Biogenetic study of the emissions of species: Pinus radiata, Eucalyptus globulus Labill and Alnus acuminata in Riobamba canton, Ecuador

Background: Air pollution is one of the biggest problems in the world, and it is generated by industrial production, vehicular flow and use of fossil fuels, leaving aside other important emission sources such as vegetation. The aim of this research is to quantify the emissions of natural volatile organic compounds produced by the forest species: Eucalyptus globulus L., Pinus radiata and Alnus acuminata in Riobamba, Ecuador. Methods: Identification of plant coverings in the years 2014 and 2017was performed using geographic information systems tools, complemented with the application of the Guenther model for the calculation of monoterpenes and other organic volatile compounds; thus, to analyze the relationship between meteorological variables and concentrations of volatile organic compounds and nitrogen dioxide per species. Results: Mathematical calculation of emissions in Riobamba showed that Eucalyptus globulus L. registered higher emissions in the years 2014-2017, followed by Pinus radiata and Alnus acuminata. These emissions are due to the vegetation cover covering each species. The analysis of volatile organic compounds in forest plantations in air is directly related to the emissions represented in the environment and correlated with the meteorological variables of temperature, global solar radiation and wind velocity. The proposed method manages to estimate concentrations of monoterpenes and volatile organic compounds for the two examined seasons, presenting the influence of the species introduced in this study such as Eucalyptus globulus L. and Pinus radiata, with a reduction in their emissions (less area found in the year 2017, with respect to 2014). However, the emission of Alnus acuminata can be quantified only in 2017, since in 2014 no records of this species were found. Conclusions: Volatile organic compound concentrations in the air are directly related to the emissions represented spatially and correlated with the meteorological variables of temperature, global solar radiation and wind velocity.


Introduction
The atmosphere contains many gases which, when presented in concentrations higher than normal, are poisonous to humans, animals and are harmful to plants; gases such as nitrogen oxides (NOx), sulphur (SOx), hydrocarbons, carbon monoxide (CO) and a wide variety of volatile organic compounds (VOCs) are considered primary pollutants, because they are emitted directly from a source. Secondary contaminants are formed by means of chemical reactions from the primary pollutants; ozone (O 3 ) is found in this group 1 . In recent years, Ecuador has been more interested in emissions of natural origin, giving rise to inventories of volatile organic compounds nationwide, obtaining 1,855,600 tons/year in 2010 2 . The Ministry of Environment, Ecuador (MEE) has developed emission inventories in the districts Ambato, Riobamba, Santo Domingo de los Colorados, Latacunga, Ibarra, Manta, Portoviejo, Esmeraldas and Milagro, giving a space to the biogenic emissions representing 3.3% of the total emissions in Riobamba 3 . Riobamba is located at an altitude of 2750 m above sea leave; it is in the Sierra Central region and constitutes the capital of Chimborazo 4 . The population of the rural areas of the Ecuadorian Highlands, including Riobamba, has been dedicated to agroforestry crops with commercial purposes 5 . Some of these plant species are exotic, which in addition to causing negative effects to the soil, emit polluting gases that react in the atmosphere, giving rise to the formation of new compounds that may have negative effects on humans 6 .
In this context, the objective of this study is to make an approximate quantification of the emissions of natural volatile organic compounds from the species Pinus radiata, Eucalyptus globulus L. and Alnus acuminata in the district, by the variation of plant coverings obtained based on spectral signatures, temperature analysis of the years 2014-2017 and application of the emission model proposed by Guenther.

Definition of monitoring plots
Based on the area occupied by each species, plots of circular form are arranged with an area of 500 m 2 each 7 applying the equation of finite populations to obtain the sample size 8 : Where: n represents the sample size; Z, 95% confidence level of = 1.96; N, study population; E, estimation error = 0.05; p, probability of success = 0.5; q, probability of failure = 0.5.
Sampling was carried out for 3 days (October 8, 9 and 10, 2018), 3750 spectral signatures of the three species under study were obtained with the Spectrum-Field Spec 4 radiometer, this in seven plots of Eucalyptus globulus L. four of Pinus radiata and four of Alnus acuminata ( Table 1).
The spectral signatures were treated statistically with SAMS 3.2 software; for the correction of jumps, the Jump Correction tool was used, which corrects the level of reflectance in the signature. The spectra that were found out of the trend of the vegetative states of the small trees (those up to 20 cm in diameter) and high trees (those exceeding 20 cm in diameter) of the three species under study, were eliminated with the help of the software Minitab 18, obtaining the standard deviation grouped to rule out significant differences between the spectra grouped by plots.

Obtaining spectral signatures
The contact probe was used to analyze the spectral signatures of plants with the spectrum-radiometer Field Spec 4, selecting five samples distributed in a plot; each sample represents a tree, from which five leaf subsamples were taken from the canopy.
Spectral signatures were analyzed using View Spec Pro 6.2 and Minitab 18 software; the consistency of the spectral signature reflectance levels is also statistically verified using the SAMS software, discarding those that do not present a similar trend to the metadata group.

Multitemporal study
The field assessment of the normalized difference vegetation index (NDVI) is calculated from the average wavelength between 640 to 670 nm and 850 to 880 nm, and in satellite images using bands 4 and 5 of the Landsat 8 Medi satellite to Equation 2 9 .
Where NIR is the atmospherically corrected reflectance corresponding to the near infrared and R is the atmospherically corrected reflectance corresponding to the red. To obtain the result, the maximum likelihood classification algorithm 11 is applied using ENVI 5.3 software. This is a comparison of the effects of the satellite image with those taken as training areas, thus assigning the pixels to the class to which they most likely belong. The resulting classification was exported to shapefile format.

Temperature study
For the study of temperature, data from the automatic meteorological stations of: ESPOCH, UNACH, San Juan, Alao, Tunshi, Quimiag, and Urbina were used. In addition, to determine the hourly temperature, linear regression of the form ax + b was used 12 . These data are interpolated with the universal kriging method 13 , generating hourly temperature maps for the years 2014-2017 20 .

Biogenic VOC (BVOC) emissions calculation
Emissions were calculated based on the temperature schedules generated for each month. It uses the biomass density values and emission factors for monoterpenes and BVOC proposed by Guenther, described in the Underlying data. Table 1.

Monoterpenes
The time emissions of monoterpenes were calculated by means of the formulas posed by Guenther 2 .
is standard emission factor of monoterpenes associated with J category soil use (µg/g.h), FBDj is density of foliar biomass of the J class of soil use (g/m 2 ), Ais area of each cell (900 m 2 ) and M(T): environmental correction factor belonging to the temperature (Equation 4).
Where: β is an empirical coefficient (0.09°K -1 ); T is leaf temperature (equal to environmental temperature in °K); T s is standard temperature (303 °K), Daily emissions are obtained using Equation 5.
Monthly emissions are obtained using Equation 6.
The calculation of the annual emissions of monoterpenes is obtained through Equation 7. Other BVOC These are calculated with Equation 8, which was also used previously for the calculation of monoterpenes, considering the variation of emission factors 2 .
Where: E BVOC (k, time) is the hourly emission of BVOC in each K cell (µg/h) and is the standard emission factor of BVOC associated with the J category of soil use (µg/g.h).
Daily, monthly and annual emissions of other volatile organic compounds are obtained by Equation 5, Equation 6 and Equation 7, respectively.

Measurement of volatile organic compounds (COV) and NO 2
The experimental application consisted of measurements of VOC and NO 2 concentrations, using the Aeroqual S-500 gas analyzer equipment between 11:00 and 15:00, due to the higher daily temperatures occurring in this range.

Statistical analysis
The analysis of the existing correlation of the meteorological variables (temperature, solar radiation and wind velocity) with VOC concentrations was performed using Pearson's productmoment correlation coefficient 14 . Two-way ANOVA with post hoc Turkey's test were used for the statistical analysis of the NO 2 concentrations, grouping the variables temperature and global solar radiation. For the graphical analysis, the moving average method of order 3 was applied to obtain a smoothing of the curves 14 . In addition, the Pearson correlation method was used to assess the linear correlation between the VOC concentrations in the plantations of each plant species with the following variables: temperature, global solar radiation and wind speed. 15 . Statistical analyses were performed using Minitab v18 (Minitab, Inc.).

Results and discussion
Evaluation of representative spectral signatures In the spectral signature of Eucalyptus globulus L., the reflectance level in the vegetative states (sapling and timber) does not differ significantly in the range comprising the NDVI; the highest peak of the sapling state has a reflectance of 72.18% and a timber of 72.27% 15 . This is due to the similarity in the structure of the leaves in the two vegetative states.
The spectral signature of Pinus radiata shows that the reflectance levels of the sapling state are slightly higher (83.82% in the highest peak), while in the timber the highest peak corresponds to 82.36% of reflectance.
In the spectral signature of Alnus acuminata, the representative spectral signature in the sapling state shows that the highest peak has 83.13% reflectance at wavelength 880 nm, which is located in the range comprising the NDVI.

Comparative analysis of the NDVI
The NDVI values obtained in the field are elevated values approximated to 1, being dense and healthy vegetation 16 . The highest value is presented in the sapling state of Alnus acuminata (0.887) and the lower result of the index corresponds to the species of Pinus radiata (0.795), whereas in the timber state, the highest value (0.808) corresponds to Eucalyptus Globulus L. and the lowest (0.819) to Pinus radiata (see Underlying data: Table 2).
Using the information from the NDVI maps of the satellite images, it was determined that in the year 2014 the minimum value (0.303) was of Pinus radiata in the timber state, and the highest in sapling state (0.622). In the year 2017, the highest value (0.537) corresponding to Pinus radiata in timber state and the lowest (0.384) is of Alnus acuminata.

Variation of vegetal cover in the years 2014-2017
The plant coverings in the years 2014-2017 for the three forest species were obtained through a supervised classification, taking advantage of the spectral difference found in the NDVI values. The reliable results were verified by means of the maximum likelihood algorithm reflected in the confusion matrix, surpassing the value of 0.85 in the Kappa coefficient, considered an almost perfect classification according to Landi and Koch 17 .
The variation in the area covered by forest species is important ( Figure 4, Figure 5), especially in Eucalyptus globulus L., which has suffered greater deforestation, and in the last three years it has decreased 469.22 ha, and Pinus radiata has reduced 228.11 ha in the same time. Since in the year 2014 no plantations were found; in 2017 there were 44.31 ha, located mainly in the Cacha parish due to the existing deforestation programs.
The annual gross deforestation in Riobamba shows that species planted for commercial purposes, such as Pinus radiata, contributes towards 76.04 ha/year of deforestation. Eucalyptus globulus L. is deforested by 156.41 ha/year, in contrast to Alnus acuminata, the area of which has increased by 14.77 ha/year. These values represent an important part of the average annual gross deforestation in Chimborazo 16 , which reaches 928 ha/year.

Temperature variations
The temperature variations with respect to time obtained from geostatistical analysis ( Figure 6) shows that in the year 2017 the average hourly temperatures are slightly higher than in the year 2014, emphasizing from 13:00 to 15:00 hours, time with the highest temperatures.
Temperature behaves similarly in the two years (Figure 7), but it is evident that in 2014 August is the month with the lowest average temperature, reaching 9.64°C, whereas, in 2017 July has the lowest average monthly temperature (10.01°C). The highest monthly average temperature values recorded in 2014 correspond to February (11.78°C); however, November 2017 has a higher value (12.16°C), thus, conditioning the emissions of natural VOC.        The total emissions of BVOC in 2017 in Eucalyptus Globulus L. was 26.29 ton/year and in Pinus radiata was 13.87 ton/year. These two species emit larger amounts of BVOC than Alnus acuminata, which only reached 0.571 ton/year.

VOC concentrations in plantations of Pinus radiata
According to Pearson correlation coefficient analysis with a confidence level of 99%, VOC concentrations in plantations of Pinus radiata. have a positive significant linear correlation that is higher with temperature (R 2 =0.725) and global solar radiation (R 2 =0.535) (Figure 12), indicating that as temperature and radiation increase, VOC emissions also increase.
Wind velocity has a significant negative linear correlation (R 2 =0.528) with VOC emissions (Figure 12), i.e., when wind velocity is lower, gases tend to accumulate in the planting area, thus increasing the concentration.
VOC concentrations in plantations of Eucalyptus globulus L. VOC concentrations in plantations in Eucalyptus globulus L. show a positive significant linear correlation with temperature variables (R 2 =0.80) and global solar radiation (R 2 =0.609), and a significant negative linear correlation with wind velocity (R 2 =-0.569) ( Figure 13).
The linear relationship between VOC and meteorological variables is lower in Pinus radiata compared to Eucalyptus globulus L., demonstrating a lesser influence of the meteorological variables on VOC emissions in Pinus radiata plantations.

VOC concentrations in plantations of Alnus acuminata
Concentrations in the air were nil; this behavior can be related to the lower presence of existing biomass and low values of emission factors, especially monoterpenes.

Analysis of NO 2 concentrations
Two-way ANOVA showed that the average concentrations of NO 2 do not differ from each other when related to the variables of temperature and solar radiation, in plantations of Pinus radiata, Eucalyptus globulus L, and Alnus acuminata (Underlying data: Table 3).
The trend between the concentrations of NO 2 and the variables temperature and global solar radiation is similar in plantations of Pinus radiata, Eucalyptus globulus L., and Alnus acuminata, demonstrating the relationship between the behavior of the climatic variables and NO 2 concentrations in an environment with low anthropogenic intervention.

Conclusions
The representative spectral signatures of each species were obtained. The reflectance values were similar in the vegetative and timber states, allowing the generalization in each species, finding that the maximum level of reflectance in Eucalyptus Globulus L. is 72.2%, in Pinus radiata is 83.8% and in Alnus acuminata is 83.1%. The spectral difference found among the species allowed the obtaining of the NDVI vegetation index, which served as a basis for an optimum dissolution among classes, identifying the exact geographical location of each plant species. Eucalyptus globulus L. is the species with the highest emissions in both years due to the greater number of plantations, followed by Pinus radiata and Alnus acuminata. At the general level, Eucalyptus globulus L. and Pinus radiata record a decrease in emissions in 2017 when compared with 2014, which is linked to deforestation; unlike Alnus acuminata, which exhibited a small increase in plantations due to existing reforestation plans, so increased in emissions in the same way.

Grant information
The author(s) declared that no grants were involved in supporting this work.

General comments:
The article is within the scope of the F1000 Research Open for Science. The abstract clearly explains the goal and the methodology as well as the flow of the paper. The method is well defined and well-illustrated as it is. The authors are using modern concepts, pertinent equations and numerical comparisons between relevant variables which are clear to understand; besides, the authors detailed the data validation process with tools for spectral signatures.
Although the authors quantify the emissions of natural volatile organic compounds produced by the forest species: L., and in Riobamba, Ecuador, in the Eucalyptus globulus Pinus radiata Alnus acuminata "Results and Discussions" the authors need to discuss the limitations of this study.

Specific comments
In Table 1, lines: 2, 4 y 5 the terms "High tress" should be replaced by "High trees".
The text included in all figures should be improved with more resolution and with uniform font sizes.
On page 5, "Measurement of volatile organic compounds (COV) and NO " should be replaced by "Measurement of volatile organic compounds (VOC) and NO ".