipitación estimada por satélite TRMM y su aplicación en la modelación hidrológica del rio Ramis Puno Perú. Revista de Investigaciones Altoandinas, 17(2), 221-228.
• Lujano, E., Sosa, J. D., Lujano, R., & Lujano, A. (2020). Evaluación del desempeño de modelos hidrológicos GR4J, HBV y SOCONT para el pronóstico de caudales medios diarios en la cuenca del río Ramis, Perú. Revista Ingeniería UC, 27(2), 189-199.
• Moriasi, D. N., Gitau, M. W., Pai, N., & Daggupati, P. (2015). Hydrologic and water quality models: performance measures and evaluation criteria. Transactions of the ASABE, 58(6), 1763-1785. http://doi.org/10.13031/trans.58.10715
  » http://doi.org/10.13031/trans.58.10715
• Muñoz-Sabater, J., Dutra, E., Agustí-Panareda, A., Albergel, C., Arduini, G., Balsamo, G., Boussetta, S., Choulga, M., Harrigan, S., Hersbach, H., Martens, B., Miralles, D. G., Piles, M., Rodríguez-Fernández, N. J., Zsoter, E., Buontempo, C., & Thépaut, J. N. (2021). ERA5-Land: a state-of-the-art global reanalysis dataset for land applications. Earth System Science Data, 13(9), 4349-4383. http://doi.org/10.5194/essd-13-4349-2021
  » http://doi.org/10.5194/essd-13-4349-2021
• Nash, J. E., & Sutcliffe, J. (1970). River flow forecasting through conceptual models part I: a discussion of principles. Journal of Hydrology (Amsterdam), 10(3), 282-290. http://doi.org/10.1016/0022-1694(70)90255-6
  » http://doi.org/10.1016/0022-1694(70)90255-6
• Nguyen, P., Shearer, E. J., Tran, H., Ombadi, M., Hayatbini, N., Palacios, T., Huynh, P., Braithwaite, D., Updegraff, G., Hsu, K., Kuligowski, B., Logan, W. S., & Sorooshian, S. (2019). The CHRS Data Portal, an easily accessible public repository for PERSIANN global satellite precipitation data. Scientific Data, 6(1), 180296. PMid:30620343. http://doi.org/10.1038/sdata.2018.296
  » http://doi.org/10.1038/sdata.2018.296
• Pelosi, A., & Chirico, G. B. (2021). Regional assessment of daily reference evapotranspiration: can ground observations be replaced by blending ERA5-Land meteorological reanalysis and CM-SAF satellite-based radiation data? Agricultural Water Management, 258, 107169. http://doi.org/10.1016/j.agwat.2021.107169
  » http://doi.org/10.1016/j.agwat.2021.107169
• Perrin, C., Michel, C., & Andréassian, V. (2003). Improvement of a parsimonious model for streamflow simulation. Journal of Hydrology (Amsterdam), 279(1-4), 275-289. http://doi.org/10.1016/S0022-1694(03)00225-7
  » http://doi.org/10.1016/S0022-1694(03)00225-7
• Probst, E., & Mauser, W. (2022). Evaluation of ERA5 and WFDE5 forcing data for hydrological modelling and the impact of bias correction with regional climatologies: A case study in the Danube River Basin. Journal of Hydrology. Regional Studies, 40, 101023. http://doi.org/10.1016/j.ejrh.2022.101023
  » http://doi.org/10.1016/j.ejrh.2022.101023
• Rahmati Ziveh, A., Bakhtar, A., Shayeghi, A., Kalantari, Z., Bavani, A. M., & Ghajarnia, N. (2022). Spatio-temporal performance evaluation of 14 global precipitation estimation products across river basins in southwest Iran. Journal of Hydrology. Regional Studies, 44, 101269. http://doi.org/10.1016/j.ejrh.2022.101269
  » http://doi.org/10.1016/j.ejrh.2022.101269
• Roffe, S. J., & van der Walt, A. J. (2023). Representation and evaluation of southern Africa’s seasonal mean and extreme temperatures in the ERA5-based reanalysis products. Atmospheric Research, 284, 106591. http://doi.org/10.1016/j.atmosres.2022.106591
  » http://doi.org/10.1016/j.atmosres.2022.106591
• Sadeghi, M., Nguyen, P., Naeini, M. R., Hsu, K., Braithwaite, D., & Sorooshian, S. (2021). PERSIANN-CCS-CDR, a 3-hourly 0.04 global precipitation climate data record for heavy precipitation studies. Scientific Data, 8(1), 157. PMid:34162874. http://doi.org/10.1038/s41597-021-00940-9
  » http://doi.org/10.1038/s41597-021-00940-9
• Satgé, F., Ruelland, D., Bonnet, M. P., Molina, J., & Pillco, R. (2019). Consistency of satellite-based precipitation products in space and over time compared with gauge observations and snow-hydrological modelling in the Lake Titicaca region. Hydrology and Earth System Sciences, 23(1), 595-619. http://doi.org/10.5194/hess-23-595-2019
  » http://doi.org/10.5194/hess-23-595-2019
• Servicio Nacional de Meteorología e Hidrología – SENAMHI. (2023). Retrieved in 2023, June 10, from https://www.senamhi.gob.pe/?&p=descarga-datos-hidrometeorologicos
  » https://www.senamhi.gob.pe/?&p=descarga-datos-hidrometeorologicos
• Tarek, M., Brissette, F. P., & Arsenault, R. (2020). Evaluation of the ERA5 reanalysis as a potential reference dataset for hydrological modelling over North America. Hydrology and Earth System Sciences, 24(5), 2527-2544. http://doi.org/10.5194/hess-24-2527-2020
  » http://doi.org/10.5194/hess-24-2527-2020
• Tomas-Burguera, M., Vicente-Serrano, S. M., Beguería, S., Reig, F., & Latorre, B. (2019). Reference crop evapotranspiration database in Spain (1961–2014). Earth System Science Data, 11(4), 1917-1930. http://doi.org/10.5194/essd-11-1917-2019
  » http://doi.org/10.5194/essd-11-1917-2019
• van Buuren, S., & Groothuis-Oudshoorn, K. (2011). mice: multivariate imputation by chained equations in R. Journal of Statistical Software, 45(3), 1-67. http://doi.org/10.18637/jss.v045.i03
  » http://doi.org/10.18637/jss.v045.i03
• Vera, L., Villegas, E., Oria, C., & Arboleda, F. (2021). Control de calidad de datos de estaciones meteorológicas e hidrológicas automáticas en el centro de procesamiento de datos del SENAMHI Peru: Servicio Nacional de Meteorología e Hidrología del Perú (SENAMHI).
• Wang, N., Liu, W., Wang, H., Sun, F., Duan, W., Li, Z., Li, Z., & Chen, Y. (2021). Improving streamflow and flood simulations in three headwater catchments of the Tarim River based on a coupled glacier-hydrological model. Journal of Hydrology (Amsterdam), 603, 127048. http://doi.org/10.1016/j.jhydrol.2021.127048
  » http://doi.org/10.1016/j.jhydrol.2021.127048
• Woldesenbet, T. A., Elagib, N. A., Ribbe, L., & Heinrich, J. (2017). Gap filling and homogenization of climatological datasets in the headwater region of the Upper Blue Nile Basin, Ethiopia. International Journal of Climatology, 37(4), 2122-2140. http://doi.org/10.1002/joc.4839
  » http://doi.org/10.1002/joc.4839
• Xu, C., Wang, W., Hu, Y., & Liu, Y. (2024). Evaluation of ERA5, ERA5-Land, GLDAS-2.1, and GLEAM potential evapotranspiration data over mainland China. Journal of Hydrology. Regional Studies, 51, 101651. http://doi.org/10.1016/j.ejrh.2023.101651
  » http://doi.org/10.1016/j.ejrh.2023.101651
• Xu, X., Frey, S. K., & Ma, D. (2022). Hydrological performance of ERA5 and MERRA-2 precipitation products over the Great Lakes Basin. Journal of Hydrology. Regional Studies, 39, 100982. http://doi.org/10.1016/j.ejrh.2021.100982
  » http://doi.org/10.1016/j.ejrh.2021.100982
• Xue, C., Wu, H., & Jiang, X. (2019). Temporal and spatial change monitoring of drought grade based on ERA5 analysis data and BFAST method in the belt and road area during 1989-2017. Advances in Meteorology, 2019(1), 4053718. https://doi.org/10.1155/2019/4053718
  » https://doi.org/10.1155/2019/4053718
• Yuan, X., Zhang, X., & Tian, F. (2020). Research and application of an intelligent networking model for flood forecasting in the arid mountainous basins. Journal of Flood Risk Management, 13(3), e12638. http://doi.org/10.1111/jfr3.12638
  » http://doi.org/10.1111/jfr3.12638
• Zambrano-Bigiarini, M. (2020). HydroGOF: goodness-of-fit functions for comparison of simulated and observed hydrological time series. R package version 0.4-0 Retrieved in 2023, June 10, from https://cran.r-project.org/web/packages/hydroGOF/hydroGOF.pdf
  » https://cran.r-project.org/web/packages/hydroGOF/hydroGOF.pdf
• Zubieta, R., Laqui, W., & Lavado, W. (2018). Hydrological modeling using observed and satellite precipitation datasets for the Ilave River basin, 2011-2015 period, Puno, Peru. IMTA-TC, 9(5), 85-101. http://dx.doi.org/10.24850/j-tyca-2018-05-04
  » http://dx.doi.org/10.24850/j-tyca-2018-05-04

Edited by

• Editor in-Chief: Adilson Pinheiro
  Associated Editor: Fernando Mainardi Fan

Publication Dates

• Publication in this collection
  10 Feb 2025
• Date of issue
  2025

History

• Received
  11 July 2024
• Reviewed
  02 Nov 2024
• Accepted
  06 Dec 2024