Revista Brasileira de Geografia Física v.17, n.06 (2024) 3974-3991.

Performance Assessment of Precipitation Estimates for Eastern Amazonia

Edmir dos Santos Jesus¹, Douglas Batista da Silva Ferreira², Antonio Vasconcelos Nogueira Neto³, Nilzele de Vilhena
Gomes Jesus4, Rômulo Augusto Jucá Oliveira5, Pedro Alberto Moura Rolim6, Cláudia Priscila Wanzeler da Costa7,
Renata Gonçalves Tedeschi8, Vânia dos Santos Franco9, Ana Paula Paes dos Santos10, Naurinete de Jesus da Costa
Barreto11
1
Dsc., Weatherman, Researcher. Vale Technological Institute (ITV), Street Boaventura da Silva, 955, Nazaré, Zip Code 66055-090, Belém/PA, (91)
980122541, Corresponding author: [email protected] (https://orcid.org/0000-0002-4383-5353); ²Dsc., Weatherman, Researcher,
[email protected] (https://orcid.org/0000-0002-8868-7235); ³Dsc., Oceanographer, Researcher, [email protected]
(https://orcid.org/0000-0002-3055-5950); 7Dsc., Weatherwoman, Researcher, [email protected] (https://orcid.org/0000-0002-3001-7378); 8Dsc.,
Physical, Researcher, [email protected] (https://orcid.org/0000-0002-9312-0030); 9Dsc., Weatherwoman, Researcher, [email protected]
(https://orcid.org/0000-0002-1970-7634); 10Dsc., Weatherwoman, Researcher, [email protected] (https://orcid.org/0000-0002-7736-7818); 11Dsc.,
Weatherwoman, Researcher, [email protected] (https://orcid.org/0000-0001-5167-6228). 4Dsc., Weatherwoman, Analyst, Management and
Operational Center of the Amazon Protection System (CENSIPAM), Av. Júlio Cesar, Val de Cans, SN, Belém/PA, [email protected]
(https://orcid.org/0000-0002-9744-5799); 6Msc., Analyst, CENSIPAM, [email protected] (https://orcid.org/0000-0002-0863-8557); 5Dsc. in
Science, Institute for Research and Development in Geosciences (GET), Toulouse - França, [email protected] (https://orcid.org/0000-
0001-5090-1817)
Artigo recebido em 31/12/2023 e aceito em 02/06/2024
ABSTRACT
The objective of this study was to evaluate the performance of precipitation estimators in relation to data observed by rain
gauges in the eastern Amazon, covering the states of Pará, Tocantins, and Maranhão, from January 2017 to December
2021. Precipitation data from the National Water Agency, the National Institute of Meteorology, and the Vale
Technological Institute were used, comprising a total of 14 measurement stations. The precipitation estimates evaluated
came from the CHIRPS, CMORPH-HR, CMORPH-BLD, ERA5, and GSMAP products, at a spatial resolution of 0.25°
latitude and longitude. Daily, monthly, and annual averages were calculated for subjective comparison; and the Nash-
Sutcliffe coefficients, accuracy coefficient, Pearson correlation, performance index, and BIAS were used for objective
analysis. The results showed that most estimators underestimated precipitation, except for CHIRPS. It was observed that
the monthly average precipitation presented better estimates in the dry season. CMORPH-BLD was among the products
that presented the best performance based on the indices. The differences found between the precipitation estimates show
the importance of evaluations of this type, as they can allow the most appropriate estimates to be chosen more carefully
in case of absence or failure in the data set. Topography and proximity to the coast were factors that influenced the
accuracy of precipitation estimates in eastern Amazonia.
Keywords: CMORPH, Topography, Efficiency, Amazon, Accuracy

Avaliação do Desempenho das Estimativas de Precipitação para o Leste da

Amazônia
RESUMO
O objetivo deste estudo foi avaliar o desempenho de estimadores de precipitação em relação aos dados observados por pluviômetros
no leste da Amazônia, abrangendo os estados do Pará, Tocantins e Maranhão no período de janeiro de 2017 a dezembro de 2021. Foram
utilizados dados de precipitação provenientes da Agência Nacional de Águas, do Instituto Nacional de Meteorologia e do Instituto
Tecnológico Vale, compondo ao todo 14 postos de medida. As estimativas de precipitação avaliadas foram provenientes dos produtos
CHIRPS, CMORPH-HR, CMORPH-BLD, ERA5 e GSMAP, em uma resolução espacial de 0,25° de latitude e longitude. Foram
calculadas as médias diária, mensal e anual para comparação subjetiva; e os coeficientes de Nash-Sutcliffe, coeficiente de exatidão,
correlação de Pearson, índice de desempenho e BIAS, para uma análise objetiva. Os resultados mostraram que a maioria dos
estimadores subestimaram a precipitação, exceto o CHIRPS. Foi observado que a média mensal da precipitação apresentou melhores
estimativas no período seco. Sendo o CMORPH-BLD entre os produtos que apresentou o melhor desempenho baseado nos índices. As
diferenças encontradas entre as estimativas de precipitação mostram a importância de avaliações deste tipo, de forma que podem
permitir a escolha das estimativas mais adequadas e de forma mais criteriosa em caso de ausência ou falha no conjunto de dados. A
topografia e a proximidade do litoral foram fatores que influenciaram na acurácia das estimativas de precipitação no leste da Amazônia.
Palavras-chave: CMORPH, Topografia, Eficiência, Amazônia, Acurácia

3974
Jesus E. S., Ferreira, D. B. S., Neto, A. V. N., Jesus, N. V. G., Oliveira, R. A. J., Rolim, P. A. M., Costa, C. P. W., Tedeschi, R. G.,
Franco, V. S., Santos, A. P. P., Barreto, N. J. C.
 Revista Brasileira de Geografia Física v.17, n.06 (2024) 3974-3991.
Introduction which increases the Transportable Moisture Limit
Despite the large number of (TML) and makes it impossible to load ore onto
meteorological stations in Brazil, the total number ships. Other processes, such as scheduling
of public institutions: the National Institute of maintenance of plants, railways and recovery of
Meteorology (INMET), the National Center for yards, for safety reasons, must be carried out in
Monitoring and Alerts for Natural Disasters good weather, without rain, visibility of over 10 km
(CEMADEN), the National Water Agency (ANA), and low risk of lightning.
the Center for Weather Forecasting and Climate For the reasons explained above,
Studies (CPTEC), among others, is still knowledge of the spatial-temporal distribution of
insufficient, given the size of Brazil's territory. In precipitation depends on a dense network of
the Amazon, the lack of observational records is observational records and large data samples with
even more striking, as it is a region with greater high temporal frequency (Huffman et al. 2020).
logistical complexity for the installation and However, this reality is not observed in the
maintenance of sensors, in addition to the Amazon, which can be explained by the
difficulties of connecting to the internet for online inaccessibility of remote areas, the maintenance
transmission of data. and conservation of sensors, or the high costs of
Specifically, the precipitation variable, due installing and maintaining meteorological radars,
to the high spatial-temporal variability in tropical which are capable of identifying the nearby
regions, is one of the most relevant among atmosphere in detail.
observational records, both for regulating the An alternative way to obtain rainfall
regional climate (De Souza et al., 2009; Alvares et information is through remote sensing. Satellite-
al., 2013), and for various types of industries, such based sensors have helped to understand rainfall
as mining, for example. Because it occurs in open patterns in areas with little or no sampling
areas, mining is episodically exposed to (Sanchez-Moreno et al., 2014). In recent years,
atmospheric phenomena that induce precipitation, many studies have used precipitation estimates
which influence operational activities related to through remote sensing as a data source, comparing
mineral production and the safety of people them with conventional measurements in the most
(Ferreira et al. 2024). diverse parts of the planet, such as in China (Sun et
In Brazil, important mining activity takes al., 2018; An et al., 2020), Iran (Alijanian et al.,
place in the eastern Amazon, between the states of 2017 and Ziveh et al., 2021), the United States
Pará and Maranhão, where seasonal rainfall is one (Zeweldi and Gebremichael, 2009; Behrangi et al.,
of the main characteristics that define the regional 2011), Africa (Habib et al., 2012; Satgé et al.,
climate. During the rainy season, which occurs 2020), Australia (Chua, Kuleshov and Watkins,
between the months of December and May, the 2020) and Brazil (Sodré and Rodrigues, 2013;
Intertropical Convergence Zone - ITCZ (Ferreira et Cordeiro and Blanco, 2021; Santos, 2021; Franco
al., 2015), the South Atlantic Convergence Zone - et al., 2022, Dos Santos Silva et al., 2023, Da Silva,
SACZ (Carvalho et al., 2004), the Upper Level 2024).
Cyclonic Vortex Edge - VCAN (Reboita et al., Helmi and Abdelhamed (2023) evaluated
2017) and Instability Lines - LI (Cohen et al., 1995) daily, monthly, annual, and annual maximum
are formed. On the other hand, during the dry precipitation records in arid regions of Saudi
season, rainy events are induced by small-scale Arabia. The daily analysis revealed a low
atmospheric systems, which generate correlation for all estimators (below 0.31),
thermodynamic factors and local convection, improving slightly in the annual and annual
which cause rain in the form of showers, through maximum analysis and reaching its highest value
the manifestation of IL and convective clusters (0.58) in the monthly analysis. Quantitative
(Fisch et al., 1998; Lopes et al., 2013). In the statistical measures included Pearson's correlation
southeastern portion of Pará, where the ore mining coefficient (r), Nash-Sutcliffe (NS) efficiency
process takes place, the action of frontal systems coefficient. The best performance for all statistical
coming from the Southern Region of Brazil measures was found at the altitude of 500 to 750 m
intensifies the regional rainfall volume, as well as in the central and northern parts of the study area
the electrical activity in the atmosphere (Amorim for all satellites except minor anomalies. The
Neto et al., 2015). CMORPH dataset presented the lowest centered
The mining process, storage in yards, root mean square error (RMSEc) for all analysis
transportation by rail and loading of ore from ships periods with the best results in the monthly
are negatively impacted by moisture from rain, analyses.
3975
Jesus E. S., Ferreira, D. B. S., Neto, A. V. N., Jesus, N. V. G., Oliveira, R. A. J., Rolim, P. A. M., Costa, C. P. W., Tedeschi, R. G.,
Franco, V. S., Santos, A. P. P., Barreto, N. J. C.
 Revista Brasileira de Geografia Física v.17, n.06 (2024) 3974-3991.
Zuh and Liu (2024) demonstrated that the specific manner in the eastern Amazon, more
performance of satellite-based estimation products specifically in a strategic region for mining
varied significantly when evaluated by different activity. Due to the different algorithms used for
metrics and across spatial and temporal scales. the different satellite products, one of the databases
However, the morphing product technique is expected to present results that allow its use for
(CMORPH-BLD) showed better performance in practical operational planning purposes. The
summer (rainy season) than in winter (dry season) findings of this study may represent an alternative
over mainland China. reference for rainfall monitoring in locations with
Paca et al., (2023) analyzed the no in situ records, in the areas of a mining
performance of remote sensing products in company's mine, railway and port.
estimating precipitation and evapotranspiration in
the urban area of Belém-PA, with high and stable Materials and methods
precipitation, and found that, in general, CHIRPS
and CMORPH were in agreement with station data Study Area
and previous studies. Resulting in overestimation Figure 1 shows the study area located in the
for higher precipitation values and slightly eastern Amazon, covering the states of Pará,
underestimation for lower precipitation values in a Tocantins and Maranhão, where one of the main
monthly time interval. CHIRPS performed to some mining-logistics corridors in the world is located,
extent better than CMORPH in all stations. notably between the Carajás region, in southeastern
Da Silva Santos et al., (2024) showed that Pará, and the Ponta da Madeira Maritime Terminal
CHIRPS estimates provided good responses (TMPM), in São Luís, in the state of Maranhão,
compared to observed data with most values above connected by the Carajás Railway (EFC) (VALE,
0.9 as the coefficient of dispersion; Nash-Sutcliffe 2023). The study area has a very rugged
above 0.86. CHIRPS was more accurate in topography, ranging from mean sea level on the
detecting estimated rainfall and events, especially Maranhão coast to altitudes close to 700 meters,
those occurring before and after the rainy season such as in the Serra dos Carajás in Pará (Souza-
and in the rainy season itself, according to the Filho et al., 2016).
values of the correct proportion close to 1 and with To evaluate the performance of
the frequency polarization index close to 1. precipitation estimated by satellite and with
CHIRPS showed adequate capacity to represent the reanalysis data, daily precipitation data measured
spatiotemporal variation of precipitation. by 14 rain gauges throughout the study area (Figure
In view of the above, most previous studies 1) from the National Water Agency (ANA), the
carried out a spatial approach to validate National Institute of Meteorology (INMET) and
precipitation estimates by remote sensing. the automatic meteorological station of the Vale
However, it is important to know the punctual Technological Institute (ITV) were used as a
response of such databases, since, in practice, this reference, in the period from January 1, 2017 to
information is used for more local purposes. December 31, 2021, as this is a time interval in
In this context, this study aimed to evaluate which the data presented better quality in terms of
the performance of five precipitation estimators in outliers and a lower percentage of failures in the
relation to data observed by rain gauges in a time series.

Figure 1. Geographical location of the listed rainfall stations used in this study, distributed near the Carajás
Railway to Porto Ponta da Madeira in São Luís-MA.

3976
Jesus E. S., Ferreira, D. B. S., Neto, A. V. N., Jesus, N. V. G., Oliveira, R. A. J., Rolim, P. A. M., Costa, C. P. W., Tedeschi, R. G.,
Franco, V. S., Santos, A. P. P., Barreto, N. J. C.
 Revista Brasileira de Geografia Física v.17, n.06 (2024) 3974-3991.

In order to compare observed data with spatial and temporal resolution, based on
estimated data, five precipitation estimation precipitation estimates in the microwave and
products were evaluated, described below: infrared bands. It uses the Lagrangian interpolation
method to combine precipitation estimates from the
1) Climate Hazards Group InfraRed passive microwave (PMW) channel with infrared
Precipitation with Stations (CHIRPS): is an (IR) channel estimates from geostationary
estimate composed of different sources of satellites, where PMW measurements are
information, such as, The Climate Hazards propagated by IR-derived motion vectors (Joyce et
Group's Precipitation Climatology (CHPClim); al., 2010; Li, Yang, and Hong, 2013). This strategy
satellite observations with thermal infrared combines the advantages of high precision of
spectroscopy, quasi-global geostationary from microwave data with the high temporal resolution
the National Oceanic and Atmospheric of IR data. CMORPH has several versions with
Administration (NOAA), Climate Prediction implementations and adjustments. In this study,
Center (CPC), National Climatic Data Center two versions were used:
(NCDC); NOAA's Coupled Forecast System, 4) CMORPH-HR: is a version with corrections
version 2 (CFSv2); and observational data from based on rainfall information called BIAS-
weather stations (Funk et al., 2015). These data CORRECTED CMORPH, whose primary
are available at: version is combined with data from the CPC on
https://data.chc.ucsb.edu/products/CHIRPS-2.0 the continent and the GPCP on the ocean.
Available at
2) ERA 5: is a product of the fifth generation of https://www.ncei.noaa.gov/data/cmorph-high-
global climate atmospheric reanalysis data resolution-global-precipitation-
implemented by the European Centre for Medium estimates/access/daily/0.25deg/
Range Weather Forecasts (ECMWF) and is based
on the Integrated Forecast System (IPS), where 5) CMORPH-BLD (Blended Analysis): is a
precipitation is defined as the accumulated liquid mixed product of satellite and rain gauge data,
and frozen water, composed of rain and snow, whose correction is made with the daily data of
that falls on the Earth's surface. It is the sum of the CPC through the probability and density
large-scale precipitation and convective function interpolation method. Available at
precipitation. Large-scale precipitation is https://ftp.cpc.ncep.noaa.gov/precip/CMORPH_
generated by the cloud scheme in the IPS. RT/BLD.
Convective precipitation is generated by the
convection scheme in the IPS, which represents Although some of these precipitation
convection at spatial scales smaller than the grid products have versions with high spatial resolution,
(Dee et al., 2011; Hersbach et al., 2020). The data it was decided to use the 0.25° x 0.25° latitude and
are available at longitude resolution in this study, the lowest
(https://cds.climate.copernicus.eu/cdsapp#!/datas resolution in all estimators, so that the
et/reanalysis-era5-single-levels?tab=overview). discrepancies in the results obtained were not
related to the different sizes of the estimate grid,
3) Global Satellite Mapping of Precipitation similar to the study by Sun et al., (2018) which
(GSMAP): It is an estimate using data from evaluated the performance of precipitation
passive multiband microwave and infrared estimated by several multi-satellite precipitation
radiometers on the Global Precipitation products for the Fujiang River Basin, in southwest
Measurement Core Observatory satellite and a China.
constellation of other satellites, provided by the To equalize the time series, the quality of
Japan Aerospace Exploration Agency, available the measured data was checked to identify days
at http://sharaku.eorc.jaxa.jp/GSMaP_crest. with failures in conventional measurements. This
Further details on the GSMAP estimate are resulted in an average reduction of 5% of the total
described in Okamoto et al., (2005). of 1826 days with data throughout the series,
leaving 1733 days with valid measurements. It is
We used the Climate Prediction Center worth mentioning that other studies that evaluated
Morphing Technique (CMORPH) estimates, which the performance of precipitation estimates, in
are global precipitation analyses with very high relation to conventional measurement, used data
3977
Jesus E. S., Ferreira, D. B. S., Neto, A. V. N., Jesus, N. V. G., Oliveira, R. A. J., Rolim, P. A. M., Costa, C. P. W., Tedeschi, R. G.,
Franco, V. S., Santos, A. P. P., Barreto, N. J. C.
 Revista Brasileira de Geografia Física v.17, n.06 (2024) 3974-3991.
samples similar to this one, such as Li et al. (2013) annual precipitation of each location was added
who analyzed five years (from 2008 to 2012) of and then divided by the number of locations, that
data for the Yangtze River, in China; Satgé et al. is, 3.
(2020), who used four years of data (2000 to 2003)
for West Africa; and Behrangi et al. (2011) who 2nd to assess the accuracy of precipitation
used a six-year sample (2003 to 2008) in their estimation products with reference to rain gauge
analyses for the Illinois Basin, in the United States. measurements in the 14 study locations, the
After adjusting the time series, 2 analyses following statistical metrics were used:
were performed:
1) Nash-Sutcliffe coefficient or index (NS)
1st Subjective comparison of the annual described by Zhang et al. (2007), which indicates
and monthly averages from 2017 to 2021: identify the adjustment of the estimated data and can vary
and compare the spatiotemporal distribution of the from -∞ to 1 (Equation 1). The NS is associated
annual and monthly averages of measured and with the efficiency of the estimate. And according
estimated precipitation over the 14 locations to Silva et al. (2012), the index values, calculated
distributed throughout the study region (Figure 1). through Equation 1, are classified according to
In this analysis, it was decided to organize these Table 1.
locations into 4 groups defined by proximity. Thus, 𝑖∑𝑛 (𝑂 −𝑃 )²
𝑖
𝑁𝑆 = 1 − [ ∑𝑖=1
𝑛 (𝑂
−𝑂̅)²
] (1)
the locations of Carajás, Parauapebas and Serra 𝑖=1 𝑖
Pelada make up Group 1 (G1); Fazenda Alegria,
Marabá and Esperantina make up G2; Açailândia, On what: 𝑃𝑖 is the estimated precipitation (mm);
Reta Km 32, Vale do Pindaré and Tucumã as G3; 𝑂𝑖 is the observed precipitation (mm); 𝑂̅ is the
Alto Alegre, Pindaré Mirim and Miranda as G4 and average observed precipitation (mm) e 𝑛 is the
finally, São Luís separately because it is close to number of observations.
the Brazilian coast.
As an example of how the average annual
precipitation of G1 was obtained, the average

Table 1. Classification of NS index values, according to Silva et al., 2012.

NS Value Performance
> 0,75 Adequate
0,36 to 0,75 Satisfactory
< = 0,36 Inadequate

2) Coefficient or index of agreement or between the estimate and the observation.

accuracy (d): indicates the degree of agreement (Equation 3).
between the estimated and observed values. It ∑𝑛 ̅ ̅
𝑖=1(𝑃𝑖 −𝑃 ).(𝑂𝑖 −𝑂)
𝑟= (3)
varies from 0 to 1 and the closer to 1, the better the √∑𝑛 ̅ 𝑛 ̅
𝑖=1(𝑃𝑖 −𝑃 )².∑𝑖=1(𝑂𝑖 −𝑂)²
agreement (Willmott et al., 1985), and is given by
Equation 2:
∑𝑛
On what: 𝑃̅ is the average estimated precipitation
𝑖=1(𝑃𝑖 −𝑂𝑖 )²
𝑑 = 1 − [∑ 𝑛 (|𝑃 ̅ |+|𝑂𝑖 −𝑂̅|)²
] (2) (mm).
𝑖=1 𝑖 −𝑂

3) Pearson correlation coefficient or index (r): To identify the intensity of the correlation between
indicates the degree of correlation between data the estimates and conventional observation, the
sets. Positive (negative) values indicate a directly classification proposed by Dancey and Reidy
(inversely) proportional linear relationship (2006) will be used, according to Table 2.

3978
Jesus E. S., Ferreira, D. B. S., Neto, A. V. N., Jesus, N. V. G., Oliveira, R. A. J., Rolim, P. A. M., Costa, C. P. W., Tedeschi, R. G.,
Franco, V. S., Santos, A. P. P., Barreto, N. J. C.
 Revista Brasileira de Geografia Física v.17, n.06 (2024) 3974-3991.
Table 2. Correlation coefficient classification (r).
Value of the modulus of r Classification
0,7 to 1,0 Strong
0,4 to 0,6 Moderate
0,1 to 0,3 Weak

4) Confidence or performance index (c): through Equation 4, which are classified according
statistical indicators correlate the estimated values to Table 3.
with those measured, where we will have the 𝑐 = 𝑑. 𝑟 (4)
confidence or performance index, calculated

Table 3. Classification of the Willmott performance index (c) proposed by Camargo and Sentelhas (1997).
c value Performance
> 0,85 Optimum
0,76 to 0,85 Very good
0,66 to 0,75 Good
0,61 to 0,65 Reasonable
0,51 to 0,60 Acceptable
< 0,50 Bad

5) BIAS: which shows the average deviation of the adopted, so that it was evaluated whether the NS
estimate in relation to the observed data, obtained a value greater than 0.36, satisfying the
indicating whether there was an overestimation classification (from satisfactory to adequate). This
(positive values) or underestimation (negative index was chosen to begin the definition due to the
values) (Equation 5). The closer to zero, the better fact that it has the lowest value among the other
the result. criteria. Then, the performance index value (c) was
1
𝐵𝐼𝐴𝑆 = ∑𝑛𝑖=1(𝑃𝑖 − 𝑂𝑖 ) (5) verified to be greater than 0.5 (from acceptable to
𝑛
optimum); if so, the r index obtained a moderate to
Using an electronic spreadsheet, the
strong correlation (above 0.4). If there was more
observed and estimated daily data were subjected
than one estimator that met these conditions, the
to statistical analysis of the Nash-Sutcliffe
one with the highest d value (accuracy coefficient)
coefficients, Willmott performance, correlation,
was accepted, identifying the best precipitation
accuracy and BIAS, according to the methodology.
estimate suitable for use. If the estimator did not
To define the precipitation estimate that would
meet the conditions described above, it was not
obtain the best accuracy in relation to the observed
recommended for use. And so the analysis was
data, the sequence described in the flowchart in
done for each location, for each day of the month
Figure 2 was followed. Initially, criteria were
and year compared with the selected estimator data.

YES YES
Start r > 0.4
NS > 0.36? c > 0.5?

NO NO

There is a performance
THE ESTIMATOR
tie between the
IS NOT estimators?
RECOMMENDED

YES NO

Choose estimator with THE PROPER

highest value “d” ESTIMATOR

End

Figure 2. Flowchart for defining the precipitation estimate obtained due to better accuracy in relation to rain
gauge measurement.
3979
Jesus E. S., Ferreira, D. B. S., Neto, A. V. N., Jesus, N. V. G., Oliveira, R. A. J., Rolim, P. A. M., Costa, C. P. W., Tedeschi, R. G.,
Franco, V. S., Santos, A. P. P., Barreto, N. J. C.
 Revista Brasileira de Geografia Física v.17, n.06 (2024) 3974-3991.
Results e discussion However, in the location of São Luís (2,550 mm),
where the highest annual precipitation averages are
Quantitative analysis observed, the ERA5 product (2,435 mm) was the
closest to this value, followed by GSMAP (2,293
Figure 3 shows the average annual mm). The lowest annual precipitation average
precipitation from 2017 to 2021, measured by rain measured occurred in group (G3), both in relation
gauges and estimated by CHIRPS, CMORPH-HR, to the measured data (1,494 mm) and in the
CMORPH-BLD, GSMAP and ERA5, for groups estimates (1,330 mm).
G1 (around 1,700mm), G2 (around 1,900mm), G3 Franco et al., (2022) agree that the highest
(around 1,490mm), G4 (around 1,980mm) and in values are seen in the region located north of the
São Luís (just above 2,550mm) annually. Railway up to Porto Ponta da Madeira, which
In this first analysis, it can be seen that the corresponds to the rainiest quarter and where in this
other estimators show a follow-up with the area that differs in the southern part of the Railway,
observed precipitation, where a gradual increase in the rains begin to decrease.
the direction from southwest to northeast of the The annual analysis carried out for these
section (called Northern Corridor) between the locations divided into groups demonstrates the
study locations can be seen. Another observation comparison of observed and estimated
made was that the average annual precipitation precipitation, which agree with the climatology
estimated by CHIRPS stands out among the other described by Alvares et al., (2013) and Dos Santos
products, coming closest to the observed value. Silva et al., (2023).

Figure 3. Average annual precipitation measured by rain gauge and estimated by CHIRPS, CMORPH-HR,
CMORPH-BLD, ERA5 and GSMAP for the locations analyzed in this study subdivided into groups. Data
period: January 2017 to December 2021.

Figure 4 shows the average monthly from November to April. According to the
precipitation observed and estimated by observed data, the month with the highest average
precipitation products over the locations precipitation volume was February in G1, G2 and
distributed in G1, G2, G3, G4 and in São Luís. It G3, and March in G4 and São Luís. However, the
was observed that for both the recorded and data estimated by GSMAP in G1, G2, G3 and G4
estimated data, in the locations located near the and CMORPH-BLD in G3 were discrepant in this
coast (G4 and São Luís) the rainy season occurs sense, indicating limitations for the estimate of
from December to May (austral spring and maximum average precipitation in relation to the
summer), while in the interior points of the rain gauge data. During these months, the
continent (G1, G2 and G3) the average monthly occurrence of precipitation is mainly related to the
precipitation exceeded 100 millimeters per month action of large-scale meteorological systems such
3980
Jesus E. S., Ferreira, D. B. S., Neto, A. V. N., Jesus, N. V. G., Oliveira, R. A. J., Rolim, P. A. M., Costa, C. P. W., Tedeschi, R. G.,
Franco, V. S., Santos, A. P. P., Barreto, N. J. C.
 Revista Brasileira de Geografia Física v.17, n.06 (2024) 3974-3991.
as the ITCZ and VCAN (De Souza and Rocha, greater differences between the conventional
2006; Lopes et al., 2013). measurements and the estimated ones when
According to the average monthly compared to the same analysis carried out to
precipitation for the period analyzed, the dry identify the rainiest month. This discrepancy was
season in the study area begins in June and greater in G4, where the rain gauge indicated that
continues until November in G4 and São Luís, and August was the month with the lowest precipitation
from May to October in G1, G2 and G3, with value (18 mm), different from what was estimated
minimums below 20 mm in the rain gauge records. by CMORPH-HR as being the driest month in July
However, this average minimum value presents a and for CMORPH-BLD, September was the month
large variation when compared to the estimates. In with the lowest precipitation (Figura 4).
addition, it is noted that the driest month presented

Figure 4. Average monthly precipitation measured by rain gauge and estimated by CHIRPS, CMORPH-HR,
CMORPH-BLD, ERA5 and GSMAP products for the locations analyzed in this study. Data period: January
2017 to December 2021.

3981
Jesus E. S., Ferreira, D. B. S., Neto, A. V. N., Jesus, N. V. G., Oliveira, R. A. J., Rolim, P. A. M., Costa, C. P. W., Tedeschi, R. G.,
Franco, V. S., Santos, A. P. P., Barreto, N. J. C.
 Revista Brasileira de Geografia Física v.17, n.06 (2024) 3974-3991.
The estimation products generally show the Tucumã (10) as “Acceptable” and “Satisfactory” in
seasonality of precipitation over the Northern December and lastly in the locality of São Luís (14)
Corridor, especially monthly rainfall during the dry as “Very Good/Satisfactory” in January and
season. According to observational data, the “Acceptable/Satisfactory” in the two-month period
months with the highest accumulations (equal to or (Feb-Mar); as “Good/Satisfactory” in April and as
above 200 mm/month) are from January to April “Reasonable” and “Satisfactory” in May.
over groups G1 to G4 and from January to May The correlation between the data measured
over São Luís. While the period in which by the rain gauge and estimated by CMORPH-
precipitation does not exceed 50 mm/month BLD was above 0.7 and positive, that is, with a
extends from June to October over the locations of strong classification, according to the correlation
groups G1 to G4 and finally, from August to index (r) and the data from both sources were
October over São Luís. The greatest differences directly proportional.
observed during the dry season occurred due to the The values found by this estimator confirm
nature of most precipitation systems originating on its use in a satisfactory manner for the seasonality
a smaller scale (Lopes et al., 2013), which often and variability of rainfall in the region as described
cannot be recorded by a rain gauge, however they by Da Silva Santos et al. (2016) and by Helmi and
are captured by precipitation estimates. Abdelhamed (2023).
In the first analysis of the performance Figure 6 shows that precipitation was
indexes for the annual period, the results showed underestimated (negative BIAS) in most points
“Very Good to Optimum” and “Satisfactory to where the CMORPH-BLD estimator stood out,
Adequate” in most locations, differing from the except in December in the locality of Marabá (5),
quarterly period with results of “Bad to Very whose BIAS value was positive, indicating an
Good” and “Satisfactory to Adequate”. Therefore, overestimation of precipitation in this month.
it was decided to do the monthly period in order to The other estimators showed good accuracy
increase the margin in the amount of data observed in specific locations and months, as in the case of
for the qualitative and quantitative analysis. The the use of CMORPH-HR, which obtained an
other results, such as index (c) with “Bad” or Acceptable/Satisfactory result in Esperantina(6) in
“Terrible” performance and index (NS) January; Carajás(1) in the months of February and
“Inadequate”, were not selected to avoid excess May (Reasonable/Satisfactory) and in Tucumã(10)
information in the local spatial distribution. in the month of May as Acceptable/Satisfactory.
GSMAP stood out in the locality of Fazenda
Qualitative Analysis Alegria (4) in the two-month period (April-May) as
Acceptable/Good/Satisfactory. And in São Luís as
Figure 5 shows the distribution of the NS, c, “Excellent” and “Adequate” for the month of
d indices and the precipitation estimator that stood December with strong correlation and accuracy
out the most in relation to the series of daily data index (d) equal to 0.96.
analyzed in the rainy period (December to May). Figure 6 shows that these results confirm the
The CMORPH-BLD estimator presented the strong correlation with the data measured by rain
best performance from “Acceptable” to gauges in the study locations, as seen in the months
“Optimum” and from “Satisfactory” to “Adequate” of March and April as being those of greatest
in the months of the year, especially for the accumulated precipitation. They showed that of the
locations of Carajás (1) in the month of February; 14 locations, only 4 (in March) and two in April
in Parauapebas (2) in the months (Feb, Mar and responded with underestimated BIAS (-), with
May); in Fazenda Alegria (4) in the months of “Good”, “Very Good” and “Acceptable”
(Mar, Apr and Dec); in Marabá (5) in the month of performance for the Willmott index and
December as “Good” and “Satisfactory”, in the “Satisfactory” for the Nash-Sutcliffe index, and
months of February and May as “Very Good” and only the location of Marabá (5) responded with
“Adequate” and in the month of March as positive BIAS (overestimated) for the month of
“Optimum” and “Adequate” with strong December.
correlation and accuracy index (d) equal to 0.83; in

3982
Jesus E. S., Ferreira, D. B. S., Neto, A. V. N., Jesus, N. V. G., Oliveira, R. A. J., Rolim, P. A. M., Costa, C. P. W., Tedeschi, R. G.,
Franco, V. S., Santos, A. P. P., Barreto, N. J. C.
 Revista Brasileira de Geografia Física v.17, n.06 (2024) 3974-3991.

Figure 5. Monthly local distribution (December to May) of the Nash-Sutcliffe (NS) Index, Willmott's
concordance index (d) and performance (c) and of the precipitation estimators that stood out in relation to the
observed data among the study points.

3983
Jesus E. S., Ferreira, D. B. S., Neto, A. V. N., Jesus, N. V. G., Oliveira, R. A. J., Rolim, P. A. M., Costa, C. P. W., Tedeschi, R. G.,
Franco, V. S., Santos, A. P. P., Barreto, N. J. C.
 Revista Brasileira de Geografia Física v.17, n.06 (2024) 3974-3991.

Figure 6. Monthly local distribution (December to May) of the Pearson and BIAS correlation index (r) and of
the precipitation estimators that stood out in relation to the observed data between the study points.

3984
Jesus E. S., Ferreira, D. B. S., Neto, A. V. N., Jesus, N. V. G., Oliveira, R. A. J., Rolim, P. A. M., Costa, C. P. W., Tedeschi, R. G.,
Franco, V. S., Santos, A. P. P., Barreto, N. J. C.
 Revista Brasileira de Geografia Física v.17, n.06 (2024) 3974-3991.

Figures 7 and 8 show the distribution of satellite precipitation estimates over Australia,
indexes for the dry months (June to November), where the topography of the region may be a factor
where there was a variation in precipitation that contributes to the rainfall regime, as well as in
estimates that were efficient in relation to the study area.
conventional measurement. Figure 8 shows the correlation index (r)
The performance of CMORPH-BLD was category, the BIAS signal, and the estimator that
efficient, presenting NS as “Satisfactory” or obtained the best accuracy in relation to
“Adequate” and an agreement index (c) alternating conventionally measured data. Unlike the results
between “Good”, “Very Good” and “Acceptable” for the rainy months, in the dry months (from June
in Parauapebas (2) in the months of June, July and to November) there were more points in which
September; in Marabá (5), in the months of June precipitation was overestimated, such as in the
(“Good”), August (“Acceptable”), September months of July, August, and November in Serra
(“Very Good”) and November (“Good”); in Pelada (3); July and September in Fazenda Alegria
Fazenda Alegria (4) it stood out for obtaining the (4); September in Marabá (5) and Tucumã (10),
highest accuracy index value (d=0.95) for the and in August and November in the locality of São
month of September with index c (Very Good) and Luís. It is worth noting that all estimators that were
index NS (Adequate). In São Luís (14), the months efficient in their estimates presented strong
of June and July were “Acceptable” and correlation (r value above 0.70).
“Reasonable”, respectively. Furthermore, the CHIRPS estimator,
The CMORPH-HR estimator obtained which did not obtain good accuracy in the rainy
performance ranging from “Acceptable” to “Very season, responded with Acceptable/Satisfactory
Good” and from “Satisfactory” to “Adequate” for performance in the locations (Serra Pelada and
the quarter Sep-Oct-Nov in the locality of Carajás Tucumã) within the continent in the months of
(1), which corresponds to Group 1, which had an September and November (dry season), according
accuracy index (d) equal to 0.93 for the month of to Cordeiro and Blanco (2021), however, a
September. The ERA5 estimator obtained different response when compared to the coastal
performance indices ranging from “Acceptable” to region given by Santos et al., (2024).
“Very Good” in the most continental range (Serra On the coast, more specifically in the São
Pelada, Fazenda Alegria and Marabá), localities Luís region, the estimated precipitation, especially
belonging to G2 and G3 in the month of July. by CMORPH-BLD, showed good accuracy, while
The GSMAP estimator demonstrated the estimates in the locations of Reta 32 km (8),
better performance in its estimates, in relation to Vale do Pindaré (9) and Miranda (13) which are at
the five estimators analyzed, in the dry months, altitudes of 348, 161 and 46 meters, respectively,
when compared to the rainy months, standing out did not obtain good accuracy in any of the
in the precipitation estimate at Fazenda Alegria(4) estimators. These results reinforce that topography
in June (“Good; Satisfactory”); Açailândia influences the performance of precipitation
(“Reasonable; Satisfactory”) in August and estimates, corroborating the results obtained by
(“Acceptable; Satisfactory”) in September; Alto Ziveh et al., (2022) who found the greatest
Alegre(11) in August (“Acceptable; Satisfactory”); uncertainties in the 14 precipitation estimates
Parauapebas(2) and Marabá(5) in October (“Good evaluated in Iran over mountainous regions, due to
and Reasonable”) and in São Luís(14) in August the complex precipitation processes in these
(“Excellent; Adequate”) and November regions; also reinforcing the conclusions of
(“Excellent; Satisfactory”) results that are similar Alijanian et al., (2017) who found the best
to those found by Chua, Kuleshov and Watkins correlations between the estimates and
(2020) who evaluated the CMORPH and GSMaP precipitation measurements in the coastal areas of
datasets and pointed out as useful sources of the Persian Gulf.

3985
Jesus E. S., Ferreira, D. B. S., Neto, A. V. N., Jesus, N. V. G., Oliveira, R. A. J., Rolim, P. A. M., Costa, C. P. W., Tedeschi, R. G.,
Franco, V. S., Santos, A. P. P., Barreto, N. J. C.
 Revista Brasileira de Geografia Física v.17, n.06 (2024) 3974-3991.

Figure 7. Monthly distribution (June to November) of the Nash-Sutcliffe (NS) Index, Willmott's concordance
index (d) and performance (c) and of the precipitation estimators that stood out in relation to the observed data
among the study points.

3986
Jesus E. S., Ferreira, D. B. S., Neto, A. V. N., Jesus, N. V. G., Oliveira, R. A. J., Rolim, P. A. M., Costa, C. P. W., Tedeschi, R. G.,
Franco, V. S., Santos, A. P. P., Barreto, N. J. C.
 Revista Brasileira de Geografia Física v.17, n.06 (2024) 3974-3991.

Figure 8. Monthly local distribution (June to November) of the Pearson and BIAS correlation index (r) and of
the precipitation estimators that stood out in relation to the observed data between the study points.

3987
Jesus E. S., Ferreira, D. B. S., Neto, A. V. N., Jesus, N. V. G., Oliveira, R. A. J., Rolim, P. A. M., Costa, C. P. W., Tedeschi, R. G.,
Franco, V. S., Santos, A. P. P., Barreto, N. J. C.
 Revista Brasileira de Geografia Física v.17, n.06 (2024) 3974-3991.
Conclusions the scientific community and for monitoring
This study performed a qualitative purposes, in order to choose the most appropriate
evaluation of five precipitation estimation products estimates more carefully in case of absence or
for the mining chain in eastern Amazonia. Unlike failure in the observational data set.
previous studies, which performed spatial For the mining chain in the eastern
evaluations, this work's approach was specific, Amazon, it can be considered strategic to know that
considering a comparative analysis with rainfall CMORPH-BLD was the most suitable product for
stations from different institutions. monitoring rainfall for various purposes, especially
After applying combined statistical for those that require prior knowledge of recent
methods, some products were detected with better rainfall in areas where there are no local records.
performance when compared to in situ records. Among the processes that benefited are: control of
Among the main conclusions, the following can be slope stability along railways in order to mitigate
cited: landslides; monitoring of settling basins in order to
1. The comparison between the distribution of the avoid flooding in adjacent communities; control of
annual and monthly average of the analyzed period ore moisture for the safety of maritime navigation;
based on observational data and precipitation control of air quality for planning road wetting,
products revealed that the estimated data among several other activities inherent to the
demonstrate values slightly below the mineral production processes, which depend on
observational precipitation records, with the past rainfall information.
exception of the CHIRPS estimator; For future studies, it is suggested that the
2. It was observed that, in the monthly average, combination of different satellite products be
precipitation presented better estimates in the dry evaluated, aiming to obtain greater adherence to
period, when compared to the rainy period, observational records.
indicating that seasonality influenced the accuracy
of the estimates; Acknowledgments
3. Among the products that presented the best The authors would like to thank the
precipitation estimates, based on the analysis of the Foundation for Research Support and
performance indexes (Nash-Sutcliffe, Pearson, Development (FADESP) for granting Research
Willmott), was CMORPH-BLD, as it presented the Grants and the Vale Technological Institute (ITV).
highest relative frequency (56%) throughout the
period. Subdivided with “Excellent” (3%), “Very References
Good” (21%) and “Good” (24%) performance for Alijanian, Mohammadali et al. (2017) Evaluation
the locations of Marabá, Fazenda Alegria and of satellite rainfall climatology using
Parauapebas belonging to G1 and G2; CMORPH, PERSIANN‐CDR, PERSIANN,
4. The second option was the GSMAP product TRMM, MSWEP over Iran. International
with 21%, which stood out with “Excellent” Journal of Climatology, 37(14), 4896-4914.
performance for São Luís, especially in months https://doi.org/10.1002/joc.5131
with higher precipitation values, suggesting that Alvares, C. A.; Stape, J. L.; Sentelhas, P. C.;
they are the most recommended for use in the Gonçalves, J. L. M.; Sparovek, G. (2013)
analyzed region. In other locations such as Carajás Koppen’s climate classification map for Brazil.
and Esperantina, the estimators that stood out the Meteorologische Zeitschrift. 22(6). 711–728.
most were CMORPH-HR and CMORPH-BLD in https://doi.org/10.1127/0941-2948/2013/0507.
the months of January and February; Amorim Neto, A.; Satyamurty, P.; Correia, F. W.
5. In accordance with previous studies, the results (2015) Some observed characteristics of frontal
found in this work for different locations in eastern systems in the Amazon Basin. Meteorological
Amazonia presented evidence that regional terrain Applications, 22, 617-635.
characteristics, such as vegetation type, different https://doi.org/10.1002/met.1497
types of elevation, proximity to coastal regions, An, Y.; Zhao, W.; Li, C.; Liu, Y. (2020) Evaluation
among other factors, may be limiting requirements of Six Satellite and Reanalysis Precipitation
for diagnosing the accuracy of multi-satellite Products Using Gauge Observations over the
precipitation products and analysis over the study Yellow River Basin, China. Atmosphere 11,
region. 1223. https://doi.org/10.3390/atmos11111223
From a technical-scientific perspective, the Behrangi, A. et al. (2011) Hydrologic evaluation of
differences found between precipitation estimates satellite precipitation products over a mid-size
show the importance of assessments of this type for basin. Journal of Hydrology. 397(3-4), 225-
3988
Jesus E. S., Ferreira, D. B. S., Neto, A. V. N., Jesus, N. V. G., Oliveira, R. A. J., Rolim, P. A. M., Costa, C. P. W., Tedeschi, R. G.,
Franco, V. S., Santos, A. P. P., Barreto, N. J. C.
 Revista Brasileira de Geografia Física v.17, n.06 (2024) 3974-3991.
237. 24(2), 111-124. https://doi.org/10.1590/S0102-
https://doi.org/10.1016/j.jhydrol.2010.11.043 77862009000200001
Carvalho, L. M. V.; Jones, C.; Liebmann, B. (2004) De Souza, E. B.; & Rocha, E. J. P. (2006) Diurnal
The South Atlantic Convergence Zone: variation of rainfall in eastern Amazonia during
intensity, form, persistence, and relationships rainy season: mean characteristics and extreme
with intraseasonal to interannual activity and events. Revista Brasileira de Meteorologia,
extreme rainfall. Journal of Climate, v.17, p.88- 21(3), 142-152.
108. https://doi.org/10.1175/1520- Dee, D. P.; Uppala, S. M.; Simmons, A. J.;
0442(2004)017%3C0088:TSACZI%3E2.0.CO Berrisford, P.; Poli, P.; Kobayashi, S.; Andrae,
;2 U.; Balmaseda, M. A.; Balsamo, G.; Bauer, P.;
Cohen, J. C. P.; Dias, M. A. F. S.; Nobre, C. A. Bechtold, P.; Beljaars, A. C. M.; Van De Berg,
(1995) Environmental Conditions Associated L.; Bidlot, J.; Bormann, N.; Delsol, C.; Dragani,
with Amazonian Squall Lines: a case study. R.; Fuentes, M.; Geer, A. J.; Haimberger, L.;
Monthly Weather Review, v. 123, p. 3163-3174. Healy, S. B.; Hersbach, H.; Hólm, E. V.;
https://doi.org/10.1175/1520- Isaksen, L.; Kållberg, P.; Köhler, M.;
0493(1995)123%3C3163:ECAWAS%3E2.0.C Matricardi, M.; Mcnally, A. P.; Monge-Sanz, B.
O;2 M.; Morcrette, J. J.; Park, B. K.; Peubey, C.; De
Cordeiro, A. L. M.; Blanco, C. J. C. (2021) Rosnay, P.; Tavolato, C.; Thépaut, J.N.; Vitart,
Assessment of satellite products for filling F.: (2011). The ERA-Interim reanalysis:
rainfall data gaps in the Amazon region. Natural configuration and performance of the data
Resource Modeling. 34: e12298. assimilation system, Q. J. Roy. Meteor. Soc.,
https://doi.org/10.1111/nrm.12298 137, 553–597. https://doi.org/10.1002/qj.828
Costa, J.; Pereira, G.; Siqueira, M. E.; Cardozo, F.; Dos Santos Silva, F. D.; Da Costa, C. P. W.; Dos
Silva, V. V. (2019). Validação dos dados de Santos Franco, V.; Gomes, H. B.; Da Silva, M.
precipitação estimados pelo CHIRPS para o C. L.; Dos Santos Vanderlei, M. H. G.; Costa,
Brasil. Revista Brasileira de Climatologia, 24 R. L.; Da Rocha Júnior, R. L.; Cabral Júnior, J.
(15),228-243. B.; Dos Reis, J. S. et al. (2023) Intercomparison
http://dx.doi.org/10.5380/abclima.v24i0.60237 of Different Sources of Precipitation Data in the
Chua, Zhi-Weng; Kuleshov, Yuriy; Watkins, Brazilian Legal Amazon. Climate, 11, 241.
Andrew. (2020) Evaluation of Satellite https://doi.org/10.3390/cli11120241
Precipitation Estimates over Australia. Remote Ferreira D. B.; De Souza, E. B.; De Moraes, B. C.;
Sensing, 12(4), 678. Meira-Filho, L.G. (2015) Spatial and temporal
https://doi.org/10.3390/rs12040678 variability of rainfall in eastern Amazon during
Dancey, Christine P.; Reidy, John. (2006) the rainy season. Scientific World Journal.
Estatística Sem Matemática para Psicologia: 2015:209783.
Usando SPSS para Windows. Artmed, 608p. https://doi.org/10.1155/2015/209783.
ISBN 8536306882, 9788536306889 Ferreira, D. S., Paes, A. P., da Costa, C. W.,
Da Silva, S. F. (2024). Validation of the CHIRPS Tedeschi, R., Silva Júnior, R., Júnior, W. N.,
precipitation estimate in a Brazilian Cerrado Neto, A. N., Kuhn, P., Franco, V., Jesus, E.,
area. Revista Brasileira de Geografia Física, Souza-Filho, P. W., de Souza, E., Sodré, G.,
17(2),860-881. Silva, F., Saraiva, I., Carvalho, E., Rocha, R., &
https://doi.org/10.26848/rbgf.v17.2.p860-881 Alves, R. (2024). The Role of Meteorology in
Da Silva Santos, M. R., Vitorino, M. I., Braga, C. The Mining Chain in Northern Brazil.
C., de Oliveira Souto, J. I., Moura, M. N., & European Journal of Applied Sciences, 12(4),
Ferreira Junior, P. P. (2016). Padrões de 08-30. https://doi.org/10.14738/aivp.124.17165
precipitação mensal na mesorregião Fisch, G.; Marengo, J. A.; Nobre, C. A. (1998)
metropolitana de Belém: uso do CMORPH e Uma revisão geral sobre o clima da Amazônia.
componentes principais. Revista Brasileira De Acta Amazônica, 28(2), 101-126,
Geografia Física, 9(7), 2301–2311. https://doi.org/10.1590/1809-43921998282126
https://doi.org/10.5935/1984-2295.20160164 Franco, Vania dos Santos, Cláudia Priscila
De Souza, E. B.; Lopes, M. N. G.; Rocha, E. J. P. Wanzeler da Costa, Fabricio Felippe Dos
et al. (2009) Precipitação climatológica sobre a Santos e Helber Barros Gomes. (2022)
Amazônia oriental durante o período chuvoso: “Recomendação de bases de dados de
observações e simulações regionais com o precipitação para estudos climáticos no
RegCM3. Revista Brasileira de Meteorologia, Corredor Norte da Vale.”. Relatório Técnico ao
3989
Jesus E. S., Ferreira, D. B. S., Neto, A. V. N., Jesus, N. V. G., Oliveira, R. A. J., Rolim, P. A. M., Costa, C. P. W., Tedeschi, R. G.,
Franco, V. S., Santos, A. P. P., Barreto, N. J. C.
 Revista Brasileira de Geografia Física v.17, n.06 (2024) 3974-3991.
Instituto Tecnológico Vale (ITV). Okamoto, K.; Iguchi, T.; Takahashi, N.; Iwanami,
https://doi.org/10.29223/prod.tec.itv.ds.2022.4 K.; Ushio, T. (2005) The global satellite
4.franco mapping of precipitation (GSMaP) project.
Funk, C.; Peterson, P.; Landsfeld, M.; Pedreros, D.; Proc. of IGARSS, 3414-3416.
Verdin, J.; Shukla, S.; Husak, G.; Rowland, J.; https://doi.org/10.1109/IGARSS.2005.1526575
Harrison, L.; Hoell, A.; Michaelsen, J. (2015). Paca, V. H. D. M.; De Souza E. B.; Queiroz, J. C.
The climate hazards infrared precipitation with B.; Espinoza-Dávalos, G. E. (2023) Assessment
stations – A new environmental record for of Precipitation and Evapotranspiration in an
monitoring extremes, Sci. Data, 2, 1–21. Urban Area Using Remote Sensing Products
https://doi.org/10.1038/sdata.2015.66 (CHIRP, CMORPH, and SSEBop): The Case of
Helmi, A. M.; Abdelhamed, M. S. (2023) the Metropolitan Region of Belem, Amazon;
Evaluation of CMORPH, PERSIANN-CDR, Water, 15(19), 3498;
CHIRPS V2.0, TMPA 3B42 V7, and GPM https://doi.org/10.3390/w15193498
IMERG V6 Satellite Precipitation Datasets in Rahmati, A.; Bakhtar, A.; Shayeghi, A.; Kalantari,
Arabian Arid Regions. Water; 15(1):92. Z.; Bavani, A. M.; Ghajarnia, N. (2021) Spatio-
https://doi.org/10.3390/w15010092 temporal performance evaluation of 14 global
Hersbach, H.; Bell, B.; Berrisford, P. et al. (2020) precipitation estimation products across river
The ERA5 global reanalysis. QJR Meteorology basins in southwest Iran. ESS Open Archive.
Soc..146:1999-2049. September 03,.
https://doi.org/10.1002/qj.3803 . https://doi.org/10.1002/essoar.10507895.1
Huffman, G. J. et al. (2020). Integrated Multi- Reboita, M. S., Campos, B., Santos, T., Gan, M. A.,
satellite Retrievals for the Global Precipitation & Carvalho, V. (2017). Análise sinótica e
Measurement (GPM) Mission (IMERG). In: numérica de um VCAN no Nordeste do Brasil
Levizzani, V., Kidd, C., Kirschbaum, D.B., (Synoptic and numerical analysis of a VCAN in
Kummerow, C.D., Nakamura, K., Turk, F.J. the Brazilian Northeast). Revista Brasileira de
(eds) Satellite Precipitation Measurement. Geografia Física, 10(1), 041–059.
Advances in Global Change Research, v.67. https://doi.org/10.5935/1984-2295.20170004
Springer, Cham. Roca, R; Alexander, L. V.; Potter, G.; Bador, M.;
https://doi.org/10.1007/978-3-030-24568-9_19 Jucá, R. A.; Contractor, S.; Bosilovich, M. G.;
Joyce, R.J., Xie, P., Yarosh, Y., Janowiak, J.E., Cloché, S. (2019) FROGS: a daily 1°×1°
Arkin, P.A. (2010). CMORPH: A “Morphing” gridded precipitation database of rain gauge,
Approach for High Resolution Precipitation satellite and reanalysis products. Earth Syst. Sci.
Product Generation. In: Gebremichael, M., Data, 11, 1017-1035.
Hossain, F. (eds) Satellite Rainfall Applications https://doi.org/10.5194/essd-11-1017-2019
for Surface Hydrology. Springer, Dordrecht. Sanchez-Moreno, J. F.; Mannaerts, C.; Jetten, V.
https://doi.org/10.1007/978-90-481-2915-7_2 (2014). Influence of topography on rainfall
Joyce, R.J.; Janowiak, J.E.; Arkin, P.A.; Xie, P. variability in Santiago Island, Cape Verde.
(2004) CMORPH: A method that produces International Journal of Climatology, n. 34,
global precipitation estimates from passive 1081-1097. https://doi.org/10.1002/joc.3747
microwave and infrared data at high spatial and Santos, M. R. S., Paula, E. M. S. de, Rabelo, D. R.,
temporal resolution. Journal & Pimentel, M. A. da S. (2024). Variação
Hydrometeorology. 5, 487-503. Espaço-Temporal (1981-2020) da Chuva na
https://doi.org/10.1175/1525- Região Costeira do Estado do Pará – Amazônia
7541(2004)005%3C0487:CAMTPG%3E2.0.C Oriental. Caminhos de Geografia, 25(99), 167–
O;2 178. https://doi.org/10.14393/RCG259970567
Li, Zhe; Yang, Dawen; Hong, Yang. (2013) Multi- Satgé, Frédéric et al. (2020) Evaluation of 23
scale evaluation of high-resolution multi-sensor gridded precipitation datasets across West
blended global precipitation products over the Africa. Journal of Hydrology, 581, 124412.
Yangtze River. J. Hydrology. 500, 157-169. https://doi.org/10.1016/j.jhydrol.2019.124412
https://doi.org/10.1016/j.jhydrol.2013.07.023 Silva Júnior, R. O.; Queiroz, J. C. B.; Ferreira, D.
Lopes, M. N. G.; De Souza, E. B.; Ferreira, D. B. B. S.; Tavares, A. L.; Souza Filho, P. W. M.;
S. (2013) Climatologia regional da precipitação Guimarães, J. T. F.; Rocha, E. J. P. (2017).
no estado do Pará. Revista Brasileira de Estimativa de precipitação e vazões médias para
Climatologia, 12, 84-102. a bacia hidrográfica do Rio Itacaiúnas (BHRI),
https://doi.org/10.5380/abclima.v12i1.31402 Amazônia Oriental, Brasil. Revista Brasileira
3990
Jesus E. S., Ferreira, D. B. S., Neto, A. V. N., Jesus, N. V. G., Oliveira, R. A. J., Rolim, P. A. M., Costa, C. P. W., Tedeschi, R. G.,
Franco, V. S., Santos, A. P. P., Barreto, N. J. C.
 Revista Brasileira de Geografia Física v.17, n.06 (2024) 3974-3991.
de Geografia Física. 10(5). 1638-1654. https://www.vale.com/pt/logistica. Acesso em:
https://doi.org/10.26848/rbgf.v.10.5.p1638- 18 nov. 2023.
1654 Willmott, C. J.; Ackleson, S. G.; Davis, R. E.;
Silva, B. M.; Montenegro, S. M. G. L.; Silva, F. B. Feddema, J. J.; Klink, K. M.; Legates, D. R.;
Da; Araújo Filho, P. F. (2012) Chuvas intensas Rowe, C. M.; O’Donnell, J. (1985) Statistics for
em localidades do estado de Pernambuco. the evaluation and comparison of models.
Revista Brasileira de Recursos Hídricos. 17(3), Journal of Geophysical Research, Ottawa, 90
135-147. (5), 8995-9005.
Sodré, G. R. C.; & Rodrigues, L. L. M. (2013) https://doi.org/10.1029/JC090iC05p08995
Comparison Between Estimation of Xie, P., Joyce, R., Wu, S., Yoo, S.-H., Yarosh, Y.,
Precipitation Observed by the CMORPH Sun, F., And Lin, R. (2017) Reprocessed, Bias-
Technique and INMET Meteorological Stations Corrected CMORPH Global High Resolution
in Different Regions of Brazil. Brazilian Precipitation Estimates from 1998, J.
Journal of Physical Geography, 6(2), 301-307. Hydrometeorology, 18, 1617-1641,
https://doi.org/10.26848/rbgf.v6.2.301-307 https://doi.org/10.1175/JHM-D-16-0168.1
Souza-Filho, P. W. M.; De Souza, E. B.; Silva Xie, P.; Xiong, A.Y. (2011) A conceptual model
Júnior, R. O.; Nascimento, W. R.; De for constructing high-resolution gauge-satellite
Mendonça, B. R. V.; Guimarães, J. T. F.; merged precipitation analyses. J. Geophysical.
Dall’agnol, R.; Siqueira, J. O. (2016) Four Res. Atmos., 116.
decades of land-cover, land-use and https://doi.org/10.1029/2011JD016118
hydroclimatology changes in the Itacaiúnas Zeweldi, D. A.; Gebremichael, M. (2009)
River watershed, southeastern Amazon. Evaluation of CMORPH precipitation products
Journal of Environmental Management. 167, at fine space–time scales. Journal of
175-184. Hydrometeorology, 10(1), 300-307.
https://doi.org/10.1016/j.jenvman.2015.11.039 https://doi.org/10.1175/2008JHM1041.1
Sun, W.; Ma, J.; Yang, G.; Li, W. (2018) Statistical Zhang, H.G.; Fu, S.H.; Fang, W.H.; Imura, H. &
and Hydrological Evaluations of Multi-Satellite Zhang, X. C. (2007) Potential effects of climate
Precipitation Products over Fujiang River Basin change on runoff in the Yellow River Basin of
in Humid Southeast China. Remote Sens., 10, China. Journal of the ASABE, 50: 911-918.
1898. https://doi.org/10.3390/rs10121898 https://doi.org/10.13031/2013.23155
Ushio, T., Kubota, T., Shige, S., Okamoto, K., Ziveh, Akbar Rahmati et al. (2022) Spatio-
Aonashi, K., Inoue, T., Takahashi, N., Iguchi, temporal performance evaluation of 14 global
T., Kachi, M., Oki, R., Morimoto, T., And precipitation estimation products across river
Kawasaki, Z.: (2009) A Kalman Filter basins in southwest Iran. Journal of Hydrology:
Approach to the Global Satellite Mapping of Regional Studies, 44, 101269.
Precipitation (GSMaP) from Combined Passive https://doi.org/10.1016/j.ejrh.2022.101269
Microwave and Infrared Radiometric Data, J. Zhu, S.; Liu, Z. (2024) Comprehensive quantitative
Meteorol. Soc. Jpn., 87A, 137-151, assessment of the performance of fourteen
https://doi.org/10.2151/jmsj.87a.137 satellite precipitation products over Chinese
VALE. Saiba mais sobre os negócios da Vale: mainland. Clim Dyn.
minério de ferro e pelotas. https://doi.org/10.1007/s00382-024-07237-8

3991
Jesus E. S., Ferreira, D. B. S., Neto, A. V. N., Jesus, N. V. G., Oliveira, R. A. J., Rolim, P. A. M., Costa, C. P. W., Tedeschi, R. G.,
Franco, V. S., Santos, A. P. P., Barreto, N. J. C.