565 Atmósfera Atmósfera 0187-6236 2395-8812 Universidad Nacional Autónoma de México México claudio.amescua@atmosfera.unam.mx 56582679013 10.20937/ATM.52966 Sin sección Prediction of atmospheric corrosion from meteorological parameters:Case of the atmospheric basin of the Costa Rican Western Central Valley Rodríguez-Yáñez Javier E. jrodriguezy@uned.ac.cr Rivera-Fernández Erick Alvarado-González Daniel Abdalah-Hernández Mariela Quirós-Quirós Rafael Urban Ecology Research Laboratory, Costa Rica Distance University, 474-2050, Mercedes de Montes de Oca, San José, Costa Rica. Costa Rica Urban Ecology Research Laboratory, Costa Rica Distance University School of Physics and Center for Geophysical Research, University of Costa Rica, 11501-2060, San Pedro de Montes de Oca, San José, Costa Rica. Costa Rica School of Physics and Center for Geophysical Research, University of Costa Rica Advanced Computing Laboratory, National High Technology Center, 1174-1200, Pavas, San José, Costa Rica. Costa Rica School of Physics and Center for Geophysical Research, University of Costa Rica Advanced Computing Laboratory, National High Technology Center, 1174-1200, Pavas, San José, Costa Rica. Costa Rica Advanced Computing Laboratory, National High Technology Center, School of Geography, University of Costa Rica, 11501-2060, San Pedro de Montes de Oca, San José, Costa Rica. Costa Rica School of Geography, University of Costa Rica jrodriguezy@uned.ac.cr Enero 2023 36 1 171 182 26 08 2020 20 04 2021 Abstract

The assessment of atmospheric corrosion is currently based on the study of atmospheric basins (AB). Modeling atmospheric corrosion in many cases involves the measurement of several meteorological and atmospheric pollution parameters, specifically temperature, relative humidity, chloride, and sulfur dioxide, making the estimation process more complex. However, for the Western Central Valley (WCV) in Costa Rica, a low-pollution AB, it is possible to develop simplified atmospheric corrosion models based on a small number of atmospheric parameters. In this paper, the meteorological variables of the study region were analyzed in terms of their dependency on altitude and their applicability in the development of a simplified model to predict the corrosion rate (Vcorr). The output of the predictive model was compared with the standard model of ISO 9223:2012, showing a higher skill and giving reliable results for a wide interval of altitudes.

 Resumen

La evaluación de la corrosión atmosférica se basa actualmente en el estudio de las cuencas atmosféricas (AB, por sus siglas en inglés). La modelización de la corrosión atmosférica implica en muchos casos la medición de varios parámetros meteorológicos y de contaminación atmosférica, concretamente la temperatura, la humedad relativa, el cloruro y el dióxido de azufre, lo que hace más complejo el proceso de estimación. Sin embargo, para el Valle Central Occidental (WCV, por sus siglas en inglés) de Costa Rica, un AB de baja contaminación, es posible desarrollar modelos simplificados de corrosión atmosférica, basados en un pequeño número de parámetros atmosféricos. En este trabajo se analizaron las variables meteorológicas de la región de estudio en cuanto a su dependencia de la altitud y su aplicabilidad en el desarrollo de un modelo simplificado para predecir la tasa de corrosión (Vcorr). El resultado del modelo de predicción se comparó con el modelo estándar de la norma ISO 9223:2012, mostrando una mayor habilidad y dando resultados fiables para un amplio intervalo de altitudes.

 Keywords atmospheric basin corrosion meteorological parameters basic modeling redalyc-journal-id 565 <bold>1. Introduction</bold>

The Costa Rican Western Central Valley (WCV) is limited to the northwest by the Central Volcanic mountain range, to the east by the hills of Ochomogo, and to the southwest by the foothills of the Talamanca mountain range. The average altitude of the WCV is about 1000 m above sea level (m.a.s.l.), with values up to 2000 m.a.s.l. in some places (IMN, 2008).

The region’s average annual precipitation is 2300 mm, with two clearly marked seasons: a rainy season from May to November and a dry season from December to April. During the rainy season, a relative minimum of precipitation occurs in July-August, known as the veranillo or mid-summer drought (Magaña et al., 1999). Temperature and relative humidity show small changes, with average values of 20 ºC and 75 %, respectively, varying with altitude and the time of the year (Solano and Villalobos, 2000; IMN, 2008). The predominant winds are the trade winds that blow from the northeast to the southwest. The Central Volcanic mountain range allows the passage of trade winds over the mountains or through mountain passes (mainly the gaps of La Palma and Desengaño). The seasonal variation of wind intensity, associated with the Caribbean low-level jet, determines most of the climatic changes in the WCV (Muñoz et al., 2002; Amador, 2008). Another characteristic pattern of the wind in the WCV is related to mountain-valley circulations. These are generated by temperature differences between both geographic landforms during the day, causing the wind to move from the mountain to the valley in the morning and in the opposite direction in the afternoon (Zárate, 1978; Barrantes et al., 1985). Due to the climatic and orographic conditions of the WCV, this area can be considered an Atmospheric Basin (AB). Here the climatic variability, circulation, and dispersion of air pollutants are driven mainly by the prevailing direction of winds and orography (Zárate, 1980; Muñoz et al., 2002; Caetano and Iniestra, 2008; García-Reynoso et al., 2009; Herrera et al., 2012; SEMARNAT-INECC, 2016).

Metallic materials are thermodynamically unstable under atmospheric conditions, which causes their deterioration, giving rise to the atmospheric corrosion phenomenon. A usual way to evaluate this process is by obtaining the loss of material in a fundamental state with respect to the time of exposure to the environment. This way allows the corrosion rate estimation as a loss of mass per unit area (mg m–2 y–1) or as a loss of sheet thickness (µm y–1) during the exposure time. Low carbon steel is the most common material for visualizing corrosión, similar to ASTM A36/A36M-19 (ASTM International, 2019).

Previous research by Morcillo et al. (2011) and the guidelines of the ISO 9223:2012 standard state that temperature (T), relative humidity (RH), chlorides (Cl-), and sulfur oxides (SO2) are the primary agents that contribute to corrosion. Temperature and relative humidity provide enough conditions for the material oxidation process, while pollutants modify the oxides formed on steel surfaces, producing distortion in the crystalline network and structural weakness. Consequently, materials turn more labile and porous, becoming more susceptible to corrosive attacks (Feliu et al., 1993; Morcillo et al., 2013; Díaz et al., 2018).

Temperature and relative humidity variables are obtained from meteorological data, and the deposition of both chlorides and sulfur oxides is measured or estimated by air pollution evaluation methods. These parameters permit the characterization of the effects of pollution variation on the corrosion rate in the WCV (Morcillo et al., 1998; Roberge et al., 2002; Morcillo et al., 2012; Robles, 2013; Garita-Arce et al., 2014; Rodríguez-Yáñez et al., 2015; Alcántara et al., 2015; Morcillo, 2017).

Garita-Arce et al. (2014) proposed a series of models for obtaining corrosion rates depending on the available information. For Costa Rica, maps of applicability of the corrosion models are based on the work of Rodríguez-Yáñez et al. (2015). These maps present an increasing complexity, starting from the Brooks deterioration index to the ISO 9223 (ISO, 1992) standard and using linear equations to model corrosion rate. The updated ISO 9223:2012 (ISO, 2012a) standard includes a new model for estimating corrosion. It considers the annual corrosion as a function of climatic and pollution variables.

 <bold>2. Data and methods</bold>

Hourly meteorological data for several stations within the WCV were provided by the National Meteorological Institute and the Costa Rican Electricity Institute (IMN and ICE, respectively, by their Spanish acronyms). Figure 1 shows the location of the area of study. The analysis period was from 2010 to 2019 (see Table I). Only meteorological stations with less than 5 % of missing data were considered. From these data, the variables of interest were T and RH. The time of wetness (TOW), defined as the fraction of hours per year where T is greater than 0 ºC and RH is greater than or equal to 80 %, was also obtained.

Location of the WCV and the weather stations used for data collection. Fig. 1 Location of the WCV and the weather stations used for data collection. Fig. 1 Location of the WCV and the weather stations used for data collection.

Location of the meteorological stations available in the WCV. Table I Location of the meteorological stations available in the WCV.
Name North latitude West longitude Altitude (m.a.s.l) Start date End date
RECOPE, Ochomogo 09º 53’ 40.2” 83º 56’ 19.4” 1546 01/01/2010 31/01/2018
CIGEFI 09º 56’ 11.0” 84º 02’ 43.0” 1210 01/01/2010 31/01/2018
IMN, Aranjuez 09º 56’ 16.6” 84º 04’ 10.8” 1181 01/01/2010 31/01/2018
Juan Santamaría International airport 09º 59’ 26.5” 84º 12’ 52.9” 913 01/01/2010 15/02/2018
Fraijanes lake 10º 08’ 14.4” 84º 11’ 36.6” 1720 01/01/2010 31/01/2018
RECOPE, La Garita 10º 00’ 19.0” 84º 17’ 45.0” 740 12/01/2010 31/01/2018
West Pavas, near to airport 09º 57’ 26.3” 84º 08’ 51.6” 997 01/01/2010 05/03/2015
Santa Bárbara 10º 02’ 00.0” 84º 09’ 57.0” 1070 01/01/2010 31/01/2018
Belén 09º 58’ 30.0” 84º 11’ 08.0” 926 01/01/2010 02/11/2017
Burio hill, Aserrí 09º 50’ 25.3” 84º 06’ 45.6” 1811 14/11/2012 12/10/2017
Altos Tablazo, Higuito 09º 50’ 12.4” 84º 03’ 18.1” 1660 28/06/2012 31/05/2017
Chitaria hill, Santa Ana 09º 53’ 30.1” 84º 11’ 37.3” 1717 01/03/2011 11/10/2017
Poás volcano lake 10º 11’ 21.0” 84º 13’ 55.2” 2598 05/09/2011 12/02/2018
Cedral hill, Escazú 09º 51’ 41.1” 84º 08’ 45.2” 2255 29/08/2012 12/10/2017
School of Agricultural Sciences, Santa Lucía, Heredia 10º 01’ 22.8” 84º 06’ 42.4” 1257 19/11/2013 31/01/2018
West Tobías Bolaños airport, Pavas 09º 57’ 26.3” 84º 08’ 51.6” 981 05/03/2015 10/10/2017
San Luis 10º 00’ 54.0” 84º 01’ 36.0” 1341 01/09/2018 31/9/2019
Colima 09º 57’ 09.0” 84º 05’ 24.2” 1145 01/01/2010 12/31/2017

<bold> <italic>2.1 Atmospheric conditions and air pollution</italic> </bold>

Within the studied period, the number of both warm (El Niño) and cold (La Niña) phases of the El Niño-Southern Oscillation (ENSO) is similar (NOAA, 2019). ENSO episodes are linked to inter-annual variations in circulation and precipitation patterns over Costa Rica (Fernández et al. 2013).

On the other hand, the volcanic activity was low and spaced in time. It was mainly associated with eruptions of the Poás and Turrialba volcanoes in 2015. During most of these eruptive events, winds were not blowing toward the WCV (OVSICORI, 2015). Under these conditions, several authors have shown that pollution levels of chlorides and sulfur oxides in the WCV are about 5 mg m–2 d–1 and 10 mg m–2 d–1, respectively (Robles, 2013; Herrera-Murillo et al., 2014). These levels correspond to a low polluted atmosphere.

<bold> <italic>2.2 Corrosion models</italic> </bold>

According to the results of Garita-Arce et al. (2014) for the low carbon steel, an estimation curve of the corrosion rate at the WCV can be derived from meteorological parameters, specifically RH and TOW. Corrosion rate (Vcorr) for the WCV is a linear function of TOW or RH, as given by the following equations:

where,

Vcorr: corrosion rate, µm y–1,

TOW: fraction average annual hours of wetness, dimensionless,

RH: average annual relative humidity, %.

The above equations were determined for an altitude of 1108 m.a.s.l. in the La Sabana sector of San José, Costa Rica.

The ISO 9223:2012 standard contains guidelines to determine Vcorr. One of them is to utilize the corrosion categories listed in Table II. A second alternative is to obtain it by employing model equations. In this standard, the set of equations to calculate the corrosion of carbon steel is:

where,

f st: factor of Steel

Vcorr: corrosion rate in the first year, µm y–1,

T: annual average temperature, ºC,

RH: annual average relative humidity, %,

Pd: annual average SO2 deposition, mg m–2 d–1,

Sd: annual average Cl- deposition, mg m–2 d–1.

Corrosion rate for different corrosivity categories for low carbon steel. Table II Corrosion rate for different corrosivity categories for low carbon steel.
Corrosion category Corrosion rate (µm y–1) Corrosivity level
C1 ≤ 1.3 Very low
C2 1.3 to 25 Low
C3 25 to 50 Medium
C4 50 to 80 High
C5 80 to 200 Very High
CX 200 to 700 Extreme

<bold> <italic>2.3. Data processing</italic> </bold>

Annual averages of RH, T, and TOW were obtained for each location with the corresponding standard deviation. A relatively high correlation between these parameters and altitude (A), in m.a.s.l., was found. The determination correlation coefficient (R2) is a useful statistic for testing linearity. An acceptable value of R2 must be higher than 0.75 (Wilks, 2011).

General data analysis was performed with Python. The corrosion rate was determined and represented in two maps. One was generated with a simplified model, and the other with the new method formulated by the ISO 9223:2012 standard. Contour lines for the WCV were taken from the Atlas of Costa Rica with a spatial resolution of 10 m (Ortiz-Malavasi, 2015), and the maps were created with ArcGIS software.

The simplified and the ISO 9223:2012 models were compared between them and against experimental measurements. The latter were obtained according to the procedures in the ISO 9223:2012, ISO 9225:2012, and ASTM G1-03 (ISO, 2012a; ISO, 2012b; ASTM International, 2017) standards.

 <bold>3. Results and discussion</bold> <bold> <italic>3.1. Meteorological parameters</italic> </bold>

Most stations were located below 1500 m.a.s.l. and inside the valley. There were less data points in medium- altitude locations a (1500 m.a.s.l. to 2000 m.a.s.l.), and only three in the most elevated parts (above 2000 m.a.s.l.). The geographical distribution of the meteorological observations was not homogeneous. Data points were scarce in the northern section of the WCV, particularly around the foothills of the Barva volcano.

Figures 2, 3, and 4 represent the relations between temperature, relative humidity, and time of wetness with altitude and their standard deviation.

Variation of (a) annual average and (b) standard deviation of temperature with altitude. Fig. 2 Variation of (a) annual average and (b) standard deviation of temperature with altitude. Fig. 2 Variation of (a) annual average and (b) standard deviation of temperature with altitude.

Variation of (a) annual average and (b) standard deviation of RH with altitude. Fig. 3 Variation of (a) annual average and (b) standard deviation of RH with altitude. Fig. 3 Variation of (a) annual average and (b) standard deviation of RH with altitude.

Variation of (a) annual accumulated average and (b) standard deviation of time of wetness with altitude. Fig. 4 Variation of (a) annual accumulated average and (b) standard deviation of time of wetness with altitude. Fig. 4 Variation of (a) annual accumulated average and (b) standard deviation of time of wetness with altitude.

Based on the information presented in figure 2, the equation that describes the temperature variation between 700 m.a.s.l. and 2600 m.a.s.l. is:

where,

T: annual average temperature, ºC,

A: altitude, m.a.s.l.

The value of the vertical temperature gradient in the WCV from equation (4) is remarkably similar to the global mean value of –6.5 ºC km–1 in the lower atmosphere Sendiña-Nadal and Pérez-Muñuzuri, 2006). The standard deviation tends to increase with altitude at a rate of 2 % for every 1000 m. This behavior is possibly related to a more significant seasonal variability of temperatures in mountainous regions of the WCV (Solano and Villalobos, 2000). Usually, the standard deviation of temperature in the WCV is less than 0.5ºC.

In figure 3, it can be observed that relative humidity increases linearly with an altitude between 700 m.a.s.l. and 2600 m.a.s.l. This relation is expressed as:

(5)

where,

RH: annual average relative humidity, %,

A: altitude, m.a.s.l.

At higher elevations, the air usually reaches the saturation point, and clouds may form. In the WCV, relative humidity was near the 100 % value at 2600 m.a.s.l. (Murphy and Hurtado, 2013). Above this altitude, equation (5) is not applicable.

The standard deviation of RH is usually less than 5 %, with a slight tendency to decrease with altitude. In the lowest areas of the valley, the seasonal variability of RH is usually determined by the differences between dry and rainy seasons (Abdalah-Hernández et al., 2019).

Campos and Castro (1992) explained that RH could quickly vary from 20 % to 100 % at higher elevations of the eastern part of the WCV. This variation is due to dry upper troposphere air and the transport of humid air masses through the mountain pass from the Caribbean region. According to these authors, RH fluctuations decrease at lower elevations.

The relation between time of wetness and altitude from 700 m.a.s.l. to 2600 m.a.s.l., as shown in figure 4, is as follows

where,

TOW: average fraction of annual hours of wetness per year, dimensionless,

A: altitude, m.a.s.l.

The TOW increases while the deviation standard decreases with altitude (Rodríguez-Yáñez et al., 2015). This proportionality is due to more saturated air at higher elevations, where RH increases and gets close to 100 %.

The variance of the TOW in the lower valley is associated with changes between the dry and rainy seasons. There is also greater homogeneity in the TOW at higher elevations since the RH values tend to be above 80 % for extended periods.

<bold> <italic>3.2. Estimation of Vcorr using the linear and ISO models</italic> </bold>

Using the linear relationships from the previous section (equations 4 to 6) is possible to calculate the variation of Vcorr with altitude by only considering meteorological variables (equations 1 and 2). The derived equations for corrosion rate, based on RH and TOW, are:

In both cases, the estimated R2 values are greater than 0.75 and therefore acceptable.

A similar process is applied to the ISO 9223:2012 equation (equation 3). The factors Pd and Sd are estimated by Herrera et al. (2012) approximation. In this case:

The simplified equation of ISO 9223:2012 is governed by the effect of RH and increases exponentially with the altitude.

<bold> <italic>3.3. Model output comparison</italic> </bold>

Field corrosion data was collected from September 2018 to September 2019. Table III shows the location and altitude of the measurement sites, as well as the annual average of the atmospheric variables and pollutants at each point. Atmospheric corrosion, from observations and models (equations 1 to 3 and 7 to 9), is shown in Table IV.

Annual mean atmospheric parameters and levels of pollutants in three different sites. Table III Annual mean atmospheric parameters and levels of pollutants in three different sites.
Site North latitude West longitude A (m.a.s.l.) RH (%) TOW (fraction) T (ºC) Cl– (mg m–2 d–1) SO2 (mg m–2 d–1)
CIGEFI 09º 56’ 11” 84º 02’ 43” 1210 79.92 0.5701 20.29 3.82 6.06
San Luis 10º 00’ 54” 84º 01’ 36” 1341 90.19 0.8126 18.36 3.52 8.58
Santa Ana 09º 52’ 54” 84º 11’ 14” 1772 83.46 0.6643 17.41 4.00 7.02

Atmospheric corrosion in three different sites, using<xref ref-type="disp-formula" rid="e1"> equations 1</xref> to <xref ref-type="disp-formula" rid="e3">3</xref>, <xref ref-type="disp-formula" rid="e7">7</xref> to <xref ref-type="disp-formula" rid="e9">9</xref>, and actual values, in µm y<sup>–1</sup>. Values in <italic>italics</italic> correspond to overestimations of the corrosion rate compared to actual measurements (bold black quantities). Other modeled corrosion rates showing small differences (less than 5 %) with respect to observations are underlined. Table IV Atmospheric corrosion in three different sites, using equations 1 to 3, 7 to 9, and actual values, in µm y–1. Values in italics correspond to overestimations of the corrosion rate compared to actual measurements (bold black quantities). Other modeled corrosion rates showing small differences (less than 5 %) with respect to observations are underlined.
Site Vcorr Vcorr Vcorr Vcorr Vcorr Vcorr Vcorr
model RH model TOW ISO 9223 Observed model RH model TOW ISO 9223
(eq. 1) (eq. 2) (eq. 3) (eq. 7) (eq. 8) (eq. 9)
CIGEFI 19.2 20.8 20.8 17.4 16.7 18.6 23.8
San Luis 28.9 26.5 30.0 17.3 18.2 19.6 25.2
Santa Ana 22.5 22.9 24.9 16.6 23.4 23.2 31.0

Pollutant levels were similar to those utilized for the approximations with equations 7 to 9; however, RH and TOW did not follow the expected tendency with altitude. The reason for this behavior is probably the position of the stations in the WCV since they were located under the influence of different climatic conditions. Particularly, the station in San Luis is constantly influenced by humid air masses transported from the Caribbean region (Campos and Castro, 1992).

Figure 5 illustrates the predicted corrosion from equations 7 to 9 and their comparison with actual measurements.

Vcorr vs. Altitude, for a model using a) <xref ref-type="disp-formula" rid="e7">equation 7</xref> (RH, dotted line), b) equation 8 (TOW, dashed line), c) <xref ref-type="disp-formula" rid="e9">equation 9</xref> (ISO 9223, solid line). Actual measurements are represented by cross markers (+). Fig. 5 Vcorr vs. Altitude, for a model using a) equation 7 (RH, dotted line), b) equation 8 (TOW, dashed line), c) equation 9 (ISO 9223, solid line). Actual measurements are represented by cross markers (+). Fig. 5 Vcorr vs. Altitude, for a model using a) equation 7 (RH, dotted line), b) equation 8 (TOW, dashed line), c) equation 9 (ISO 9223, solid line). Actual measurements are represented by cross markers (+).

The Vcorr values from the simple linear models based on RH and TOW (equations 7 and 8, respectively) increased rapidly with altitude and corresponded to C2 and C3 corrosivity categories, according to the ISO 9223:2012 standard (Table II). The corrosion in the ISO 9223:2012 model shows a more significant increase with altitude. This increase is due to the exponential behavior of equation 9, governed by the RH variable. The difference between observed values and the ISO 9223:2012 equation is around 20 % at medium altitudes. In the mountain (Santa Ana), this contrast reaches the order of 100 %.

Model outputs in comparison with observations presented notable differences. When the atmospheric parameters are used (equations 1 and 2), the deviations are greater in places with high values of RH (RH > 80 %) or TOW (TOW > 0.6), for example, in San Luis. When considering the simple models with altitude dependency (equations 7 and 8), there is generally a good estimation of Vcorr in the region between 1000 m.a.s.l. and 1400 m.a.s.l. Overestimation greater than 10% occurs at altitudes higher than 1500 m.a.s.l. It is important to consider that most of the population of the WCV lives within the 900 m.a.s.l to 1400 m.a.s.l altitude range (INEC, 2011).

The model-based of ISO 9223:2012 (equations 3 and 9) presented higher values with respect to observations. This problem was already pointed out in some studies carried out in the tropical atmosphere (Correa-Bedoya et al., 2007; Corvo et al., 2008, Morcillo et al., 2012; Morcillo et al., 2013, Garita-Arce et al., 2014; Rios et al., 2017; Vera et al., 2017).

In figure 6, the shown maps complement each other. It is possible to apply both the RH- and the TOW-based models (Fig. 6a or 6b, respectively) in the central region of the WCV (from 900 m.a.s.l to 1400 m.a.s.l.) and the ISO 9223:2012 model (Fig. 6c) overestimated the corrosion in all altitudes.

Map of Vcorr using the a) RH, b) TOW, and c) ISO models. Fig. 6 Map of Vcorr using the a) RH, b) TOW, and c) ISO models. Fig. 6 Map of Vcorr using the a) RH, b) TOW, and c) ISO models.

<bold>4. Conclusions</bold>

Simple models based on RH, TOW, and the ISO standard were used to determine Vcorr in the WCV. The measurements indicated intermediate (C2) levels of Vcorr in the region, with values ranging from 16.6 to 17.4 µm y–1. Almost all the proposed models presented a level of Vcorr within the C2 category.

The linear equations dependent on atmospheric parameters only (equations 1 and 2) always overestimated the value of Vcorr with differences between 10 % and 70 %. Conversely, the altitude-dependent models (equations 7 and 8), presented a more realistic estimation, between 1000 m.a.s.l and 1400 m.a.s.l.

The equation of the ISO 9223:2012 standard (equation 3) and its simplified form based on the altitude (equation 9) showed an exponential behavior governed by the increase of RH with altitude.

The results demonstrated that linear equations with regular parameters, such as RH, are more practical for determining Vcorr at low altitudes in the WCV, specifically in the range of 1000 m.a.s.l to 1400 m.a.s.l., where most of the population of the WCV lives. Vcorr calculations out of this range showed larger discrepancies with respect to observations (more than 10 %).

The ISO model overestimated the Vcorr values in the WCV for all altitudes. In some cases, the differences were greater than 20 %.

The WCV is not a climatically uniform region, as it shows significant spatial and temporal variations. This variability suggests the need for more assessments considering these factors, given the dependency of corrosion with RH. Future analyses should encompass mountain profiles, climatic sub-regions, and climatic periods (for example, dry and rainy seasons).

 Acknowledgments

Special thanks to the National Meteorological Institute and the Costa Rican Electricity Institute for providing the meteorological data of the Western Central Valley. The actual corrosion data was obtained from measuring sites located at facilities of the National Power and Light Company and the Ministry of Security. This study was funded by the National Council of University Presidents (CONARE) of Costa Rica as part of a collaborative project among the State Distance University (UNED-VINVES-6-10-50), the University of Costa (UCR-VI-805-B8-650), the National High Technology Center (CeNAT-VI-269-2017), the National University (UNA- SIA: 0600-17), and the Costa Rica Institute of Technology (ITCR-VIE 1490-021).

Rodríguez-Yáñez acknowledges the support of the Innovation and Human Capital for Competitiveness Program (PINN) of the Ministry of Science, Technology, and Telecommunications (MICITT) of Costa Rica.

 <bold>References</bold> Abdalah-Hernández M, Rodríguez-Yáñez J, Alvarado-González D. 2020. Gaussian mixture analysis of basic meteorological parameters: Temperature and relative humidity. Tecnología en Marcha 33:5-12. https://doi.org/10.18845/tm.v33i5.5068 Abdalah-Hernández M Rodríguez-Yáñez J Alvarado-González D Gaussian mixture analysis of basic meteorological parameters: Temperature and relative humidity Tecnología en Marcha 2020 33 5 12 10.18845/tm.v33i5.5068 Alcántara J, Chico B, de la Fuente D, Morcillo M. 2015. Airborne chloride deposit and its effect on marine atmospheric corrosion of mild steel. Corrosion Science 97:74-88. https://doi.org/10.1016/j.corsci.2015.04.015 Alcántara J Chico B de la Fuente D Morcillo M Airborne chloride deposit and its effect on marine atmospheric corrosion of mild steel Corrosion Science 2015 97 74 88 10.1016/j.corsci.2015.04.015 Amador J. 2008. The Intra-Americas Sea low-level jet. Annals of the New York Academy of Sciences 1146 (1):153-188. https://doi.org/10.1196/annals.1446.012 Amador J The Intra-Americas Sea low-level jet Annals of the New York Academy of Sciences 2008 1146 (1) 153 188 10.1196/annals.1446.012 ASTM International. 2017. ASTM G1-03(2017)e1 Standard Practice for Preparing, Cleaning, and Evaluating Corrosion Test Specimens. American Society for Testing and Materials International. West Conshohocken, PA, USA. https://doi.org/10.1520/G0001-03R17E01 (accessed 2018 June 20) ASTM International ASTM G1-03(2017)e1 Standard Practice for Preparing, Cleaning, and Evaluating Corrosion Test Specimens 2017 West Conshohocken, PA, USA American Society for Testing and Materials International 2018/06/20 10.1520/G0001-03R17E01 ASTM International. 2019. ASTM A36 / A36M-19 Standard Specification for Carbon Structural Steel. American Society for Testing and Materials International. West Conshohocken, PA, USA. https:// doi.org/10.1520/A0036_A0036M-19 (accessed 2018 June 20) ASTM International ASTM A36 / A36M-19 Standard Specification for Carbon Structural Steel 2019 West Conshohocken, PA, USA American Society for Testing and Materials International 2018/06/20 10.1520/A0036_A0036M-19 Barrantes J, Liao A, Rosales A. 1985. Atlas climatológico de Costa Rica. San José, Costa Rica: Instituto Meteorológico Nacional. Available at https://www.imn.ac.cr/atlas-climatologico (accessed 2019 Jun 12) Barrantes J Liao A Rosales A Atlas climatológico de Costa Rica 1985 San José, Costa Rica Instituto Meteorológico Nacional 2019/06/12 https://www.imn.ac.cr/atlas-climatologico Caetano E, Iniestra R. 2008. Identificación de cuencas atmosféricas en México . Instituto de Ecología y Cambio Climático de México. Available at https://docplayer.es/58286477-Identificacion-de-cuencas-atmosfericas-en-mexico.html (accessed 2018 April 15) Caetano E Iniestra R Identificación de cuencas atmosféricas en México 2018 Instituto de Ecología y Cambio Climático de México 2018/04/15 https://docplayer.es/58286477-Identificacion-de-cuencas-atmosfericas-en-mexico.html Campos M, Castro V. 1992. El clima a sotavento de las montañas de Costa Rica. Tópicos Meteorológicos y Oceanográficos 2:1-21. Available at: https://hdl.handle.net/10669/551 (accessed 2021 March 3) Campos M Castro V El clima a sotavento de las montañas de Costa Rica Tópicos Meteorológicos y Oceanográficos 1992 2 1 21 2021/03/03 https://hdl.handle.net/10669/551 Corvo F, Pérez T, Martín Y, Reyes J, Dzib L, González-Sánchez J and Castañeda A. 2008. Time of wetness in tropical climate: Considerations on the estimation of TOW according to ISO 9223 standard. Corrosion Science 50:206-219. https://doi.org/10.1016/j.corsci.2007.06.012 Corvo F Pérez T Martín Y Reyes J Dzib L González-Sánchez J Castañeda A Time of wetness in tropical climate: Considerations on the estimation of TOW according to ISO 9223 standard Corrosion Science 2008 50 206 219 10.1016/j.corsci.2007.06.012 Correa-Bedoya E, Botero-Vega C, Restrepo AH, Delgado J, Castaño JG, Echevarria F. 2007. Corrosión del acero al carbono, acero galvanizado y aluminio en diferentes atmósferas colombianas, Scientia et Technica, 13(36): 7-12. Correa-Bedoya E Botero-Vega C Restrepo AH Delgado J Castaño JG Echevarria F Corrosión del acero al carbono, acero galvanizado y aluminio en diferentes atmósferas colombianas Scientia et Technica 2007 13 (36) 7 12 Díaz I, Cano H, Lopesino P, de la Fuente D, Chico B, Jiménez J, Medina SF, Morcillo M. 2018. Five-year atmospheric corrosion of Cu, Cr and Ni weathering steels in a wide range of environments. Corrosion Science 141:146-157. https://doi.org/10.1016/j.corsci.2018.06.039 Díaz I Cano H Lopesino P de la Fuente D Chico B Jiménez J Medina SF Morcillo M Five-year atmospheric corrosion of Cu, Cr and Ni weathering steels in a wide range of environments Corrosion Science 2018 141 146 157 10.1016/j.corsci.2018.06.039 Feliu S, Morcillo M, Feliu S. 1993. The prediction of atmospheric corrosion from meteorological and pollution parameters. I. Annual corrosion. Corrosion Science 34:403-4014. https://doi.org/10.1016/0010-938X(93)90112-T Feliu S Morcillo M Feliu S The prediction of atmospheric corrosion from meteorological and pollution parameters. I. Annual corrosion Corrosion Science 1993 34 403 414 10.1016/0010-938X(93)90112-T Fernández W, Campos M, Amador JA. 2003. Second progress report of the AIACC LA06 project: Assessment of impacts and ad¬aptation measures for the water resources sector due to extreme events under climate change conditions in Central America. AIACC. Available at https://sedac.ciesin.columbia.edu/aiacc/progress/LA06_July03.pdf (accessed 2019 June 11) Fernández W Campos M Amador JA Second progress report of the AIACC LA06 project: Assessment of impacts and ad¬aptation measures for the water resources sector due to extreme events under climate change conditions in Central America 2003 AIACC 2019/06/11 https://sedac.ciesin.columbia.edu/aiacc/progress/LA06_July03.pdf García-Reynoso JA, Ruiz-Suárez LG, García-Escalante JS, Reséndiz-Juárez NA. 2009. Comportamiento de los contaminantes en cuencas atmosféricas: Metodología y estudio de caso (INE/A1-004/2009), Mexico: INE. García-Reynoso JA Ruiz-Suárez LG García-Escalante JS Reséndiz-Juárez NA Comportamiento de los contaminantes en cuencas atmosféricas: Metodología y estudio de caso (INE/A1-004/2009) 2009 Mexico INE Garita-Arce L, Rodríguez-Yáñez J, Robles-Cordero J. 2014. Modelado de la velocidad de corrosión de acero de baja aleación en Costa Rica. Revista Ingeniería 24:79-90. https://doi.org/10.15517/RING.V24I2.14624 Garita-Arce L Rodríguez-Yáñez J Robles-Cordero J Modelado de la velocidad de corrosión de acero de baja aleación en Costa Rica Revista Ingeniería 2014 24 79 90 10.15517/RING.V24I2.14624 Herrera-Murillo J, Rodríguez S, Rojas-Marín J, Báez A. 2012. Relations between bulk precipitation, PM10 composition and meteorological conditions in the metropolitan area of Costa Rica. The Open Journal of Atmospheric Sciences 6:19-32. https://doi.org/10.2174/1874282301206010019 Herrera-Murillo J Rodríguez S Rojas-Marín J Báez A Relations between bulk precipitation, PM10 composition and meteorological conditions in the metropolitan area of Costa Rica The Open Journal of Atmospheric Sciences 2012 6 19 32 10.2174/1874282301206010019 Herrera J, Rojas JF, Beita VH, Chaves M. 2014. Composición química de muestras de depositación total colectadas en el área metropolitana de Costa Rica en 2012. Ciencias Ambientales 48:30-38. https://doi.org/10.15359/rca.48-2.3 Herrera J Rojas JF Beita VH Chaves M Composición química de muestras de depositación total colectadas en el área metropolitana de Costa Rica en 2012 Ciencias Ambientales 2014 48 30 38 10.15359/rca.48-2.3 IMN. 2008. Clima, variabilidad y cambio climático en Costa Rica. San José, Costa Rica: Comité Regional de Recursos Hidráulicos. Available at http://cglobal.imn.ac.cr/documentos/publicaciones/CambioClimatico/climaVariabilidadCambioClimaticoCR.pdf (accessed 2019 April 17) IMN Clima, variabilidad y cambio climático en Costa Rica 2008 San José, Costa Rica Comité Regional de Recursos Hidráulicos 2019/04/17 http://cglobal.imn.ac.cr/documentos/publicaciones/CambioClimatico/climaVariabilidadCambioClimaticoCR.pdf INEC. 2011. Anuario estadístico 2010. San José, Costa Rica: Instituto Nacional de Estadística y Censos, Ministerio de Hacienda. Available at https://www.inec.cr/sites/default/files/documentos/anuario_estadistico/publicaciones/reanuarioestad2010-02.pdf (accessed 2019 May 27) INEC Anuario estadístico 2010 2011 San José, Costa Rica Instituto Nacional de Estadística y Censos, Ministerio de Hacienda 2019/05/27 https://www.inec.cr/sites/default/files/documentos/anuario_estadistico/publicaciones/reanuarioestad2010-02.pdf ISO. 1992. Corrosion of metals and alloys - Corrosivity of atmospheres - Classification (ISO 9223:1992). ISO. Available at https://www.iso.org/standard/16855.html (accessed 2018 June 20) ISO Corrosion of metals and alloys - Corrosivity of atmospheres - Classification (ISO 9223:1992) 1992 ISO 2018/06/20 https://www.iso.org/standard/16855.html ISO. 2012a. Corrosion of metals and alloys - Corrosivity of atmospheres - Classification, determination and estimation (ISO 9223:2012). ISO. Available at https://www.iso.org/standard/53499.html (accessed 2018 June 20) ISO Corrosion of metals and alloys - Corrosivity of atmospheres - Classification, determination and estimation (ISO 9223:2012) 2012 ISO 2018/06/20 https://www.iso.org/standard/53499.html ISO. 2012b. Corrosion of metals and alloys – Corrosivity of atmospheres – Measurement of pollution (ISO 9225:2012). ISO. Available at https://www.iso.org/standard/53501.html (accessed 2018 June 20) ISO Corrosion of metals and alloys – Corrosivity of atmospheres – Measurement of pollution (ISO 9225:2012) 2012 ISO 2018/06/20 https://www.iso.org/standard/53501.html Magaña V, Amador JA, Medina S. 1999. The midsummerdrought over Mexico and Central America.Journal of Climate 12 (6):1577-1588. https://doi.org/10.1175/1520-0442(1999)012%3C1577:TMDOMA%3E2.0.CO;2 Magaña V Amador JA Medina S The midsummer drought over Mexico and Central America Journal of Climate 1999 12 (6) 1577 1588 10.1175/1520-0442(1999)012<1577:TMDOMA>2.0.CO;2 Morcillo M, Almeida E, Rosales B, Uruchurtu J, Marrocos M. 1998. Corrosión y protección de metales en las atmósferas de Iberoamérica. Parte I: Mapas iberoamericanos de corrosión atmosférica (MICAT). Madrid, Spain: CYTED. Morcillo M Almeida E Rosales B Uruchurtu J Marrocos M Corrosión y protección de metales en las atmósferas de Iberoamérica. Parte I: Mapas iberoamericanos de corrosión atmosférica (MICAT) 1998 Madrid, Spain CYTED Morcillo M, De la Fuente D, Díaz I, Cano H. 2011. Atmospheric corrosion of mild steel. A review. Revista Metalurgia 47: 426–444. http://doi.org/10.3989/revmetalm.1125 Morcillo M De la Fuente D Díaz I Cano H Atmospheric corrosion of mild steel. A review Revista Metalurgia 2011 47 426 444 10.3989/revmetalm.1125 Morcillo M, Chico B, de la Fuente D, and Simancas J. 2012. Looking back on contributions in the field of atmospheric corrosion offered by the MICAT Ibero-American testing network. International Journal of Corrosion 2012:1-24. https://doi.org/10.1155/2012/824365 Morcillo M Chico B de la Fuente D Simancas J Looking back on contributions in the field of atmospheric corrosion offered by the MICAT Ibero-American testing network International Journal of Corrosion 2012 2012 1 24 10.1155/2012/824365 Morcillo M, Chico B, Díaz I, Cano H, de la Fuente D. 2013. Atmospheric corrosion data of weathering steels. A review. Corrosion Science 77:6-24. https://doi.org/10.1016/j.corsci.2013.08.021 Morcillo M Chico B Díaz I Cano H de la Fuente D Atmospheric corrosion data of weathering steels. A review Corrosion Science 2013 77 6 24 10.1016/j.corsci.2013.08.021 Morcillo M. 2017. Fundamental and research frontier of atmospheric corrosion. Materials, Madrid, Spain. https://doi.org/10.3390/books978-3-03842-642-4 Morcillo M Fundamental and research frontier of atmospheric corrosion 2017 Madrid, Spain Materials 10.3390/books978-3-03842-642-4 Muñoz AC, Fernández W, Gutiérrez JA, Zárate E. 2002. Variación estacional del viento en Costa Rica y su relación con los regímenes de lluvia. Tópicos meteorológicos y oceanográficos 9: 1-13. Muñoz AC Fernández W Gutiérrez JA Zárate E Variación estacional del viento en Costa Rica y su relación con los regímenes de lluvia Tópicos meteorológicos y oceanográficos 2002 9 1 13 Murphy GM, Hurtado, RH. 2013. Agrometeorología. Buenos Aires, Argentina: Editorial Facultad de Agronomía, Universidad de Buenos Aires. 57-70. Murphy GM Hurtado RH Agrometeorología 2013 57 70 Buenos Aires, Argentina Editorial Facultad de Agronomía, Universidad de Buenos Aires NOAA. 2019. Cold and warm episodes by season. National Weather Service of the United States. Available at: http://origin.cpc.ncep.noaa.gov/products/analysis_monitoring/ensostuff/ONI_v5.php (accessed 2019 October 27) NOAA Cold and warm episodes by season 2019 National Weather Service of the United States 2019/10/27 http://origin.cpc.ncep.noaa.gov/products/analysis_monitoring/ensostuff/ONI_v5.php Ortiz-Malavasi E. 2015. Atlas digital de Costa Rica 2014. Instituto Tecnológico de Costa Rica. Cartago, Cartago, Costa Rica. Available at http://hdl.handle.net/2238/7477 (accessed 2019 July 25) Ortiz-Malavasi E Atlas digital de Costa Rica 2014 2015 Instituto Tecnológico de Costa Rica Cartago, Cartago, Costa Rica 2019/07/25 http://hdl.handle.net/2238/7477 OVSICORI. 2015. Estado de los volcanes. Observatorio Vulcanológico y Sismológico de Costa Rica. Available at http://www.ovsicori.una.ac.cr/index.php/extension/estado-de-los-volcanes (accessed 2019 October 21) OVSICORI Estado de los volcanes 2015 Observatorio Vulcanológico y Sismológico de Costa Rica 2019/10/21 http://www.ovsicori.una.ac.cr/index.php/extension/estado-de-los-volcanes Ríos-Rojas JF, Aperador-Rodríguez D, Hernández-García EA, Arroyave-Posada CE. 2017. Annual atmospheric corrosion rate and dose-response function for carbon steel in Bogotá. Atmósfera 30:53-61. https://doi.org/10.20937/atm.2017.30.01.05 Ríos-Rojas JF Aperador-Rodríguez D Hernández-García EA Arroyave-Posada CE Annual atmospheric corrosion rate and dose-response function for carbon steel in Bogotá Atmósfera 2017 30 53 61 0.20937/atm.2017.30.01.05 Roberge PR, Klassen RD, Haberecht PW. 2002. Atmospheric corrosivity modeling-A review. Materials and Design 23:321-330. https://doi.org/10.1016/S0261-3069(01)00051-6 Roberge PR Klassen RD Haberecht PW Atmospheric corrosivity modeling-A review Materials and Design 2002 23 321 330 10.1016/S0261-3069(01)00051-6 Robles J. 2013. Evaluation of atmospheric corrosion in three geographic areas of Costa Rica by electrochemical and gravimetric techniques. Bachelor thesis, Universidad Nacional de Costa Rica. Robles J Evaluation of atmospheric corrosion in three geographic areas of Costa Rica by electrochemical and gravimetric techniques 2013 Universidad Nacional de Costa Rica Bachelor thesis Rodríguez-Yáñez JE, Garita-Arce L, Saborío-Leiva E. 2015. Mapas estimativos de la corrosión atmosférica de acero de baja aleación en Costa Rica. Cuadernos de Investigación 7:181-191. https://doi.org/10.22458/urj.v7i2.1144 Rodríguez-Yáñez JE Garita-Arce L Saborío-Leiva E Mapas estimativos de la corrosión atmosférica de acero de baja aleación en Costa Rica Cuadernos de Investigación 2015 7 181 191 10.22458/urj.v7i2.1144 SEMARNAT-INECC. 2016. Diagnóstico sobre la calidad del aire en cuencas atmosféricas de México. Instituto de Ecología y Cambio Climático de México. Mexico City, Mexico. Available at: http://biblioteca.semarnat.gob.mx/janium/Documentos/Ciga/Libros2013/CD002516.pdf (accessed 2019 January 17) SEMARNAT-INECC Diagnóstico sobre la calidad del aire en cuencas atmosféricas de México 2016 Mexico City, Mexico Instituto de Ecología y Cambio Climático de México 2019/01/17 http://biblioteca.semarnat.gob.mx/janium/Documentos/Ciga/Libros2013/CD002516.pdf Sendiña-Nadal I, Pérez-Muñuzuri, V. 2006. Fundamentos de meteorología. Galicia, Spain: Universidad Santiago de Compostela. Sendiña-Nadal I Pérez-Muñuzuri V Fundamentos de meteorología 2006 Galicia, Spain Universidad Santiago de Compostela Solano J., Villalobos R. (2000). Regiones y subregiones climáticas de Costa Rica. Instituto Meteorológico Nacional de Costa Rica. Available at: https://www.imn.ac.cr/documents/10179/20909/Regionalizaci%C3%B3n+clim%C3%A1tica+de+Costa+Rica (accessed 2018 January 17) Solano J Villalobos R Regiones y subregiones climáticas de Costa Rica 2000 Instituto Meteorológico Nacional de Costa Rica 2018/01/17 https://www.imn.ac.cr/documents/10179/20909/Regionalizaci%C3%B3n+clim%C3%A1tica+de+Costa+Rica Vera R, Troconi de Rincón O, Bagnara M, Romero N, Araya R, Ossandón S. 2017. Tropical/non-tropical marine environments impact on the behaviour of carbon steel and galvanised steel. Materials corrosion 69:614-625. https://doi.org/10.1002/maco.201709873 Vera R Troconi de Rincón O Bagnara M Romero N Araya R Ossandón S Tropical/non-tropical marine environments impact on the behaviour of carbon steel and galvanised steel Materials corrosion 2017 69 614 625 10.1002/maco.201709873 Wilks DS. 2011. Statistical Methods in the Atmospheric Sciences. United States: Academic Press. Wilks DS Statistical Methods in the Atmospheric Sciences 2011 United States Academic Press. Zárate E. 1978. Comportamiento del viento en Costa Rica. Nota técnica de investigación Nº 2. Instituto Meteorológico Nacional de Costa Rica . Zárate E Comportamiento del viento en Costa Rica. Nota técnica de investigación Nº 2 1978 Instituto Meteorológico Nacional de Costa Rica Zárate E. 1980. El viento en San José. Instituto Meteorológico Nacional de Costa Rica. Zárate E El viento en San José 1980 Instituto Meteorológico Nacional de Costa Rica