IImpact of COVID-19 on Air Quality in Israel
Sarit AgamiDepartment of EconomicsHebrew University, Mount Scopus, Jerusalem, Israelemail:[email protected] 14, 2020
Abstract
The COVID-19 pandemic has caused, in general, a sharp reduction intraffic and industrial activities. This in turn leaded to a reduction in airpollution around the world. It is important to quantity the amount of thatreduction in order to estimate the influence weight of traffic and industrialactivities over the total variation of air quality. The aim of this paper is toevaluate the impact of the COVID-19 outbreak on air pollution in Israel,which is considered one of the countries with a higher air pollution thanother Western countries. The results reveal two main findings: 1. Duringthe COVID-19 outbreak, relative to its earlier closest period, the pollutionfrom transport, based on Nitrogen oxides, had reduced by 40% on average,whereas the pollution from industrial, based on Grand-level ozone, hadincreased by 34% on average. Relative to 2019, the COVID-19 outbreakcaused a reduction in air pollution from transport and industrial as well.2. The explanation percent of the time period of COVID-19 is at most22% over the total variation of each pollutant amount.
Air pollution causes morbidity, death, and economic damage. The sources of airpollution in Israel include man made air pollution sources such as transporta-tion, power plants, factories, as well as natural sources, such as dust storms.In addition to the pollution that comes from Israel, the migration of pollutionbrings additional pollution from Europe or the desert. Other sources of pollutionare domestic air pollution, soil pollution, quarries, and the result of pesticidespraying. Severe air pollution exists in the Bay of Haifa, Tel Aviv and GushDan as well as in Jerusalem. Haifa and its surroundings have air pollution fromfactories and transport alike, when the air pollution is a product of a number offactories operating in a small area, producing high and unusual air pollution interms of the quantities and types of toxins emitted into the air. Common pollu-tants in Haifa Bay are Sulfur dioxide (SO ), Nitrogen oxides (NO x ), Particulatematter (PM), and Volatile Organic Compounds (VOCs), such as Benzene. In1 a r X i v : . [ phy s i c s . s o c - ph ] J u l el Aviv and Gush Dan, air pollution is mainly from transportation; the areasuffers from severe air pollution caused mainly by vehicles on which the trans-port is based in and around Tel Aviv-Yafo, as well as buses and trucks stuck intraffic jams, and airport. Transportation emits NO , NO x , Grand-level ozone(O ), Carbon monoxide (CO), and particulate matter 10 micrometers or less indiameter (PM ).The first tackle with the COVID-19 pandemic was reduction of the economicactivity. Particularly a sharp reduction in road traffic, air traffic, shipping andindustrial activities. This in turn leaded to a reduction in greenhouse gas emis-sions and air pollution around the world. For example, one of the largest dropsin pollution levels could be seen over the city of Wuhan, in central China, whichwas put under a strict lockdown in late January. The city of 11 million peo-ple serves as a major transportation hub and is home to hundreds of factoriessupplying car parts and other hardware to global supply chains. According toNasa, NO levels across eastern and central China have been 10%-30% lowerthan normal. NO levels also dropped in South Korea, which has long struggledwith high emissions from its large fleet of coal-fired power plants but also fromnearby industrial facilities in China.In this paper we give an initial investigation of the influence of COVID-19time period on air quality in Israel. The global COVID-19 epidemic in Israelbegan to spread towards the end of February 2020. As part of the Israeli Min-istry of Health’s deal with the epidemic, starting in mid-March, regulationsimposed on the public. At first, a local closure on places where the virus isspreading had taken, as well as movement the public sector to the emergencyand the private sector to a limited 70% service. Branches that worked in or-der - up to 100%, while reducing as much as possible, were, for example, theenergy sector (including electricity, natural gas, oil, water), food industries (in-cluding agriculture, supermarkets, transport, storage and more), all freight andstorage services, the ports and shipping companies, and workplace dealing withconstruction or infrastructure work. Industries that only some employers couldwork with 100% were, for examples, various economic sectors, and economyand industry. The public transport activity was also reduced to a quarter ofits normal size, train traffic was disabled and the number of passengers in thetaxi was limited. It was decided to ban crowds of over 10 people, and to shutdown all recreation and leisure. Next, the regulations became tougher: peoplewere ordered the avoidance of leaving home, except for emergency situationsthat required it. Later on, they asked leaving the place of residence no longerthan 100 meters and for a short time. On April 19, for the first time, easingrestrictions applied to the public took effect. In the beginning of May, relief wasadded (for example, people were allowed to go more than 100 meters). Basedon these dates, the restrictions time period is March 1 , 2020 until May 1, 2020.During writing this paper, the available data we had was until May 2, 2020.Therefore, throughout this paper, we refer to the time period of March 1 , 2020until May 2, 2020 as the COVID-19 period. In order to examine any influenceof this time period on air quality, needs to compare this period with its earlierclosest period, with its parallel period a year ago, and with its parallel period2ver various years ago. If any influence exists, we should quantity its amount,and its relative weight over the total variation of a given pollutant, as well. Inthis paper, we consider the earlier closest period to be the period of January 1,2020 until February 29, 2020. For the comparison with a year ago, we comparethe whole period of January 1, 2020 until May 2, 2020, with the same period at2019. Throughout this paper, we refer to the period of January 1 until May 2,for a given year, as the full period. We compare the full period over the years2000-2020, as well.This paper is organized as follows. Section 2 describes the data frameworkused to evaluate the air quality during the full period over the considered years.Section 3 presents the impact of the COVID-19 period on air quality, whileSection 4 examines this impact along with weather variables. Section 5 gives adiscussion and a summary of the analysis results. The data we use in this paper was downloaded from the Ministry of the Envi-ronment web [1]. We consternate on the two regions of Haifa and Gush-Dan,which, as was mentioned in Section 1, have highly air pollution in Israel. Thelist of the considered stations in Haifa and Gush-Dan, that includes their names,names in the paper, type, and area type, is described in Table 1. The generalstations are stations located in a representative area, at the height of the roofsof buildings or in open rural areas. These stations are not near specific emissionsources, such as industrial plants or roads. The transport stations are highwaystations along major transport routes. Measuring at this height makes monitor-ing of the transport stations to best represent the concentration of pollutantsexposed to pedestrians, cafes and drivers in the city.3able 1:
List of Stations . HaifaName - Full
Kiryat Haim Regavim Kiryat Shprintzak Hogim Ahuza General Haifa Igud Kfar Hasidim Nave Shaanan
Name - in Paper
Regavim Shprintzak Hogim Ahuza Igud Hasidim Shaanan
Station Type
General General General General Industrial General General
Area Type
Urban Urban Urban Urban Industrial Rural Urban
Name - Full
Nesher Park ha-Carmel Kiryat Ata Kiryat Binyamin Kiryat Tivon Kiryat Yam Atzmaut
Name - in Paper
Nesher Carmel Ata Binyamin Tivon Yam Atzmaut
Station Type
General General General General General General Traffic
Area Type
Urban Rural Urban Suburban Suburban Urban Urban
Gush-DanName - Full
Remez Rail Station Wolfson Rail Station Komemiyut Rail Station Yoseftal Rishon Lezion Amiel Ironi Dalet Levinsky
Name - in Paper
Remez Wolfson Komemiyut Yoseftal Rishon Amiel IroniD Levinsky
Station Type
Traffic Traffic Traffic Traffic Traffic Traffic Traffic Traffic
Area Type
Urban Indoor Indoor Indoor Urban Urban Urban Urban
Name - Full
Kvish 4 Yefet Yaffo Yad Avner Holon Hamashtela Petah Tikva Road Antokolsky Ehad ha-Am
Name - in Paper
Kvish4 Yefet Avner Holon Mashtela PT Antok Am
Station Type
Traffic Traffic General General General General General Traffic
Area Type
Urban Urban Urban Urban Suburban Urban Urban Urban he resolution of the data is half-hour in a day, and the considered air pol-lutants are PM (PM and PM . ), NO x , NO , NO, CO, O , SO , Benzene,Toluene, and Ethyl Benzene (EthylB). In addition, we use the weather vari-ables wind-direction (WD), wind-speed (WS), temperature (Tmp), and relative-humidity (RH). We consider at each region the pollutants and weather variableshaving no more than 10% missing values each one, at a specific station. In or-der to deal with unreasonable negative values of the pollutants, we convertednegative values smaller than -1 to zero, while we treated as missing data thenegative values that are larger than -1. The measurement units of each variableare described in Tables 2, 3.Table 2: Measurement Units of the Pollutants Variables .Variable PM PM . CO NO x NO NO O SO Benzene Toluene EthylBUnits µ g/ m µ g/ m ppm ppb ppb ppb ppb ppb ppb ppb ppbTable 3: Measurement Units of the Considered Weather Variables . Variable WD WS Tmp RHUnits deg m/sec c degrees %
We follow three comparisons in order to examine the influence of COVID-19time period on air quality:1. A comparison of COVID-19 period (March 1, 2020 until May 2, 2020) withits earlier closest period.2. A comparison of the full period (January 1 until May 2), over the years 2020and 2019.3. A comparison of the full period (January 1 until May 2), over the years2000-2020.Each comparison is based on means comparison of a given pollutant at a specificstation. The calculation of the mean omitted the missing values. For testing thesignificant of the difference, we used the t-test for two means comparison. Inorder to evaluate the relative contribution of the COVID-19 period to the totalvariation of a given pollutant, at a specific station, we used a linear regressionmodel. This model includes an indicator variable, ind , which is defined to be 0if the date of the record belongs to the earlier closest period, and 1 if the date ofthe record belongs to the COVID-19 period. That is, this variable distinguishesbetween the COVID-19 period and the other period. Then, the model is of theform Y = α + βind , where Y is the amount of a given pollutant at a specific5tation. The resulted R of the model obtains the desired result. By this we canunderstand the explanation percent of the total variation of a specific pollutantas contributed by transport and industrial activates. Here we compare, for a given pollutant at a specific station, its mean at theCOVID-19 period with its mean at the earlier closest period. Most of the pol-lutants had a reduction in their means at the COVID-19 period relative to theearlier closest period. But some of them had increasing. In most cases, thereduction/increasing was significant. We report on the significant results only.The relative change of each pollutant, which is defined as ( mean ( covid − − mean ( earlier )) /mean ( earlier )) ∗ R thatresulted by the above regression model.6able 4: COVID-19 period (March 1, 2020 - May 2, 2020) vs. its earlier closest period (January 1, 2020 -February 29, 2020), Haifa .Regavim Shprintzak Hogim Ahuza Igud Hasidim Shaanan Nesher Carmel Ata Binyamin Tivon YamNO -39.4 -35.64 -31.83 -36.62 -30.47 -30.54 -32.38 -25.69 -33.54 -21.8 -9.21 -33.65 -44.94 R R x -42.94 -38.53 -30.77 -41.19 -32.3 -27.07 -31.89 -30.13 -32.14 -22.48 -8.88 -37.17 -51.34 R -53.47 -3.24 -41.57 -27.69 9.08 -36.27 -23.8 -16.14 -24.31 25.21 R R R R R R . R R able 5: COVID-19 period (March 1, 2020 - May 2, 2020) vs. its earlier closest period (January 1, 2020 -February 29, 2020), Gush-Dan . Remez Wolfson Komemiyut Yoseftal Rishon Amiel IroniD Levinsky Kvish4 Yefet Avner Holon Mashtela PT Antok AmNO -47.6 -17.31 -44.26 -44.98 -40.91 -30.94 -39.71 -49.02 -35.58 -43.99 -47.15 -36.4 -45.39 -48.19 -40.98 -37.7 R R x -54.84 -57.05 -52.73 -48.57 -33.02 -43.6 -58.31 -44.46 -52.07 -45.15 -34.64 -46.16 -54.14 -40.47 -44.79 R R -8.6 5.99 -1.85 R R . R R number of common trends are seen, as follows. The pollutants No, NO x , and NO are all compounds of Nitrogen oxide. Gen-erally, the most common source of Nitrogen oxide air pollution is internal com-bustion engines for motor vehicles, and power plants. Particulary in Israel, themain source of nitrogen oxides is air pollution from land vehicles such as cars,buses and trucks. During the COVID-19 period, these pollutants of NO, NO x ,and NO had a consistent reduction relative to the earlier closest period, over allthe considered stations in Haifa and Gush-Dan. The reduction is, on average,-33.35 in Haifa, and -46.33 in Gush-Dan. Note that in Gush-Dan the reductionis higher than in Haifa; This is coincide with the fact that the main source of airpollution in Gush-Dan is from transport. In more details, the reduction for eachcomponent of these pollutants in Haifa is: NO : -31.21, NO: -36, NO x :-32.83,and in Gush-Dan: NO : -40.63, NO: -51.94,NO x : -47.33. The explanation per-cent of the total variation of each of these pollutants by the period indicator isranged between 0 to 0.09 in Haifa, and between 0.01 to 0.22 in Gush-Dan. Par-ticulary, the percent explanation of Nitrogen oxide by the period in Gush-Danis a little higher than in Haifa, which again coincide with the above fact. ) Ground-level ozone is a secondary pollutant created by photochemical reactionbetween primary pollutants such as nitrogen oxides, and hydrocarbons (volatileorganic compounds) in the presence of solar radiation. These pollutants areemitted from industrial pollution and pollution from transport. In contraryto Nitrogen oxide, O had increased during the COVID-19 period relative tothe closest period, over all the considered stations in Haifa and Gush-Dan. Theincreasing is, on average, 20.86 in Haifa, and 46.24 in Gush-Dan. This surprisingresult is due to increasing in promotion of transport and energy infrastructures[2], and due to increasing in home renovations [3]. That is, industrial pollution.The explanation percent of O by the period indicator is ranged between 0.07to 0.18 in Haifa, and between 0.13 to 0.22 in Gush-Dan. Carbon monoxide emissions usually occur as a result of incomplete combustionprocesses of natural gas, oil and coal. The main industrial activities contributingto carbon monoxide emissions are: metal processing, power generation, metaland coal mining, food production, gas and oil production, chemical production,cement and lime production, plaster and concrete production, oil production andrefining. Other sources of carbon monoxide emissions include human sources -transportation (vehicles, aircraft, ships), construction equipment, home heating,cigarettes and fires. And natural resources - volcanoes, forest burning, lightning.One of the most important sources of massive exposure to carbon monoxide in9umans is tobacco smoking. During the COVID-19 period, CO had decreasedrelative to the earlier closest period over all the considered stations in Gush-Dan, by -36.93 on average. This can be due to the reduction in transport. Theexplanation percent of CO by the period indicator is ranged between 0.13 to0.22. But, in the two stations in Haifa for which CO was measured, one stationhad a reduction of -11.42 in CO, with R = 0 .
05, whereas the other station hadan increasing of 4.96, with R = 0 .
01, both general stations. The last increasingcan be due to increasing in industrial pollution, as was noted above. and PM . Particulate matter has natural and human sources. Natural sources that emitparticles with volcanoes, dust storms, forest and grass fires, vegetative pollen,and seawater spray. Human particle-forming operations include combustion offossil fuel in vehicles and power plants; Industrial activities such as mining, andcombustion processes in various industries; Burning wood for heating and cook-ing, and burning vegetation (for example in felled agriculture and burning); Aswell as dust caused by soil and desertification due to non-sustainable farming.On a global average, human sources of particulate matter are only about 10% ofthe total particle in the atmosphere. PM is divided into coarse particles, PM ,and fine particles, PM2.5. The origin of the coarse PM particles may be dust(from a natural or human source such as construction), agriculture, mining,fly ash, spores and plant particles. The source of the finer respiratory parti-cles PM . is usually gases emitted as a result of fire and industrial combustionprocesses that become particles in the atmosphere. During the COVID-19 pe-riod, PM had increased relative to the closest period, over all the consideredstations in Haifa and Gush-Dan. This, can be due to increasing in industrialpollution, as was noted above. The increasing is on average 55.28 on Haifa,and 40.1 in Gush-Dan. The explanation percent of PM by the period indi-cator is very small and between 0.01 to 0.03 over Haifa and Gush-Dan. Fromthe other hand, PM . had increased in most stations except for two stations,which had reduction. Over the stations with increasing in PM . , the increasingin Haifa was 22.63 on average, and in 22.79 on average in Gush-Dan. For thetwo stations with reduction: the reduction in Haifa was -4.24 (general station),and in Gush-Dan it was -16.67 (traffic station). This is coincide with the factthat Gush-Dan is a traffic city, therefore the reduction is due to the reductionin transport, whereas the increasing in PM for most stations considered is dueto increasing in industrial pollution. The R is at most 3% over Haifa andGush-Dan. Human sources of atmospheric sulfur dioxide emissions are the use of fossil fuels- mainly coal and oil, as well as in the smelting and production processes ofmetals and minerals from sulfur-containing lead. The most important humansource of sulfur dioxide is the burning of sulfur-containing quartz fuels for home10eating, electricity generation at power plants, and motorized vehicles. In moststations in Haifa and Gush-Dan, SO decreased during the COVID-19 relativeto the closet period, but some stations had increasing in SO . The reductionis on average -28.31 on Haifa, and -5.22 in Gush-Dan. Over the stations withincreasing in SO , the increasing is on average 17.14 in Haifa, and 5.99 in Gush-Dan. The explanation percent of SO by the period indicator is very small andis 0.03 at most over Haifa and Gush-Dan. Volatile organic compounds (VOCs) are organic chemicals with high vapor pres-sure at room temperature. Man-made emission sources are divided into twomain groups: combustion processes (air pollution from transport, industrialfuel combustion, vegetation fires and more), and evaporation processes (suchas storage tanks, various products in the home or office and more). The mainsources of man-made emissions can be divided into industrial, transport (andgas stations) and emissions in the home environment. In our data, the pollutantsBenzene, Toluene, and EthylB are components of VOCs, which were measuredin one station in Haifa (general urban station). during the COVID-19, relativeto the closet period, they all had reduction of -49.5, -25.65, and -55.83, respec-tively. The explanation percent of these three pollutants is 0.13, 0.01, and 0.04,respectively.
Here we compare the full period (January 1 until May 2) over 2019 and 2020.This comparison is in terms of the relative difference in pollutant amount only,without examining the period contribution to this difference. Each compari-son is based on means comparison of a given pollutant at a specific station.The calculation of the mean omitted the missing values. The relative dif-ference for a given pollutant at a specific station is defined as ( mean − mean ) /mean . Tables A1, 2 in the Appendix present the results. Inaddition, we used the t-test for two means comparison for testing the signif-icance of the reduction/increasing for each pollutant. The resulted p -value isdescribed in Tables A1, A2 in the Appendix. For testing the significant of thedifference, we use the t-test for two means comparison. A number of commontrends are seen, as follows. In Gush-Dan, significant reduction in PM andPM . observed in all considered stations; significant reduction observed in thepollutants CO, NO , NO x , and NO in all considered stations except for onestation (Yad-Avner) which has no difference between 2019 and 2020. In Haifa,significant reduction observed in the pollutants PM , PM . , NO , NO x ,NO,and SO , but a significant increasing observed in the pollutants O , EthylB,Toluene, and Benzene. These are measures for one station only. The percent ofchange in Haifa is: NO x : -28.09, NO: -36.24, NO :-24.95, SO : -23.09, PM :-25.96, PM . : -30.32, and in Gush-Dan: NO x : -29.27, NO: -40.16, NO : -24.78,CO: -8.07, PM : -31.56, PM . : -28.38.11 .3 Comparison of the full period over the years 2000-2020 In order to understand if the observed change between 2019 and 2020 duringthe full period (January 1 until May 2) is due to COVID-19 only, or it is a partof a trend in the last years, we examined the behaviour of the pollutants in thefull period over the years 2000-2020. Figure 1 and Figure 2 present the trendlines for each pollutant at the various stations in Haifa and Gush-Dan, respec-tively. These trend lines are based on average of the measures at a specific year,where the calculation of the average omitted the missing values. Places in thegraph where the trend line is truncated are due to missing values in the wholeperiod over the specific year. For Gush-Dan, the graphs are based on measuresin the stations Remez, Amiel, IroniD, Yad-Avner, and Holon. For Haifa, therewas no unique station that had measures over the whole period over the years2000-2020. Therefore the graphs are based on mixed of stations: for the years2000-2010, the measures were taken from the station Market; for the years 2011-2020 the measures were taken from the station Azmaut, but from the stationPark Carmel when no measures in station Azmaut, and from Nesher when nomeasures in the former stations. For the PM measures, there were some yearswith measures of PM , and some other years with measures of PM . , there-fore we combined these data into one graph. That is, the graphs presented forHaifa are only sketch to see the trend of each pollutant over the years 2000-2020.12igure 1: Air pollution in Haifa over the years 2000-2020, in the period January 1 - May 2 igure 2: Air pollution in Gush-Dan over the years 2000-2020, in the period January 1 - May 2 he general trends that are seen in the graphs are as follows. For Gush-Dan, a decreasing trend is observed in NO x , NO , No; the trend in PM iszigzag, that is, jumping up and down; a decreasing trend in CO, but thereis some increasing trend in the three recent years as observed in the stationsIroniD and Yad-Avner. For Haifa, an increasing trend is observed in NO x ,NO , and No started at the year 2011, but it had decreased in the last years;increasing trend in O , which had started to decrease at the year 2016, butstill it is high; for PM, a decreasing trend started around the year 2013 andcontinue until now; some decreasing trend in SO started around the year 2007,but there is some increasing trend in the last years. For all these pollutants inHaifa there is reduction in 2020, except of O which increases. Using Multiplecomparisons of means (parametric t-test or non-parametric Wilcoxon test) for agiven pollutant for each two consecutive years, some comparisons are significantand some of them are not. That is, during the years 2000-2020, we can see someyears with a reduction in the amount of some pollutants. Looking inside theactivates that are related to transport and industry can explain these trends.The transport activities include a transfer of the diesel power plant to naturalgas, and technological improvements in the engines and fuels of trucks, busesand cars. This resulted in a marked decrease in sulfur oxides (SOx), and somedecline in nitrogen oxides (NO x ). This also can explain the reduction of airpollution in Gush Dan. However, there is no necessarily a reduction in PM andO . In addition, improvement made in the fuels and in the combustion systemin automotive engines, and catalytic converters had introduced. Therefore, theconcentration of CO in the air had been reduced. Nevertheless, most exposure tohigh concentrations of CO occurs in areas where there is a dense concentrationof slow moving vehicles, especially in busy city centers and along access routes.The reduction in air pollution from factories and power plants, especially inindustrial pollution centres as Haifa Bay, is due to a combination of legislation,public pressure and increased awareness that lead to the adoption of cleanertechnologies, cleaning and filtering technologies, beyond the use of natural gas.As a summary, overall, there is a trend of reduction in air pollution in recentyears. But although, the reduction that was observed in 2020 in some of thepollutants is even more than the decreasing trend seen before. That is, thisexpresses the influence of COVID-19 period, and its component of reductiontransport pollution, and increasing of industrial pollution. All this true whenassuming that there was no a significant change in weather, but it is mire orless the same over the recent years. That is, the change during the COVID-19period is not part of the trend but the influence of the COVID-19 period. The contribution of the period indicator for the variation of the pollutantamount, as we saw in Section 3, is 22% at most. But, needs to take into15ccount other factors that may explain this variation. These factors can explainthe difference in the pollutant amount during the COVID-19 period relative toits earlier closest period, as well. Due to the available data we have in hand,we examined the influence of the weather variables along with the indicator forperiod. We did it by using a linear regression include the indicator variable ind as in Section 3, and the weather variables wind-direction (WD), wind-speed(WS), temperature (Tmp), and relative-humidity (RH). Because of possible cor-relations between these variables, the regression model includes usually some ofthese variables and not all of them together. That is, the model is of the form Y = α + β ind + (cid:80) pj =2 β j X j , where Y is the amount of a given pollutant at aspecific station, and X j denotes the weather variables. The weather variablesincluded in each model in addition to the period indicator are presented in Ta-bles 6, 7 for Haifa and Gush-Dan, respectively. The combination of the variablesWD and WS is denoted in the tables as WDS . These tables also report the R for each model. By this we get the percent of the total variation of the pollutantas explained by the variables included in the model. The maximum explanationpercent of the total variation for each pollutant by the weather variables andthe indicator for period together is as follows: Nitrogen oxide (No, NO x , andNO ): 0.38 in Haifa, and 0.43 in Gush-Dan; Ground-level ozone (O ): 0.58 inHaifa, and 0.49 in Gush-Dan; CO: 0.26 in Haifa, and 0.21 in Gush-Dan; PM:0.11 in Haifa, and 0.10 in Gush-Dan; SO : 0.27 in Haifa, and 0.05 in Gush-Dan;VOCs: 0.33. 16able 6: Influence of COVID-19 and weather variables, period of January 1, 2020 - May 2, 2020, Haifa . Regavim Shprintzak Hogim Ahuza Igud Hasidim Shaanan Nesher Carmel Ata Binyamin Tivon YamNO wds wds wds wds,tmp wds wds wd,rh tmp wds,tmp ws,rh,tmp wds wds,tmp wds,rh R R x wds wds wds wds,tmp wds wds wds rh wds,tmp ws,rh wds wds,tmp wds,rh,tmp R wds wds wds wds wds wds,rh rh,tmp wds,rh ws,rh wd wds,tmp R wds wds wds wds wds,rh,tmp rh,tmp wds,rh,tmp wds,rh wds,rh R R . wds wds wds wds,rh,tmp rh,tmp wd,rh wds wds wd,tmp wds,rh R wds wds,tmp wds wds,rh rh wds,rh wds wds R R R R able 7: Influence of COVID-19 and weather variables, period of January 1, 2020 - May 2, 2020, Gush-Dan . Komemiyut Yoseftal Avner Holon PTNO rh,tmp rh,tmp rh,rain wd,rh wds R R x rh,tmp rh,tmp rh,rain wd,rh wds R rh wds R rh,rain wd R R . tmp R rh,tmp R Discussion
The period of COVID-19 breakout had influenced the air pollution in Israel intwo components, pollution from transport and pollution from industrial. Rel-ative to its earlier closest period, the COVID-19 period caused a reduction inpollution from transport, and an increasing in pollution from industry due toincreasing in infrastructures building and home renovations. We evaluate thepollution from transport mainly by the Nitrogen oxide (i.e., NO x , NO , andNo), and the pollution from industrial by O . Over the regions of Haifa andGush-Dan, the reduction in pollution from transport was on average between-33% to -46%, while the increasing in O was on average, between 21% to 46%.Other pollutants had the both changes of increasing and decreasing over the var-ious considered stations. That is, they captured the both influences of transportand industry. But overall, all pollutants except the VOCs decreased relative tothe same period in the last year. The influence of the period over the totalvariation of each pollutant is at most 22%, while adding weather variables re-sulted in explanation percent of 58% at most. As a conclusion, although thesharp reduction in transport, still the reduction in air pollution from it wasless than 50%, and the percent of explanation by period was not high. Addingthe weather variables increases the percent of explanation of the total variationof each pollutant, for a level which is higher than 50%. That is, there is stillhigh percent that we was not explained. Comparing with the previous period,reduction was in almost all pollutants, that is, the increasing in industrial notachieve the general reduction percent in pollution form industrial. But still thereduction was less than 50%. Two possible explanations for this. First, thereare other possible factors that may influence the air pollution, which we did notcontrol them. Second, even if we completely stop polluting, probably some ofthe pollution will sink but some will remain. That is, the decay in pollutionmay be non-linear, and therefore should try a polynomial relation for furtherresearch. 19 ppendix The following tables presents air pollution during COVID-19 period (period of January 1 , 2020 - May 2, 2020) in 2020comparing to 2019. Table 8:
Haifa Stations . Station Azmaut Regavim Regavim Park Carmel Regavim Azmaut Regavim Azmaut Regavim Azmaut Regavim NesherPollutant PM PM PM . SO SO NO x NO x NO NO NO NO O p -value < < < < < < < < < < < p -value < < < able 9: Gush-Dan Stations . Station Amiel Yad-Avner Holon Remez Remez IroniD AvnerPollutant PM PM PM . PM . CO CO CO2019 63.23 48.63 22.23 22.54 0.42 0.54 0.122020 47.44 30.08 15.71 16.36 0.36 0.53 0.13Relative Diff -24.97 -38.15 -29.33 -27.42 -14.29 -1.85 8.33 p -value < < < < < NO NO NO NO NO x NO x p -value < < < < < < < x NO x NO x NO NO NO NO NO2019 33.61 18.27 18.37 21.16 22.72 14.03 5.58 4.442020 22.20 15.27 14.87 12.67 9.12 7.81 4.67 4.66Relative Diff -33.95 -16.42 -19.05 -40.12 -59.86 -44.33 -16.31 4.95 p -value < < < < < < eferenceseferences