Assessment of military destruction in Ukraine and its consequences using remote sensing

The article raises the problem of using Earth remote sensing data to collect evidence of damages caused by the military actions of the Russian army in Ukraine. The core data set obtained by deciphering aerial photographs reflects the general current and operational situation in the affected area, reducing the subjectivity and uncertainty of damage characteristics on the ground. Earth remote sensing data visualize visible damage to the environment, which can be recognized and assessed using images of different spectral bands with appropriate resolution. Among the damage caused to the environment as a result of hostilities, according to the Earth remote sensing data, it is possible to study: degradation of land resources, destruction of natural ecosystems by fires, destruction of structures (buildings) and structures. Satellite images with a very high resolution (≤ 1 m) make it possible to quantify the degree of damage to the soil surface due to shell explosions: the amount of metal fragments and chemicals that have entered the soil, damage to the biological cover, etc. Images with a resolution of less than 30 meters in the short-wave infrared (SWIR) range allow to depict the devastating effects of forest and steppe fires caused by military actions. Using two different time frames, one before the fire and one after, will ensure the accuracy, objectivity and reliability of the evidence collected. These very high resolution Earth remote sensing images are used to investigate the destruction of structures (buildings) and constructions. The analysis of satellite images guarantees the diagnosis of the condition of the building in three degrees of destruction: completely destroyed, partially destroyed or not destroyed. In order to ensure the storage, quick search, analysis and visualization of all useful information obtained from the data of Earth remote sensing, it is necessary to use geoinformation systems.The proposed methodology for assessing military destruction is characterized by simplicity, accuracy and versatility, and does not require the use of expensive equipment.


Introduction
Constant armed conflicts on the planet throughout the entire existence of mankind lead to catastrophic consequences.So, by the end of 2021, almost 80 million people have lost their homes and been forcibly displaced around the world due to armed conflicts.The full-scale, 1254 (2023) 012132 IOP Publishing doi:10.1088/1755-1315/1254/1/012132 2 unprecedented war unleashed by the Russian Federation on the territory of Ukraine since February 24, 2022, in terms of the scale of military operations, human casualties, destruction and material losses, has virtually no analogues since the Second World War.More than 30% of the territory of Ukraine suffered losses from war-related pollution, destruction, bombings, etc.In just half a year of war, the amount of losses from Russia's military aggression, confirmed by the World Bank, amounts to more than 340 billion dollars.
In connection with this, there is an urgent issue of receiving reparations and compensations from the aggressor country for the restoration and reconstruction of Ukraine.The amount and nature of reparations must be determined in accordance with the caused damage.That is why there is a need to measure and confirm war losses.
Such information is collected on the base of the analysis of tens of thousands of public reports from citizens, government, local authorities about loss and damage across the country, as well as on the basis of public sources.However, considering the scale of the damage, it is clear that we need to look for more modern and sophisticated forms of collecting evidence of damages for legal proceedings and lawsuits against the aggressor.
One of the most promising directions in this matter is the use of Earth remote sensing technologies.Aerospace images ensure the independence and persuasiveness of the evidence obtained due to the decryption of the data of ERS.Importantly, this baseline data set reflects the overall actual and operational situation of the damage zone, reducing the subjectivity and uncertainty of damage characterization on the ground.This is facilitated by access to high-resolution images at various ranges and with a frequency of observation ranging from one to two weeks to daily monitoring.
Currently, in the literature, one can find a sufficient number of examples of the use of ERS methods in areas of armed conflicts in order to study the impact of their consequences on the geographical environment.Conventionally, all these studies can be divided into three areas: legal, applied and methodical.
Legal publications address various aspects of the use of Earth remote sensing data as evidence for legal claims.Most authors [1][2][3] point to the fact that the value that this unique form of evidence can have is still underutilized compared to its potential.The main problems associated with the use of aerial photographs as evidence in international criminal courts are the limited availability of relevant images, the cognitive bias of external organizations conducting the analysis of ERS as well as the lack of legal certainty in this area (there are no accepted forensic medical standards and methodologies).
As a result, in their works the authors consider various ways to overcome these difficulties, among which the main ones are the following: • due to the digital and technological nature of satellite evidence, it is especially important to establish its accuracy, objectivity and reliability; • the development of standardized methods of collecting, storing and interpreting the data of ERS is required; • for the use of aerial photographs as evidence in courts, it is necessary for judges to be familiar with geospatial data technologies.
Scientific publications of an applied orientation demonstrate the attention of the world public to the monitoring of armed conflicts with the aim of analyzing the observance of human rights, tracking genocides, analyzing the migration of refugees, etc.As an example, we can cite the use of remote sensing data of time series (night lights) to study the development of the crisis and refugee flows after the "Arab Spring" [4].By means of satellite analysis using Landsat 8 images (2013 and 2015), Sawalhah et al identified the impact of population growth and Syrian refugee settlements on rangeland degradation in Jordan [5].
Obtaining information in near real time provides an opportunity to develop rapid response measures in the provision of assistance and crisis management, as well as early warning systems to minimize civilian casualties in war zones [6].
The impact of armed conflicts on the ecological state of the geographical envelopes of Asia, Africa, and Latin America is considered by Hoffmann et al [2], Ordway [7], Shatnawi et al [8].Thanks to information obtained from satellite images, researchers studied how military confrontation directly affected the physical environment during the war and the post-war period.
Today, there are examples of the use of ERS data to assess damages from military conflicts by international organizations, such as the United Nations, the World Bank, Amnesty International, etc. [9].Thus, since 2003, the United Nations Satellite Center (UNOSAT) provides analysis of satellite images during humanitarian emergencies related to natural disasters, complex emergencies and military conflicts.As an example, we can cite the results of damage assessment of buildings in Kherson (Ukraine), based on images received from the WorldView-3 satellite.Analysis of images taken on October 14, 2022 revealed 88 buildings with visible damage.Of these, 4 are destroyed, 69 are substantially damaged, 4 are moderately damaged, and 11 are possibly damaged.
The methodical features of the use of ERS technologies during the monitoring of territories affected by the war were not left out of the attention of scientists.Most of the authors [10][11][12][13] note that damage assessment using satellite images is a complex, painstaking process that takes a lot of time and often requires site inspection.
In this regard, more and more researchers pay attention to the use of computer vision to identify different types of damage using aerial imagery.Using the previous stage of machine learning, it becomes possible to effectively apply the automated building damage classifier [12,14].As a practical embodiment of this approach, we can give an example of the development of the company "NeuroMarket" (Ukraine), which has created a neural network algorithm that can analyze the degree of damage caused to a building using images from satellites and drones.
Therefore, despite the significant interest in this problem in the scientific literature and world practice in the last 10 years, it should be noted that investigations on the study of environmental damage as a result of military actions are of a purely local, narrowly regional nature.
The objective of this research is to explore the potential application of Earth remote sensing techniques in areas affected by armed conflicts.The study aims to investigate and assess the impact of these conflicts on the geographical environment and the state's economy.

Different time satellite panchromatic and multispectral images with different resolutions have been used in the work (table 1). Image sources are open platforms FIRMS, EO Browser, Google Earth PRO, USGS Earth Explorer and Maxar Technologies.
It is important that all the sources of the images are official and ensure the independence and persuasiveness of the evidence obtained through the decoding of satellite images.Image processing has been carried out in the open Quantum GIS environment.

Results and discussion
A full-scale war in Ukraine causes significant damage to the state of nature and the country's economy.According to the estimates of the KSE Institute, during just six months of the war, the amount of direct lesion to the economy of Ukraine from the damage and destruction of residential and non-residential buildings and infrastructure (in monetary terms) amounted to $108.3 billion.Taking into account indirect losses (reduction in GDP, cessation of investments, outflow of labor, additional costs for defense and social support, etc.), this figure ranges from $564 billion to $600 billion.Determining all damages, documenting them, and assessing the Earth remote sensing data visualize a variety of environmental damage.But at the same time, it should be borne in mind that this is a visual monitoring method and it can be used only to assess those damages that can be visually recognized and evaluated using images of different spectrum ranges and the corresponding spatial resolution.
Thus, of all the environmental damage caused by hostilities, the following can be studied with the help of the data of ERS: degradation of land resources, destruction of natural ecosystems by fires, and destruction of structures (buildings) and constructionls.Let's consider each of these damages separately.

A. Degradation of land resources
One of the most negative and destructive consequences of military operations is the disruption of the soil ecosystem.Millions of exploding projectiles leave behind thousands of square kilometers of excavated Ukrainian land, contaminated with fragments of metal and chemicals.
Satellite images with a very high spatial resolution (≤ 1 m) make it possible to estimate the extent of disturbances in the surface soil layer.Figure 1 shows satellite images of fields in the Donetsk region, taken before and after the intervention of the Russian invaders.In the picture taken after the beginning of the military aggression, we see agricultural fields dotted with hundreds of artillery shells.Funnels of such density almost completely destroy the soil cover and make it unsuitable for use.
The size of the recess in the soil corresponds to the caliber of the projectile: 220 mm -the diameter of the funnel is about 7 m, 152 mm -4-6 m, 120 mm -2.5-3.5 m, 82 mm -about 1 m [15].Using the scale of the picture, you can easily determine not only the number of shelling, but also the caliber of the projectiles.Figure 2 shows an example of soil ecosystem disturbance calculations.On a plot of agricultural field near the town of Lyman (Donetsk region) with an area of 0.06 square km (200 m x 300 m) there are 71 funnels from explosions.Among them: 6 -from the caliber of 220 mm, 40 -152 mm, 12 -120 mm and 13 -82 mm (table 2).
Artillery projectiles produce fragments weighing 1 gram or more: caliber 82 mm -1,200-1,550 pieces, 120 mm -1,600-2,350 pieces, 152 mm -2,700-3,500 pieces, respectively [16].Using these data, it is possible to calculate that on a certain section of the field with an area of 0.06 square  km, almost two hundred thousand metal fragments with a total weight of about 2.5 tons fell into the soil.These metal fragments are dangerous.In addition to iron and carbon, they also contain sulfur and copper [15,16].Taking into account the number of projectiles and their caliber, it is possible to calculate the total amount of chemicals that have entered the environment in the studied area -this is about 35 kg of carbon, 32 kg of sulfur and 16 kg of copper (table 2).
Chemical elements that have entered the soil are very dispersed.This will contribute to their rapid entry into the soil, and from there into surface water, and in this way they enter the cycle of environmental substances and will be included in trophic chains.So, for example, copper is a heavy metal, individual compounds of which can be quite toxic [17].
In addition, all modern high-explosive and high-explosive fragmentation projectiles release an average of 1.2-1.5 cubic meters of soil per 1 kilogram of explosive [17].Based on this, as a result of the explosion of 71 shells, the mass of the explosive substance (amotol, hexane) totaled 475 kg.As a result, almost 700 m³ of soil was uprooted, disrupting its biological cover.
Such destruction of the soil cover together with other negative factors make the territories of rural lands unsuitable for use.It takes hundreds of years to restore land resources from pollution.Even now, in 300 days of the war, according to preliminary estimates by experts, no less than 10 billion dollars will be needed to restore Ukrainian land.
At the same time, it should be taken into account that satellite images can serve as indirect evidence of the negative consequences of shell explosions on the environment.During the detonation of amotol and hexogen, a number of chemical compounds are formed -carbon monoxide, carbon dioxide, nitrogen oxide, nitrogen dioxide, formaldehyde, cyanic acid vapors, nitrogen, as well as a large amount of identified and unidentified toxic organics [17].It is impossible to calculate even an approximate number of these emissions.Under their influence, surrounding soils, wood, and structures are oxidized.

B. Destruction of natural ecosystems by fires
Monitoring of fire areas is one of the most common areas of practical use of data from ERS So, for example, NASA has created Fire Information for Resource Management System (FIRMS) scale fires on the Earth's surface.The corresponding site shares near-real time data on active fires from the Moderate Resolution Imaging Spectroradiometer (MODIS) aboard the Aqua and Terra satellites and the Visible Infrared Imaging Radiometer Suite (VIIRS) aboard S-NPP and NOAA 20 (figure 3).
The resolution of Terra MODIS infrared channels is 1 km/pixel.This means that each detected fire cell is displayed as a point in the center of a 1 km × 1 km pixel.It is practically impossible to determine the objects of fires and assess the consequences of fires in images with such a spatial resolution.They can be used only to determine the spatio-temporal indicator of the fire: place and time of occurrence.In addition, with the help of these images, it is possible to record the total number of fires on a daily basis.Thus, most of the fires recorded by the Terra satellite took place in the period from mid-summer to the beginning of autumn 2022 in the territory of forests, forest belts and steppes, mainly in the zone of combat clashes.
For a detailed analysis of fires and their consequences, satellites with a resolution of less than 30 meters should be used (Landsat 8, Sentinel 2A, 2B, etc. -see table 1).
Among all natural ecosystems of Ukraine, forests suffer from hostilities mostly.The massive use of artillery and strike aircraft against military and infrastructure facilities in and near forests causes forest fires, which in dry conditions destroy tens and sometimes hundreds of square kilometers of forest.So, for example, according to the official information of the leadership of Ukraine, almost three million hectares of forests were destroyed during the eight months of the war as a result of shelling by Russian troops on the country's territory.
Mapping the destructive effects of wildfires is most effective in the short-wave infrared (SWIR)  range.For the Sentinel 2A satellite, this is a combination of channels B12, B8A, B04.In this composite, vegetation is shown in shades of green, and recently burned land is shown in brown (figure 4).For a semi-automated method of calculating the area of fire-destroyed forests, it is possible to process a raster image in the environment of a geographic information system.Figure 5 shows an example of raster classification of a satellite image of the Nedohirsky Forest in the QGIS program with subsequent area determination using the program's functionality.In order to record and calculate the consequences of forest fires, it is necessary to have two pictures taken at different times: one before the fire, the other after (figure 5).In this way, we can ensure the accuracy, objectivity and reliability of the collected evidence.The necessary combination of channels for capturing the analyzed terrain can be obtained using the EO Browser cloud platform.The functionality of the platform allows us to measure the area of destroyed forests with the usage of polygons.In order to record and calculate the consequences of forest fires, it is necessary to have two pictures taken at different times: one before the fire, the other after (figure 6).In this way, we can ensure the accuracy, objectivity and reliability of the collected evidence.The necessary combination of channels for capturing the analyzed terrain can be obtained using the EO Browser cloud platform.The functionality of the platform allows us to measure the area of destroyed forests with the usage of polygons.
Table 3 shows examples of calculations of the territories of damaged forests of Kherson and Mykolaiv regions from shelling by Russian troops using this method.In general, during the period from February 24, 2022 to the moment of the occupation of the right-bank part of Kherson region, several thousand hectares of forest were destroyed only in the territory of these two regions.Detailed monitoring is necessary to obtain more accurate data.
The consequences of agricultural land fires are recorded using a similar method.As a result of Russian shelling in Ukraine, fires were recorded on almost 8,000 km² of agricultural land with crops of various farming cultures in just six months of the war.Documentation of the fact of fires is carried out in the presence of public reports from citizens, the government, local authorities, or based on the fixation of the source of ignition from satellite images of low spatial resolution (Terra MODIS).Then images of medium spatial resolution are obtained (Landsat 8, Sentinel 2A, 2B) and the area of the fire is determined in the short-wave infrared range (SWIR) (figure 7).
Table 4 shows examples of calculations of areas with crops in Zaporizhzhia and Kherson regions.It should be noted that forest and steppe fires destroy the fertile soil layer, and along  The number of animal and plant populations is also significantly decreasing.Determining the catastrophic consequences of these destructions and estimating the economic losses is possible only by conducting special studies on the area.

C. Destruction of structures (buildings) and constructions
One of the most significant indicators of the destructive actions of the Russian army in Ukraine is the destruction and damage of civilian and military infrastructure.According to official data from Ukrinform, during 300 days of the war, the occupiers destroyed at least 412 industrial  To study the destruction of structures (buildings) and constructions, only satellite images with a very high spatial resolution (less than 1 meter -GeoEye, WorldView-3) are used.These pictures make it possible to compare the condition of buildings before and after the destruction (figure 8).
Due to the deciphering of the pictures, it is possible to assess the state of destruction of buildings.At the same time, it should be borne in mind that the analysis of satellite images, even with a very high spatial resolution, guarantees the diagnosis of the state of the building in three degrees of destruction: completely destroyed, partially destroyed or not destroyed.As an evidence base, such gradation may not be enough.In this case, the obtained data must be supplemented with additional information from verified sources (for example, an examination at the site of destruction), or using images from unmanned aerial vehicles So, for example, the companies SmartFarming and Vkursi Zemli launched the RebuildUA project, the purpose of which is to analyze and document the destruction of the infrastructure of populated areas as a result of Russia's war against Ukraine with the help of drones.The use of drones allows to collect more detailed data about the destruction: to take photos and videos of objects from different angles, from all sides of buildings.
In future, in order to ensure storage, quick search, analysis and visualization of all useful information, it is necessary to digitize it in the GIS environment.With the help of the GIS functionality on the raster image of the satellite photo, destroyed objects are recognized, their types are classified and the level of destruction is determined.Then these objects are digitized, with the creation of separate vector layers with appropriate consideration of their types and level of destruction (figure 8).In this form, the information is analyzed and stored.
Taking into account the amount of destruction in Ukraine, it is clear that recording and assessing the nature of damage as an evidence base requires a significant amount of work with the involvement of state and commercial organizations with experience in the field of practical application of ERS data.

Conclusions
1. Aerial photographs are a unique form of evidence of military destruction.The information obtained with their help reflects the general current and operational situation in the destruction zone, reducing the subjectivity and uncertainty of damage characteristics on the ground.2. Sources of satellite images are official state or commercial structures which are publicly available.This ensures the independence and persuasiveness of the obtained evidence for trials and lawsuits against the aggressor country.3. Deciphering the data of remote sensing of the Earth makes it possible to obtain an accurate and complete assessment of environmental damage due to military actions according to the following indicators: degradation of land resources, destruction of natural ecosystems by fires, and destruction of structures (buildings) and constructions.4. The proposed methodology for assessing military destruction is characterized by simplicity, accuracy and versatility, and does not require the use of expensive equipment.5.It is possible to increase the effectiveness and evidentiality of the conclusions obtained from the analysis of the ERS data by applying additional surveys at the site of destruction.

10 Figure 6 .
Figure 6.Raster classification of the satellite image of the Nedohirsky Forest in the QGIS program.The area of the burnt forest is marked by blue and black colours.

Figure 8 .
Figure 8. Visualization of buildings damage in the village of Dovgenkove (Kharkiv region) from shelling by Russian troops according to satellite images.Top left -20.08.22,top right -22.08.15.The picture below has been processed in the QGIS program [Source: Satellite Image 2022 Maxar Technologies, EPA].

Table 1 .
General indicators of the satellites which images were used in the work and the types of the determined environmental damage.
scope for obtaining further reparations and compensations from the aggressor country is a very important task for various government and scientific structures and organizations.

Table 2 .
Indicators of soil degradation of the specified territory.

Table 3 .
Areas of forests damaged by military operations based on the analysis of satellite images [Satellite: Sentinel-2.SWIR range.Source: EO Browser].

Table 4 .
Areas of agricultural crops damaged by fire from military operations according to the results of satellite image analysis [Satellite: Sentinel-2.SWIR range.Source EO Browser].For each of these facts, it is necessary to form an evidentiary base, and considering the volume of work, the ERS becomes important.