Next Article in Journal
Feasibility of Using a Silicon Preparation to Promote Growth of Forest Seedlings: Application to Pine (Pinus sylvestris) and Oak (Quercus robur)
Previous Article in Journal
Wood Decay Fungi Associated with Galleries of the Emerald Ash Borer
 
 
Font Type:
Arial Georgia Verdana
Font Size:
Aa Aa Aa
Line Spacing:
Column Width:
Background:
Article

Assessing Moisture Content and Its Mitigating Effect in an Urban Area Using the Land Surface Temperature–Vegetation Index Triangle Method

by
Karol Przeździecki
and
Jarosław Zawadzki
*
Faculty of Building Services, Hydro and Environmental Engineering, Warsaw University of Technology, Nowowiejska Street, 00-653 Warsaw, Poland
*
Author to whom correspondence should be addressed.
Forests 2023, 14(3), 578; https://doi.org/10.3390/f14030578
Submission received: 2 February 2023 / Revised: 3 March 2023 / Accepted: 11 March 2023 / Published: 14 March 2023

Abstract

:
Nowadays, climate change and heat extremes are becoming highly challenging problems in many cities across the globe. One of the solutions to overcome this problem is the use of vegetation, and, in particular, extending the range of overgrown areas, which are sometimes referred to as “urban green areas.” In this paper, the moisture condition and its mitigating effect on Land Surface Temperature in urban areas were examined in Warsaw, Poland, using satellite data. To do so, the so-called “Triangle Method” was employed. The triangle method is based on a Land Surface Temperature–Normalized Difference Vegetation Index (LST-NDVI) scatterplot to calculate the Temperature Vegetation Dryness Index (TVDI) and its modification–quadratic Temperature Vegetation Dryness Index (qTVDI). This article discusses, in detail, the usefulness of the triangle method for the analyses of built-up areas. The drought satellite indices TVDI and qTVDI compared with those of LST, NDVI, and NDBI (Normalized Difference Building Index). The study shows that the triangle method based on LST-NDVI scatterplot analysis is a promising tool for establishing moisture conditions over urban areas and for studying the effect of vegetation impact on urban heat islands. Detailed analysis shows that over an urban area, qTVDI shows better agreement with LST than classic TVDI.

1. Introduction

The decades-long rapid increase in the area and population of cities around the world has many undesirable effects such as depletion of natural areas, especially plant ecosystems, soil, air, and water pollution, and increase in noise level, which combined result in deterioration of the living conditions of city dwellers [1,2,3]. People living in modern cities are also facing the consequences of such problems as climate change and heat extremes in the cities. Urban heat affects cities and their residents in many ways, ranging from lethal heatwave incidents [4] to high building thermal loads and industrial cooling infrastructure, which ultimately lowers city-level energy use efficiency and impacts carbon emissions [5,6]. In cities with a high density of buildings or with large areas of impermeable surfaces and a high population density, the so-called urban heat islands (UHI) appear. These are areas with extremely high temperatures when compared to neighboring areas. In UHI areas, all these problems are rapidly escalating, making living conditions increasingly uncomfortable.
Urban greenery, including parks, forests, gardens, and meadows, allows for the alleviation of many of the above-mentioned problems [7]. It is worth emphasizing that not only extending the range of overgrown areas is important but also maintaining their good condition and wetness. Therefore, urban vegetation can be included among the key elements of sustainable urban development [8], and its presence and health are important factors affecting the quality of the urban ecosystem and, as a result, the quality of life of the urban population [9]. In particular, urban vegetation increases evapotranspiration (ET) and thus lowers the air temperature in cities [10]. It should be added that the consequence of climate change is not only more and more frequent droughts [11,12,13,14] but also a significant increase in the frequency of heat waves in cities, which additionally exacerbates the UHI phenomenon [15,16]. Urban vegetation allows them to be significantly reduced [7,8,9,10], especially in combination with proper spatial planning of the city [17]. The mitigating effect of urban vegetation is mainly based on evapotranspiration related to soil moisture, shading, and the reflection of solar radiation by trees [10,17,18]. It is worth emphasizing that the mitigating effect of urban vegetation depends on the season, in such a way that in the warmer months, the vegetation lowers the temperature in the city and increases it in the colder months [19].
Although the mitigating effect of urban vegetation on Urban Heat Island is well documented, the process of assessing the vegetation impact itself and the methodology of urban environment status analysis with respect to vegetation condition is either complicated and requires numerous data types, sources, and techniques [20] or oversimplified due to analyzing only basic methods as NDVI and/or LST determination. [3]. In the first case, one may encounter many cutting-edge research methods which are undoubtedly useful but often difficult to be reused as they require a lot of computer resources and a team of experts to conduct the analyses. In the second case, there still exists a need for improvement without overcomplicating.
In this paper, we are investigating the possibility of the moisture condition assessment and its mitigating effect on surface UHI [21] estimation in urban areas using the so-called triangle method based on satellite observations of vegetation [22,23]. Solving this issue is of great importance on an international scale, as such research can be consistently repeated also in other agglomerations. From this perspective, the analyses carried out in Warsaw are an exemplary case study.
The triangle method is based on an LST-NDVI scatterplot to calculate TVDI and its modification, namely quadratic qTVDI [24,25]. The triangle method was usually used in drought monitoring over vast agricultural areas [26] and the possibility of its direct application in an urban environment has not been investigated in detail. In this work, the authors, having solid experience in the method, aimed at studying the mitigating influence of the overall moisture conditions, resulting from the evapotranspiration of urban vegetation, on UHI. At the end of the Results and Discussion sections, the authors also addressed the important issue of why qTVDI is a better choice for urban area analysis than classical TVDI.
An additional goal of the presented research is its reproducibility using only open-source data and software. This approach could stimulate international satellite research related not only to UHI reduction but also to urban ‘humidity islands’.

2. Materials and Methods

2.1. Research Area

The research area is the city of Warsaw and its adjacent neighborhood, i.e., a 5 km buffer from the city border. Warsaw is the capital of Poland, Central Europe (Figure 1). The area of Warsaw city is about 517 km2. The dimensions of the city in the north-south direction are ca. 25 km and in the east-west direction ca. 20 km.
Warsaw has a considerably varied land cover. Ca. 48% is the residential area, but urban forests constitute ca. 15% of the city area. Urban parks and other green recreational areas cover about 10% [27,28,29]. That should satisfy the spatial heterogeneity requirement, which is necessary when using the triangle method [30]. Detailed description of the Warsaw area and comprehensive information about UHI problem in it could be found in [28]

2.2. Data

2.2.1. Landsat 8 OLI and TIRS Imagery

For the analyses, the authors used Landsat 8 OLI (Operational Land Imager) and TIRS (Thermal Infrared Sensor) imagery from EROS [31]. Data were Level-1 precision and terrain-corrected products (L1TP) [32]. These are the only necessary data to conduct the analysis, that is, to calculate all indices described in the Methods Section. The pixel size of the OLI Multispectral band is 30 m and the TIRS Thermal band is 100 m. The latter one is resampled to 30 m to match the multispectral bands. Such spatial resolution of satellite imagery allows us to perform reliable analyses of soil moisture in the studied area while ensuring their availability, spatial coherence, and low cost of observation.
The satellite images were taken from 2013 to 2020 on the following 8 dates: 20 June 2013, 6 July 2013, 22 May 2014, 7 June 2014, 10 August 2014, 24 August 2019, and 22 May 2020. In the presented analysis, images from tile 188/24 between May and August have been used. The scenes with considerable cloud coverage over the research area were filtered out.

2.2.2. Warsaw Administrative Border

The only complementary information was the vector file specifying the research area. In this study, it was an ESRI shapefile polygon file with the Warsaw Administrative border together with the 5 km buffer.

2.2.3. Urban Atlas 2018

In order to compare values of different satellite indices on different land use and land covers, the authors have used Urban Atlas 2018 [33] which provides reliable, high-resolution land use and land cover data. The data were downloaded from the Copernicus Land Monitoring Services database in *.gpkg file in version 13. Area of the layer was cropped to the extent of the research area, as shown in Figure 2.

2.3. Methods

2.3.1. Satellite Imagery Pre-Processing

The remote sensing images pre-processing included both radiometric and atmospheric corrections followed by conversion to the top of atmosphere (ToA) reflectance or temperature from a digital number using sensor metadata from a material template file. The atmospheric correction was carried out by means of the dark object selection method (DOS) [34] excluding Thermal Infrared (TIR) bands. Subsequently, satellite imagery was filtered for clouds and cloud shades using a quality assessment (QA) band.
Next, NDVI [35] and LST were calculated taking into account emissivity and radiation temperature obtained from the thermal band of Landsat 8 OLI/TIRS satellite according to [24,36,37] as well as Normalized Difference Built-up Index (NDBI [38,39].
To exclude pixels containing water bodies, the Modified Normalized Difference Water Index (MNDWI) was calculated using the following formula [38]:
M N D W I O L I = B a n d   3 B a n d   7 B a n d   3 + B a n d   7
Pixels with MNDWI values above 0.3 have been masked out as water pixels. Masking of water bodies was also carried out for TVDI and qTVDI the calculation of which is described in the next section.

2.3.2. TVDI and Quadratic TVDI

In this paper, the authors calculated two dryness indices, namely, classic TVDI proposed by Sandholt [22] and its quadratic variety, proposed and described in detail by the authors in the prior studies [24,25]. Both indices are based on the LST-NDVI scatterplot, which in general is widely used in moisture condition estimation over vast agricultural areas. TVDI and qTVDI [40] are both sensitive to soil moisture and moisture content available for vegetation and as such are using transpiration and evaporation phenomena which caused lowering LST over areas with good water availability.
The TVDI is given by the following equation:
T V D I = L S T L S T m i n N D V I L S T m a x N D V I L S T m i n N D V I ,
where LST is Land Surface Temperature in a pixel; LSTmin(NDVI) is minimum surface temperature for a given NDVI value specifying wet edge; LSTmax(NDVI) is maximum surface temperature for a given NDVI value specifying dry edge. LSTmin and LSTmax are calculated using the following formulas:
L S T m i n = b m i n + a m i n · N D V I
L S T m a x = b m a x + a m a x · N D V I
where amin, amax and bmin, bmax are slope and intercept for wet and dry edge, respectively.
Second drought index qTVDI was determined by means of the following formula:
q T V D I = L S T L S T m i n N D V I L S T m a x N D V I L S T m i n N D V I ,
where LST′min and LST′max specifying wet and dry edges, respectively, are given by the following linear equations:
L S T m i n = c m i n + b m i n · N D V I + a m i n · N D V I 2
L S T m a x = c m a x + b m a x · N D V I + a m i n · N D V I 2
Here, a′min, b′min, c′min and a′max, b′max, c′max are coefficients for polynomial equations for wet and dry edge, respectively.
A thorough description of Equations (6) and (7) could be found in [23].
Pixels for which LST edges functions have been determined were picked by an algorithm which divided the NDVI domain into subdomains with a length of 0.05. To estimate LSTmax in TVDI, only the pixels from the range of 0.2 to 0.9 were taken, as suggested in [41,42]. In the case of dry edge pixels having the highest LST value for given NDVI subdomains were taken and in the case of wet edge, those with lowest LST values.

2.3.3. Image Standardization and Reclassification

In order to compare values of the indices, the authors decided to standardize them first (Table 1). The reason for this was a need to put them on one scale and to investigate which areas are above the mean for the whole area and which pixels could be treated as outliers.
The standardization procedure could be described as follows:
z = x μ σ ,
where x is raw index layer value, µ is the mean value of index layer, and σ is the standard deviation of the values of index layer.
After the standardization, the layers were also reclassified as follows:
The reclassification process simplifies images and shows which parts of images or classes are far from the mean value.

3. Results

3.1. Edges Estimation

By way of an example, the detailed results of the estimation of wet and dry edges of the triangle scatterplot are shown in Figure 3. Both dry and wet edges were fitted to pixels taken from the scatterplot calculated using the algorithm described in the Methods Section. The pixels representing dry and wet edges obtained using linear or quadratic fitting are shown on the left (Figure 3a) and right (Figure 3b) sides, respectively. Although Figure 2 shows models established for a chosen date, for other observation dates, they are similar.
The determination coefficients for dry and wet edges for both classic and quadratic methods of TVDI edge are given in Table 2.
It is worth noting that the results shown in Table 2 are essential because R2 values allow for a convenient assessment and comparison of the fitting accuracy of the models used in determining TVDI and qTVDI by the triangle method. This makes it possible to compare the accuracy of the results obtained by various researchers using the triangle method in various environmental conditions. In particular, the high values of determination coefficients shown in Table 2 demonstrate that it is possible to calculate drought indices based on LST-NDVI scatterplots in urban areas and that in almost all cases an automated fitting process produces good quality results provided that the clouds and water bodies are properly masked out. It should be also stressed that obtained results of fitting the edges are significantly better than in [41] research, which is evidenced by high values of R2. Most of the R2 coefficients for the wet edge in the quadratic approach were between 0.45 and 0.6 and for the dry edge between 0.7 to 0.81 whereas in [3] mean R2 values were 0.83 and 0.79 for the wet edge in polynomial fitting and dry edge, respectively.
It might seem that the dry edge fitting using TVDI gives better results than using qTVDI, but it is mostly the effect of narrowing down the NDVI range from which pixels for dry edge estimation were taken. It will be shown in Section 3.5 that, when estimating moisture content over areas with low values of NDVI, qTVDI performs better than TVDI.

3.2. Analysis of LST-VI Scatterplots

Figure 3 shows LST-NDVI (Land Surface Temperature–Normalized Difference Vegetation Index) scatterplots with calculated wet and dry edges using both classical and second-degree polynomials for all dates of satellite observations. Pixel transparency on the figures was set to 5% to show a better pixel density. When analyzing these figures, it could be seen that the vast majority of the pixels lay in the area between the dry and wet edges, which additionally confirms the high quality of the analyses carried out in urban areas.
To summarize and compare the results presented above (Table 2 and Figure 4), it can be concluded that R2 coefficients for a wet edge are generally higher for quadratic fit than for a linear one. In the case of dry edge, R2 coefficients for quadratic fit may initially seem slightly lower than for linear one. However, during the linear fit of the dry edge case, the NDVI range was first narrowed down to 0.2–0.9, whereas in the case of a quadratic fit, all pixels were taken into account. Lower R2 values for the quadratic fit of the dry edges result mainly from outlying pixels located above the dry edges in the part of the scatterplot where the LST values are the highest.
It can be also seen in Figure 4 that the LST-NDVI scatterplots obtained in an urban area are significantly wider in comparison to those obtained in previous studies, e.g., [39], which carried out drought analysis using the triangle method in areas with natural vegetation or crops. The LST-NDVI scatterplots for urban areas are rather similar to triangles obtained over arid and semi-arid areas [38]. It could have been expected because urban areas have lower vegetation density, and, as a result, the NDVI range should indeed cover lower negative values than in the case of densely, or even loosely, vegetated areas.
The next stage was to analyze and discuss the obtained spatial distributions of the studied satellite indices.

3.3. Spatial Distributions of Studied Satelite Indices over the Studied Area

Figure 5 shows exemplary spatial distributions of all studied satellite indices over the research area on 10 August 2020. It can be seen in Figure 5 that higher TVDI and qTVDI values correspond to higher LST and NDBI values and lower NDVI values. It is generally true for all dates of the study. The figures for other data are available for additional materials attached to the paper. Figure 4 also shows that near the city center and other densely built-up city districts, LST, NDBI, TVDI, and qTVDI have the highest values, while NDVI has the lowest ones. Therefore, it can be concluded that urban vegetation not only contributes to the lower temperature of the land surface, which is already a well-researched fact, but also reduces dry areas in the city, which was demonstrated for the first time in this paper using satellite observations and the triangle method. Our results, therefore, show that the most unfavorable conditions affecting the comfort and health of the population occur in densely built-up districts without a significant share of the urban vegetation in the cover of the area. The best conditions for the comfort and health of the population are expected in the southeast part of the city, which is relatively densely covered with various urban vegetation, including urban parks and forests.
Thus spatial distributions of LST, TVDI, and qTVDI show which areas are particularly vulnerable to heatwaves and where UHI can be formed. They also indicate areas where activities aimed at improving the condition of vegetation should be carried out. By this saying, we meant not only planting more but also maintaining and watering existing vegetation.

3.4. Comparison of Different Classes of Coverage on the Basis of Urban Atlas 2018

The spatial distribution shown in Section 3.3, although very informative, is rather difficult to use in a detailed analysis. Because of that, the authors decided to compare indices retrieved from the image sensed on 24 August 2019 among different classes taken from an urban atlas, using zonal statistics within dissolved (aggregated) classes. Before, the aggregation cropped Urban Atlas shapefile consisted of 25,620 polygons and comprised only 25 polygons after the aggregation, which is equal to the number of classes present in the research area. Figure 6 presents maps with mean values calculated on the basis of aggregated urban atlas classes of polygons and standardized reclassified indices (as described in Section 2.3.3).
At first glance, two places clearly stand out on the spatial distributions of the studied remote indices, especially of NDBI and both types of TVDI. These two places, indicated by distinct red color, are Warsaw airports. The smaller one is Babice Airport and the other larger one is Warsaw Chopin Airport (both have a code 12,400 class according to the Airport in Urban Atlas). It should be noted that besides the fact that its LST is highest for this class, other indices such as NDBI, TVDI, and qTVDI also stand out. NDVI is not particularly low, which means that its low value does not stand out in comparison to other classes.
The juxtapositions of medians of calculated and standardized layers of studied indices in urban atlas classes are shown in Figure 7.

3.5. Comparing TVDI and qTVDI Results

TVDI and qTVDI represent similar patterns, which could have been expected. As it was mentioned above, one of the main tasks was to find out which of them would be more appropriate for satellite studies of soil moisture in urban areas. The main difference between these two indices can be clearly seen in Figure 3 and Figure 4, where the use of these indices gives distinctly different results for determining triangle edges, especially for low NDVI ranges. Narrowing down the NDVI range from 0.2 to 0.9 helps define a more pronounced triangle but also leads to an undesirable effect, which is the rejection of the low NDVI values that often occur in urban conditions.
Due to the fact, that previously the triangle method was used in mixed and agricultural areas, very low values were not very frequent. Thus, omitting them did not cause serious problems. However, in the case of urban areas, low NDVI values are frequent, and, for this reason, they are of key importance for the final results. It is worth emphasizing that, so far, this problem has not been studied and discussed in the literature.
As shown in Section 3.4, overall the mean TVDI and qTVDI values and the median closely match the mean LST, but the qTVDI showed slightly better agreement. In Figure 8, the authors presented an interesting case study of Warsaw Chopin Airport. As can be seen, plain runways, which are impregnated and made from concrete or asphalt, have lower standardized NDVI and higher standardized LST than the surrounding areas which are mostly covered by grass.
Here the difference between TVDI and qTVDI is clearly visible. Standardized TVDIs on runways have lower values, which suggests that runways were more moisturized than the surrounding lawns or had higher evaporation, which cannot be true. On the other hand, the qTVDI spatial distribution confirms the opposite, i.e., the runways are drier than the surrounding lawns. The distribution obtained by qTVDI seems logical and correct. The reason for such a situation, not previously noticed in the literature, is a polynomial construction of qTVDI, which is much more precise than a linear one, allowing an effective assessment of soil moisture content in the lower NDVI range. The presented discussion proves that in urban areas, qTVDI is a much more useful satellite index than TVDI as well as that qTVDI allows the effective use of the triangle method in typical urban areas.
It should be noted here that neither NDVI nor LST offers as good results as qTVDI. That is why it is especially promising to establish moisture conditions over urban areas.

4. Conclusions

Moisture conditions can be used as a proxy to estimate the capability of Land Surface Temperature and thus also the reduction of Heat Islands in cities. The main result of the study is that the triangle method based on LST-VI scatterplot analysis is a promising tool for establishing moisture conditions over urban areas. It should be stressed that our work presents the first effective application of the triangle method in urban areas. The second important finding is that qTVDI, which is a polynomial variation of a classic TVDI, is better suited for urban areas studies because it uses a lower range of NDVI on LST-VI scatterplots more effectively than TVDI. In the urban environment, vegetation is usually in a worse condition than in natural conditions; therefore, the satellite images are usually dominated by low NDVI values.
In this paper, the authors discussed in detail many important aspects of using the triangle method in urban conditions. Among others, the authors demonstrated that the use of the triangle method and qTVDI is useful for taking decisions on the planning of green areas in order to mitigate the urban heat islands. The triangle method, in particular its accuracy and stability, requires further research in a variety of urban conditions and in different climatic zones.
It should be also emphasized that the method presented in the paper is based on publicly available data and free software which enables us to conduct all analyses and visualize their results in many cities around the world in a coherent way, which is essential for comparative research. The results presented in the paper can therefore stimulate research on climate change and the adaptation of cities to these changes.

Author Contributions

Conceptualization, K.P. and J.Z.; methodology, K.P. and J.Z.; investigation, K.P. and J.Z.; software, K.P.; resources, K.P. and J.Z.; writing—original draft preparation, K.P. and J.Z.; review and editing, K.P. and J.Z.; supervision, J.Z.; funding acquisition K.P. and J.Z. All authors have read and agreed to the published version of the manuscript.

Funding

The research leading to these results received funding from the dean’s grant and statutory activities and of the Faculty of Building Services, Hydro and Environmental Engineering of the Warsaw University of Technology.

Data Availability Statement

The publicly available data used in this article was taken from the USGS Earth Explorer.

Conflicts of Interest

The authors declare no conflict of interest.

References

  1. United Nations. World Population Prospects, The 2019 Revision—Volume I: Comprehensive Tables; UN: New York, NY, USA, 2019; ISBN 978-92-1-004642-8. [Google Scholar]
  2. Mensah, C.A. Destruction of Urban Green Spaces: A Problem Beyond Urbanization in Kumasi City (Ghana). Am. J. Environ. Prot. 2014, 3, 1. [Google Scholar] [CrossRef]
  3. Moreno, R.; Ojeda, N.; Azócar, J.; Venegas, C.; Inostroza, L. Application of NDVI for Identify Potentiality of the Urban Forest for the Design of a Green Corridors System in Intermediary Cities of Latin America: Case Study, Temuco, Chile. Urban For. Urban Green. 2020, 55, 126821. [Google Scholar] [CrossRef]
  4. Mora, C.; Dousset, B.; Caldwell, I.R.; Powell, F.E.; Geronimo, R.C.; Bielecki, C.R.; Counsell, C.W.W.; Dietrich, B.S.; Johnston, E.T.; Louis, L.V.; et al. Global Risk of Deadly Heat. Nat. Clim. Chang. 2017, 7, 501–506. [Google Scholar] [CrossRef]
  5. Venter, Z.S.; Brousse, O.; Esau, I.; Meier, F. Hyperlocal Mapping of Urban Air Temperature Using Remote Sensing and Crowdsourced Weather Data. Remote Sens. Environ. 2020, 242, 111791. [Google Scholar] [CrossRef]
  6. Bizjak, M.; Žalik, B.; Štumberger, G.; Lukač, N. Estimation and Optimisation of Buildings’ Thermal Load Using LiDAR Data. Build. Environ. 2018, 128, 12–21. [Google Scholar] [CrossRef]
  7. Van, T.T.; Tran, N.D.H.; Bao, H.D.X.; Phuong, D.T.T.; Hoa, P.K.; Han, T.T.N. Optical Remote Sensing Method for Detecting Urban Green Space as Indicator Serving City Sustainable Development. Proceedings 2017, 2, 140. [Google Scholar] [CrossRef]
  8. Haq, S.M.A. Urban Green Spaces and an Integrative Approach to Sustainable Environment. J. Environ. Prot. 2011, 02, 601–608. [Google Scholar] [CrossRef]
  9. Liu, Y.; Meng, Q.; Zhang, J.; Zhang, L.; Jancso, T.; Vatseva, R. An Effective Building Neighborhood Green Index Model for Measuring Urban Green Space. Int. J. Digit. Earth 2016, 9, 387–409. [Google Scholar] [CrossRef]
  10. Qiu, G.; Li, H.; Zhang, Q.; Chen, W.; Liang, X.; Li, X. Effects of Evapotranspiration on Mitigation of Urban Temperature by Vegetation and Urban Agriculture. J. Integr. Agric. 2013, 12, 1307–1315. [Google Scholar] [CrossRef]
  11. Hari, V.; Rakovec, O.; Markonis, Y.; Hanel, M.; Kumar, R. Increased Future Occurrences of the Exceptional 2018–2019 Central European Drought under Global Warming. Sci. Rep. 2020, 10, 12207. [Google Scholar] [CrossRef]
  12. Zhang, R.; Zhao, C.; Ma, X.; Brindha, K.; Han, Q.; Li, C.; Zhao, X. Projected Spatiotemporal Dynamics of Drought under Global Warming in Central Asia. Sustainability 2019, 11, 4421. [Google Scholar] [CrossRef]
  13. Vicente-Serrano, S.M.; Quiring, S.M.; Peña-Gallardo, M.; Yuan, S.; Domínguez-Castro, F. A Review of Environmental Droughts: Increased Risk under Global Warming? Earth-Sci. Rev. 2020, 201, 102953. [Google Scholar] [CrossRef]
  14. Kong, D.; Gu, X.; Li, J.; Ren, G.; Liu, J. Contributions of Global Warming and Urbanization to the Intensification of Human-Perceived Heatwaves Over China. J. Geophys. Res. Atmos. 2020, 125, e2019JD032175. [Google Scholar] [CrossRef]
  15. Lemonsu, A.; Viguié, V.; Daniel, M.; Masson, V. Vulnerability to Heat Waves: Impact of Urban Expansion Scenarios on Urban Heat Island and Heat Stress in Paris (France). Urban Clim. 2015, 14, 586–605. [Google Scholar] [CrossRef]
  16. Zhang, F.; Zhang, L.-W.; Shi, J.-J.; Huang, J.-F. Soil Moisture Monitoring Based on Land Surface Temperature-Vegetation Index Space Derived from MODIS Data. Pedosphere 2014, 24, 450–460. [Google Scholar] [CrossRef]
  17. Farhadi, H.; Faizi, M.; Sanaieian, H. Mitigating the Urban Heat Island in a Residential Area in Tehran: Investigating the Role of Vegetation, Materials, and Orientation of Buildings. Sustain. Cities Soc. 2019, 46, 101448. [Google Scholar] [CrossRef]
  18. Mariani, L.; Parisi, S.G.; Cola, G.; Lafortezza, R.; Colangelo, G.; Sanesi, G. Climatological Analysis of the Mitigating Effect of Vegetation on the Urban Heat Island of Milan, Italy. Sci. Total Environ. 2016, 569–570, 762–773. [Google Scholar] [CrossRef]
  19. Chun, B.; Guldmann, J.-M. Impact of Greening on the Urban Heat Island: Seasonal Variations and Mitigation Strategies. Comput. Environ. Urban Syst. 2018, 71, 165–176. [Google Scholar] [CrossRef]
  20. Czekajlo, A.; Coops, N.C.; Wulder, M.A.; Hermosilla, T.; Lu, Y.; White, J.C.; van den Bosch, M. The Urban Greenness Score: A Satellite-Based Metric for Multi-Decadal Characterization of Urban Land Dynamics. Int. J. Appl. Earth Obs. Geoinf. 2020, 93, 102210. [Google Scholar] [CrossRef]
  21. Voogt, J.A.; Oke, T.R. Thermal Remote Sensing of Urban Climates. Remote Sens. Environ. 2003, 86, 370–384. [Google Scholar] [CrossRef]
  22. Sandholt, I.; Rasmussen, K.; Andersen, J. A Simple Interpretation of the Surface Temperature/Vegetation Index Space for Assessment of Surface Moisture Status. Remote Sens. Environ. 2002, 79, 213–224. [Google Scholar] [CrossRef]
  23. Wang, C.; Qi, S.; Niu, Z.; Wang, J. Evaluating Soil Moisture Status in China Using the Temperature–Vegetation Dryness Index (TVDI). Can. J. Remote Sens. 2004, 30, 671–679. [Google Scholar] [CrossRef]
  24. Przeździecki, K.; Zawadzki, J.J. Triangle Method Modification with Second Degree Polynomial Edges Fitting in LST NDVI Scatterplot in Lignite Mine Influence Area. Remote Sens. Appl. Soc. Environ. 2020, 20, 100404. [Google Scholar] [CrossRef]
  25. Przeździecki, K.; Zawadzki, J.J. Impact of the Variability of Vegetation, Soil Moisture, and Building Density between City. Sustainability 2023, 15, 1274. [Google Scholar] [CrossRef]
  26. Districts on Land Surface Temperature, Warsaw, Poland. Sustainability 2023, 15, 1274. [CrossRef]
  27. Bai, J.; Yu, Y.; Di, L. Comparison between TVDI and CWSI for Drought Monitoring in the Guanzhong Plain, China. J. Integr. Agric. 2017, 16, 389–397. [Google Scholar] [CrossRef]
  28. Kuchcik, M.; Dudek, W.; Błażejczyk, K.; Milewski, P.; Błażejczyk, A. Two Faces to the Greenery on Housing Estates–Mitigating Climate but Aggravating Allergy. A Warsaw Case Study. Urban For. Urban Green. 2016, 16, 170–181. [Google Scholar] [CrossRef]
  29. Gawuc, L.; Jefimow, M.; Szymankiewicz, K.; Kuchcik, M.; Sattari, A.; Struzewska, J. Statistical Modeling of Urban Heat Island Intensity in Warsaw, Poland Using Simultaneous Air and Surface Temperature Observations. IEEE J. Sel. Top. Appl. Earth Obs. Remote Sens. 2020, 13, 2716–2728. [Google Scholar] [CrossRef]
  30. Błażejczyk, K.; Kuchcik, M.; Dudek, W.; Kręcisz, B.; Błażejczyk, A.; Milewski, P.; Szmyd, J.; Pałczyński, C. Urban Heat Island and Bioclimatic Comfort in Warsaw. In Counteracting Urban Heat Island Effects in a Global Climate Change Scenario; Musco, F., Ed.; Springer International Publishing: Cham, Switzerland, 2016; pp. 305–321. ISBN 978-3-319-10424-9. [Google Scholar]
  31. Holzman, M.E.; Rivas, R.; Piccolo, M.C. Estimating Soil Moisture and the Relationship with Crop Yield Using Surface Temperature and Vegetation Index. Int. J. Appl. Earth Obs. Geoinf. 2014, 28, 181–192. [Google Scholar] [CrossRef]
  32. Department of the Interior U.S. Geological Survey. Landsat 8 (L8) Data Users Handbook; EROS: Sioux Falls, SD, USA, 2019; p. 114.
  33. Landsat Levels of Processing. Available online: https://www.usgs.gov/land-resources/nli/landsat/landsat-levels-processing (accessed on 8 April 2020).
  34. Urban Atlas 2018—Copernicus Land Monitoring Service. Available online: https://land.copernicus.eu/local/urban-atlas/urban-atlas-2018 (accessed on 30 August 2022).
  35. Chavez, P.S. Image-Based Atmospheric Corrections—Revisited and Improved. Photogramm. Eng. Remote Sens. 1996, 62, 1025–1035. [Google Scholar]
  36. Rouse, J.W.; Haas, R.H.; Schell, J.A.; Deering, D.W. Monitoring Vegetation Systems in the Great Plains with Erts. NASA Spec. Publ. 1974, 351, 309. [Google Scholar]
  37. Sobrino, J.A.; Jimenez-Munoz, J.C.; Soria, G.; Romaguera, M.; Guanter, L.; Moreno, J.; Plaza, A.; Martinez, P. Land Surface Emissivity Retrieval from Different VNIR and TIR Sensors. IEEE Trans. Geosci. Remote Sens. 2008, 46, 316–327. [Google Scholar] [CrossRef]
  38. Zha, Y.; Gao, J.; Ni, S. Use of Normalized Difference Built-up Index in Automatically Mapping Urban Areas from TM Imagery. Int. J. Remote Sens. 2003, 24, 583–594. [Google Scholar] [CrossRef]
  39. Bhatti, S.S.; Tripathi, N.K. Built-up Area Extraction Using Landsat 8 OLI Imagery. GIScience Remote Sens. 2014, 51, 445–467. [Google Scholar] [CrossRef]
  40. Liu, Y.; Wu, L.; Yue, H. Biparabolic NDVI-T s Space and Soil Moisture Remote Sensing in an Arid and Semi Arid Area. Can. J. Remote Sens. 2015, 41, 159–169. [Google Scholar] [CrossRef]
  41. Lu, Y.; Tao, H.; Wu, H. Dynamic Drought Monitoring in Guangxi Using Revised Temperature Vegetation Dryness Index. Wuhan Univ. J. Nat. Sci. 2007, 12, 663–668. [Google Scholar] [CrossRef]
  42. Liu, Y.; Yue, H. The Temperature Vegetation Dryness Index (TVDI) Based on Bi-Parabolic NDVI-Ts Space and Gradient-Based Structural Similarity (GSSIM) for Long-Term Drought Assessment Across Shaanxi Province, China (2000–2016). Remote Sens. 2018, 10, 959. [Google Scholar] [CrossRef]
Figure 1. Study area—the city of Warsaw and the 5 km buffer.
Figure 1. Study area—the city of Warsaw and the 5 km buffer.
Forests 14 00578 g001
Figure 2. Urban Atlas 2018 cropped to the extent of the research area.
Figure 2. Urban Atlas 2018 cropped to the extent of the research area.
Forests 14 00578 g002
Figure 3. Dry and wet edge fitting for both classic (a) and quadratic (b) TVDI for 25 July 2020 for the whole research area.
Figure 3. Dry and wet edge fitting for both classic (a) and quadratic (b) TVDI for 25 July 2020 for the whole research area.
Forests 14 00578 g003
Figure 4. LST-NDVI scatter plots with dry and wet edges with all pixels plotted for all analyzed dates: (a) 20 June 2013, (b) 6 July 2013, (c) 22 May 2014, (d) 7 June 2014, (e) 10 August 2014, (f) 24 August 2019, (g) 22 May 2020, (h) 25 July 2020, (i) 10 August 2020.
Figure 4. LST-NDVI scatter plots with dry and wet edges with all pixels plotted for all analyzed dates: (a) 20 June 2013, (b) 6 July 2013, (c) 22 May 2014, (d) 7 June 2014, (e) 10 August 2014, (f) 24 August 2019, (g) 22 May 2020, (h) 25 July 2020, (i) 10 August 2020.
Forests 14 00578 g004aForests 14 00578 g004b
Figure 5. Distribution maps of all analyzed indices over the research area on 10 August 2020.
Figure 5. Distribution maps of all analyzed indices over the research area on 10 August 2020.
Forests 14 00578 g005
Figure 6. Distribution maps of means of standardized, calculated, and reclassified indices on 24 August 2019, (a) LST, (b) NDVI, (c) NDBI, (d) TVDI, (e) qTVDI. [Coordinates are in meters].
Figure 6. Distribution maps of means of standardized, calculated, and reclassified indices on 24 August 2019, (a) LST, (b) NDVI, (c) NDBI, (d) TVDI, (e) qTVDI. [Coordinates are in meters].
Forests 14 00578 g006aForests 14 00578 g006b
Figure 7. Barplots of the median of standardized calculated indices on 24 August 2019: (a) LST, (b) NDVI, (c) NDBI, (d) TVDI, (e) qTVDI.
Figure 7. Barplots of the median of standardized calculated indices on 24 August 2019: (a) LST, (b) NDVI, (c) NDBI, (d) TVDI, (e) qTVDI.
Forests 14 00578 g007aForests 14 00578 g007b
Figure 8. Distribution maps of standardized LST, NDVI, TVDI, and qTVDI on 24 August 2019.
Figure 8. Distribution maps of standardized LST, NDVI, TVDI, and qTVDI on 24 August 2019.
Forests 14 00578 g008
Table 1. Reclassification table.
Table 1. Reclassification table.
Z ValueZ Reclassed
<−3−3
<−3, −2−2
<−2, −1−1
<−1, 1>0
1, 2>1
2, 3>2
>33
Table 2. The determination coefficients for wet and dry edge estimation for both classic and quadratic methods of TVDI edge estimation.
Table 2. The determination coefficients for wet and dry edge estimation for both classic and quadratic methods of TVDI edge estimation.
DateClassic TVDIQuadratic TVDI
20 June 2013Dry Edge R20.880.78
Wet Edge R20.970.97
6 July 2013Dry Edge R20.920.83
Wet Edge R20.680.90
22 May 2014Dry Edge R20.820.80
Wet Edge R20.340.77
7 June 2014Dry Edge R20.920.80
Wet Edge R20.870.90
10 August 2014Dry Edge R20.950.83
Wet Edge R20.340.80
24 August 2019Dry Edge R20.950.77
Wet Edge R20.600.81
22 May 2020Dry Edge R20.910.82
Wet Edge R20.530.62
25 July 2020Dry Edge R20.890.72
Wet Edge R20.690.85
10 August 2020Dry Edge R20.950.79
Wet Edge R20.490.87
Disclaimer/Publisher’s Note: The statements, opinions and data contained in all publications are solely those of the individual author(s) and contributor(s) and not of MDPI and/or the editor(s). MDPI and/or the editor(s) disclaim responsibility for any injury to people or property resulting from any ideas, methods, instructions or products referred to in the content.

Share and Cite

MDPI and ACS Style

Przeździecki, K.; Zawadzki, J. Assessing Moisture Content and Its Mitigating Effect in an Urban Area Using the Land Surface Temperature–Vegetation Index Triangle Method. Forests 2023, 14, 578. https://doi.org/10.3390/f14030578

AMA Style

Przeździecki K, Zawadzki J. Assessing Moisture Content and Its Mitigating Effect in an Urban Area Using the Land Surface Temperature–Vegetation Index Triangle Method. Forests. 2023; 14(3):578. https://doi.org/10.3390/f14030578

Chicago/Turabian Style

Przeździecki, Karol, and Jarosław Zawadzki. 2023. "Assessing Moisture Content and Its Mitigating Effect in an Urban Area Using the Land Surface Temperature–Vegetation Index Triangle Method" Forests 14, no. 3: 578. https://doi.org/10.3390/f14030578

Note that from the first issue of 2016, this journal uses article numbers instead of page numbers. See further details here.

Article Metrics

Back to TopTop