Spatiotemporal visibility characteristics impacted by forest and land fire over airports in sumatera and borneo Island, Indonesia

Atmospheric visibility, human health, and climate radiative forcing was adversely affected by smoke produced from large biomass burning (Saide et al. 2015, Aiken 2004). Indonesian Smoke 2015 event was the worst’s that it’s been for the last 20 years (McKirdy 2015). At least 500 Garuda Indonesia Airline and 250 Citilink flights had to be cancelled throughout September 2015, due to smoke currently affecting several areas such Jambi, Palangkaraya (Central Borneo), Pontianak (West Borneo) and Riau (Gunawan 2015). Bad weather, including low visibility, is the major factor in aircraft accidents (Jenamani, Kumar 2013). Hazard occurrence had a potential loss, herewith research is needed to support the alleviation measures that impact and affect human life (Afrianto et al. 2015).

Hazard characterization and maps plays a fundamental role in the risk assessment process and hazard management policies. Various studies have been conduct in such research: the landslide hazard analysis that very important in land planning, design and dimensioning of mitigating structure (Lari et al. 2014). The use of relatife frequencies of rockfall runout assesment, to establish hazard and risk maps at regional scale (Michoud et al. 2012). The cumulative impact of the various influencing parameters that exacerbates the seismic hazard risk of the valley to future earthquake events (Rashid et al. 2018).

Smoke is caused by suspended solid particles formed by the combustion of object. It can also directly influence horizontal visibility and creates obscured vision (CWB, n.d.). Meanwhile, the World Meteorological Organization (1992) in Heil and Goldammer (2001) define haze as a suspension of extremely small, dry particles in the atmosphere and hence does not specify a specific source. Smoke has a close relationship with aviation safety. Garuda Indonesia flight number GA 152 accident, all 234 passengers had been dead, at Medan September 26, 1997, happened in a smoky environment (Aiken 2004). Hazard weather related is one of the main causes of air accidents, excluding human errors, mechanical failures, sabotages and military operations (Janic 2000).

The impact level of the hazard on flight operation depends on the type of flight rules. Visual flight rule (VFR) require pilots to monitor weather conditions (including visibility) by vision. The best strategy for a potentially vulnerable pilot when facing a visibility weather minima situation according to flight rules is to avoid the landing (Herzegh et al. 2015). If the airport, plane, and pilot have the capacity to fly with instruments (Instrument Flight Rule/IFR), the visual limitations of the eye can be replaced by consideration of instrument information. Flying with instrument rules has less vulnerability than visual (VFR) when entering the airport with low visibility (Jenamani, Kumar 2013).

Traditionally the focus of research on aviation safety has been on analyzing accidents, investigating their causes and recomending corrective action. More recently, in addition to this reactive approach to improving aviation safety, increased emphasis has been placed on taking a proactive approach. This approach involves identifying emerging risk factors, characterizing these risk, and making recomendations with regrad to necessary improvement and what factors contributed to the accident (ICAO 2013, Oster et al. 2013). Hazard is one of the factor related to the aviation safety. In frame of proactive approach, identifying and characterizing of adverse weather (smoke visibility hazard) is one of the effort to conduct the reducing risk of the aviation.There is a need to analyse the visibility characteristics over airports in Sumatera and Borneo Island due to smoke effect caused by 2015 forest and land fire. It is useful for evaluation and taking preventive action to reduce the smoke impact on aviation safety.

The research aims to analyse the spatiotemporal impact of forest and land fire to visibility over airports in the research area. By using take-off and landing data weather observation (METAR), hotspot, and wind from numerical weather model, the research is conduct to explore monthly, daily and severity characteristic of visibility caused by smoke. This study is different from previous research because using ground actual visibility observation over airport other than remote sensing/modelling data (Xian et al. 2013, Ismanto et al. 2018) and analysing it using probability and severity analysis related to flight operation and aviation safety (Balarabe et al. 2015, Wang, Field 2004).

The study area consists of the airports in Sumatra and Kalimantan Island of Indonesia which are located 6.8°S to 6.8°N and 95°E to 120°E. The object point of the study was 47 airports that carried out actual weather observations (METAR) for aviation safety (Fig. 1). The research time period started from July 1st, 2015 at 00:00 UTC (07:00 WIB/ local time) to 31 December 2015 at 23:59 UTC (06:59 WIB/ local time). Smoke was directly reported in the METAR by weather observer, is referred to as FU (fume, France) followed by the visibility condition. The weather phenomena only limited to smoke, excluding haze phenomena. METAR is useful for flight crew prior to dispatch, enroute and adequate planning of the approach and landing. Actual weather and prevailing visibility (the greatest visibility value which is reached at least within half the horizon circle or within half of the aerodrome surface (ICAO 2011)) data has been used to be temporally and spatially analysed at each airport.

Fig. 1

The aerodrome weather observation data (METAR) varies on each airport, data availability follows the Asia Pacific regional navigation agreement. Four-digit ICAO Airport code information was downloaded from their webpage. Data density includes 30 minutes, hour or 3 hours (because of data communication problem). It is operating time distance of 6 hours, 12 hours and 24 hours. The total number of METAR observations each airport is the summation of total number of METAR observations used in this study. It used to calculate the frequentist probability of smoke occurrences.

The hotspot data was used to determine the sources of fire, which taken from the National Aeronautics and Space (LAPAN Indonesia) TERRA and AQUA satellites analysis. This daily temporal data contained the location of hotspot data (latitude and longitude) and a degree of confidence, only hotspot with the degree of confidence greater than 50%, middle confidence level hotspot (Hantson et. al. 2013), would be analysed.

Smoke from forest and land fire was strongly influenced by climate factors (including wind direction and speed) (Widodo et al. 2017). Wind from European Center for Medium-Range Weather Forecasts (ECMWF 2018) numerical weather model data, ERA5 Reanalysis, provides gridding climatological data. The average monthly winds level 925 mbar (± 762 m) data was used to analyse the propagation of smoke impact (Heil et al. 2006). This 0.5 x 0.5-degree data was downloaded from ECMWF (2018).

The research method used temporal and spatial categorical analysis of visibility severity index (VSI). First, the categories used were based on the visibility weather minima in rules of flight operations (Table 1). Second, the probability of each category, to be analysed in the monthly period, is provided according to Table 2. Visibility severity index was counted based on these two parameters, as shown in below equation. Probability was calculated based on the occurance of class-flight rule devided by total number of METAR observations in a month. Each class-flight rule have their own probability in each month. Calculation of VSI is illustrate the total hazard (visibility classes of flight rule) with their probability of occurences.

Probability theory is the body of knowledge that enables us to reason formally about uncertain events. The populist view of probability is the so-called frequentist approach whereby the probability P of an uncertain event A, written P(A), is defined by the frequency of that event based on past observations/experience (Baccini, 2001). For example in this paper, occurrence frequency of smoke happen in airport X is 70.9% at the certain month; suppose then that we are interested in the event A: ‘a randomly selected at certain time of that month on airport X, smoke is been occured’. According to the frequentist approach P(A) = 0.709.The calculation of the impact of visibility on airports was calculated based on the accumulated probability (P(h1), P(h2), P(h3), P(h4)) of each categorical visibility class. Calculated based on the equation:

VSI (Visibility Severity Index): The accumulated value of the Visibility class multiplied by the probability of events in a such period; H1, H2, H3 and H4: are hazard class based on flight rule criteria, for example: class H1 (hazard class that visibility range more than 4800 meters or 3 SM) gifted “1” as index scale because this class is lest impact on flight operation, except for VFR. The hazard index is raising related to the raising of flight operation hazard class impact (Table 1) P(h1): frequentist probability of H1 class. For example: visibility class H1 occur 60 times at observation time at certain month, the total observation time in a month are 270. Frequentist probability (P(h1)) is 60/270 (approximately 0.23). It means that hazard class H1 is rarely occurred and then to be indexed as 3 (Table 2). VSI calculation is based on all hazard visibility classes accumulation (visibility minimum criteria and frequentist probability of each class).

Aerosols from forest and land fires affected the airports’ visibility. Diurnal variation of improved visibility revealed at the period before noon until late at night (03–12 UTC, 10–19 local time). Decreased visibility dominant in the night until late afternoon. This diurnal variation arises from the variation in the production of different amounts of smoke between day and night (Saide et al. 2015); inversion level existence (Heil, Goldammer 2001); enhanced downdraft and shallower boundary layer (Ge et al. 2014) and large solar radiation differences during the day and night (variations of the atmospheric boundary layer) affecting aerosol concentrations to altitudes in the atmosphere (Tosca et al. 2011).

High to very high VSI level were revealed at airports near and/ or northern-site of a number of hotspots. The hotspots at Sumatra and Borneo Island were dominant located over 1°N–6°S. The VSI level spread until northern Sumatra, although the number of hotspots was low. The high VSI level spread to WIMA (Malikusaleh-Lhoeksumawe Airport) and to the western archipelago of Sumatra WIMB (Binaka-G.Sitoli Airport). Our finding suggests that VSI spreads impact to the airports have a close relationship with smoke transport. Revealed in Figure 5, the low-level south-easterly wind has flowed persistently during the July–November 2015, it strong propagated smoke from the south and central Sumatra to the Northern side of Sumatra. This trade wind was suggested as the dominant factor for VSI spreads, in addition (Wang et al. 2013) showed that other wind factors related to the smoke transport are land/ sea breeze, typhoon and storm over Subtropical Western Pacific, and topographic effect. While on the Island of Borneo, VSI spread tend to homogenous around the airports in the island. It reaches medium to very high VSI level. It suggests that smoke propagation was still influenced by trade wind that is south-eastern wind, south of Equator and north-eastern wind, north of Equator.

Fig. 5

Our finding revealed that the three airports, which had to have very high VSI, whose implemented instrument flight rule could, operate flight only 24 % time operation with allowed visibility minimum condition. The rest, 76% time operations, were facing visibility under 1600 m (IFR visibility below minima) at smoke climax period month. According to the CASR part 91 point 91.175. (d) No pilot operating an aircraft may land with such visibility (DGCA 2015). Flight under IFR means flight based on air traffic controller (ATC) monitoring. Before manoeuvring, aircraft must obtain a clearance from ATC to remain safely separated (Wangermann 2003). To avoid a collision, aircraft should wait until weather conditions have improved or flight into alternate airports.

While the airport with only visual flight operations (VFR), Sintang-Airport, experienced about 92% of IFR visibility below minima condition at smoke climax period month. Pilots who are operating under Visual Flight Rules (VFR) encounter with instrument meteorological conditions (IMC) have a substantial portion of the fatal accidents among general aviation aircraft (Hunter et al. 2011). Under adverse conditions/visibilitybelow minima, the pilot could not be continuing flight operations until weather conditions improve in accordance with the minimum permissible limit (DGCA 2015) or land at the nearest suitable aerodrome (ICAO 2012).