Assessment of large-scale multiple forest disturbance susceptibilities with AutoML framework: an Izmir Regional Forest Directorate case

Abstract

Disturbances such as forest fires, intense winds, and insect damage exert strong impacts on forest ecosystems by shaping their structure and growth dynamics, with contributions from climate change. Consequently, there is a need for reliable and operational methods to monitor and map these disturbances for the development of suitable management strategies. While susceptibility assessment using machine learning methods has increased, most studies have focused on a single disturbance. Moreover, there has been limited exploration of the use of “Automated Machine Learning (AutoML)” in the literature. In this study, susceptibility assessment for multiple forest disturbances (fires, insect damage, and wind damage) was conducted using the PyCaret AutoML framework in the Izmir Regional Forest Directorate (RFD) in Turkey. The AutoML framework compared 14 machine learning algorithms and ranked the best models based on AUC (area under the curve) values. The extra tree classifier (ET) algorithm was selected for modeling the susceptibility of each disturbance due to its good performance (AUC values > 0.98). The study evaluated susceptibilities for both individual and multiple disturbances, creating a total of four susceptibility maps using fifteen driving factors in the assessment. According to the results, 82.5% of forested areas in the Izmir RFD are susceptible to multiple disturbances at high and very high levels. Additionally, a potential forest disturbances map was created, revealing that 15.6% of forested areas in the Izmir RFD may experience no damage from the disturbances considered, while 54.2% could face damage from all three disturbances. The SHAP (Shapley Additive exPlanations) methodology was applied to evaluate the importance of features on prediction and the nonlinear relationship between explanatory features and susceptibility to disturbance.

Keywords

AutoML / Forest disturbances / Forest fire / Insect / Susceptibility / Wind

References

[1]	Akyuz YF, Kucukosmanoglu A (1997) Application of GIS for fire-damaged areas in Izmir. Proceedings of the XI World Forestry Congress, Turkey, 13–22 October, Vol. 1, pp.239.
[2]	Batar AK, Shibata H, Watanabe T (2021) A novel approach for forest fragmentation susceptibility mapping and assessment: a case study from the Indian Himalayan region. Remote Sens 13(20):4090. https://doi.org/10.3390/rs13204090
[3]	Bebi PSEP, Seidl R, Motta R, Fuhr M, Firm D, Krumm F, Conedera M, Ginzler C, Wohlgemuth T, Kulakowski D (2017) Changes of forest cover and disturbance regimes in the mountain forests of the Alps. For Ecol Manag 388:43–56. https://doi.org/10.1016/j.foreco.2016.10.028
[4]	Bruzón AG, Arrogante-Funes P, Arrogante-Funes F, Martín-González F, Novillo CJ, Fernández RR, Vázquez-Jiménez R, Alarcón-Paredes A, Alonso-Silverio GA, Cantu-Ramirez CA, Ramos-Bernal RN (2021) Landslide susceptibility assessment using an AutoML framework. Int J Environ Res Public Health 18(20):10971. https://doi.org/10.3390/ijerph182010971
[5]	Bugaj M, Wrobel K, Iwaniec J (2021) Model explainability using SHAP values for LightGBM predictions. In: 2021 IEEE XVIIth international conference on the perspective technologies and methods in MEMS design (MEMSTECH) (pp. 102–106). IEEE. https://doi.org/10.1109/MEMSTECH53091.2021.9468078
[6]	Chen X, Zhao W, Chen J, Qu Y, Wu D, Chen X (2021) Mapping large-scale forest disturbance types with multi-temporal CNN framework. Remote Sens 13(24):5177. https://doi.org/10.3390/rs13245177
[7]	European Commission (2013) Communication from the commission to the European parliament, the council, the European economic and social committee and the committee of the regions: a new EU forest strategy: for forests and the forest-based sectors. https://eur-lex.europa.eu/resource.html?uri=cellar:21b27c38-21fb-11e3-8d1c-01aa75ed71a1.0022.01/DOC_1&format=PDF. Accessed 5 July 2023
[8]	Corominas J, van Westen C, Frattini P, Cascini L, Malet JP, Fotopoulou S, Catani F, Van Den Eeckhaut M, Mavrouli O, Agliardi F, Pitilakis K, Winter MG, Pastor M, Ferlisi S, Tofani V, Hervás J, Smith JT (2014) Recommendations for the quantitative analysis of landslide risk. Bull Eng Geol Environ 73:209–263. https://doi.org/10.1007/s10064-013-0538-8
[9]	Dolui S (2023) Forest fire susceptibility mapping and risk assessment using integrated AHP and DEMATEL method for Purulia District, West Bengal (India). In: Water, land, and forest susceptibility and sustainability (pp 401–452). Elsevier
[10]	Fick SE, Hijmans RJ (2017) Worldclim 2: New 1-km spatial resolution climate surfaces for global land areas. Int J Climatol 37(12):4302–4315. https://doi.org/10.1002/joc.5086
[11]	FRA (2020) Global forest resources assessment 2020: terms and definitions. Food and agriculture organization of the United Nations., pp 168, Rome, Italy
[12]	Garamszegi B, Jung C, Schindler D (2022) Multispectral spaceborne proxies of predisposing forest structure attributes to storm disturbance—a case study from Germany. Forests 13(12):2114. https://doi.org/10.3390/f13122114
[13]	GDF (2023) Forestry statistics 2022. https://www.ogm.gov.tr/tr/e-kutuphane/resmi-istatistikler. Accessed 27 June 2023
[14]	Guerriero L, Di Napoli M, Novellino A, Martire DD, Rispoli C, Lee K, Bee E, Harrison A, Calcaterra D (2022) Multi-hazard susceptibility assessment using analytic hierarchy process: the Derwent Valley Mills UNESCO World Heritage Site case study (United Kingdom). J Cult Herit 55:339–345. https://doi.org/10.1016/j.culher.2022.04.009
[15]	Hadi RH, Hady HN, Hasan AM, Al-Jodah A, Humaidi AJ (2023) Improved fault classification for predictive maintenance in industrial IoT based on AutoML: a case study of ball-bearing faults. Processes 11(5):1507. https://doi.org/10.3390/pr11051507
[16]	Halvari T, Nurminen JK, Mikkonen T (2020) Testing the robustness of automl systems. arXiv preprint arXiv:2005.02649, https://doi.org/10.48550/arXiv.2005.02649
[17]	He HS, Mladenoff DJ (1999) Spatially explicit and stochastic simulation of forest-landscape fire disturbance and succession. Ecology 80(1):81–99. https://doi.org/10.1890/0012-9658(1999)080[0081:SEASSO]2.0.CO;2
[18]	Hirschmugl M, Gallaun H, Dees M, Datta P, Deutscher J, Koutsias N, Schardt M (2017) Methods for mapping forest disturbance and degradation from optical earth observation data: a review. Curr for Rep 3:32–45. https://doi.org/10.1007/s40725-017-0047-2
[19]	Hutter F, Kotthoff L, Vanschoren J (eds) (2019) Automated machine learning: methods, systems, challenges. Springer International Publishing, Cham
[20]	Iban MC, Sekertekin A (2022) Machine learning based wildfire susceptibility mapping using remotely sensed fire data and GIS: a case study of Adana and Mersin provinces. Turkey Ecol Inform 69:101647. https://doi.org/10.1016/j.ecoinf.2022.101647
[21]	Jactel H, Bauhus J, Boberg J, Bonal D, Castagneyrol B, Gardiner B, Gonzalez-Olabarria JR, Koricheva J, Meurisse N, Brockerhoff EG (2017) Tree diversity drives forest stand resistance to natural disturbances. Curr for Rep 3:223–243. https://doi.org/10.1007/s40725-017-0064-1
[22]	Kantarcioglu O, Schindler K, Kocaman S (2023) Forest fire susceptibility assessment with machine learning methods in north-east Turkiye. Int Arch Photogramm Remote Sens Spat Inf Sci 48:161–167. https://doi.org/10.5194/isprs-archives-XLVIII-M-1-2023-161-2023
[23]	Kucuk O, Bilgili E (2008) Crown fuel characteristics and fuel load estimates in young Calabrian pine (Pinus brutia Ten.) stands in Northwestern Turkey. Fresenius Environ Bull 17(12b):2226–2231
[24]	Kucukosmanoglu A, Uzmez I (2019) Forest fires in Izmir regional directorate of forestry. Appl Ecol Environ Res 17(3):6989–7001. https://doi.org/10.15666/aeer/1703_69897001
[25]	Kumar V, Faheem M, Lee KW (2022) A decade of machine learning-based predictive models for human pharmacokinetics: Advances and challenges. Drug Discov Today 27(2):529–537. https://doi.org/10.1016/j.drudis.2021.09.013
[26]	Li Z (2022) Extracting spatial effects from machine learning model using local interpretation method: an example of SHAP and XGBoost. Comput Environ Urban Syst 96:101845. https://doi.org/10.1016/j.compenvurbsys.2022.101845
[27]	Li L, Lan H, Guo C, Zhang Y, Li Q, Wu Y (2017) A modified frequency ratio method for landslide susceptibility assessment. Landslides 14:727–741. https://doi.org/10.1007/s10346-016-0771-x
[28]	Li R, Shinde A, Liu A, Glaser S, Lyou Y, Yuh B, Wong J, Amini A (2020) Machine learning–based interpretation and visualization of nonlinear interactions in prostate cancer survival. JCO Clin Cancer Inform 4:637–646. https://doi.org/10.1200/CCI.20.00002
[29]	Lindenmayer DB, Foster DR, Franklin JF, Hunter ML, Noss RF, Schmiegelow FA, Perry D (2004) Salvage harvesting policies after natural disturbance. Science 303(5662):1303–1303. https://doi.org/10.1126/science.1093438
[30]	Lundberg SM, Lee SI (2017) A unified approach to interpreting model predictions. Adv Neural Inf Process Syst 30, New York, NY, USA: Curran Associates, Inc, pp. 4765–4774, 2017, [online] Available: http://papers.nips.cc/paper/7062-a-unified-approach-to-interpreting-model-predictions.pdf
[31]	Nepstad DC, Stickler CM, Filho BS, Merry F (2008) Interactions among Amazon land use, forests and climate: prospects for a near-term forest tipping point. Philos Trans R Soc Lond Biol Sci 363(1498):1737–1746. https://doi.org/10.1098/rstb.2007.0036
[32]	?zdemir ?, ?rslü S (2019) New perspective on machine learning process: AutoML (Original in Turkish). J Manag Inf Syst 1(1):23–30
[33]	Pourghasemi HR, Kariminejad N, Amiri M, Edalat M, Zarafshar M, Blaschke T, Cerda A (2020) Assessing and mapping multi-hazard risk susceptibility using a machine learning technique. Sci Rep 10(1):3203. https://doi.org/10.1038/s41598-020-60191-3
[34]	Reichenbach P, Rossi M, Malamud BD, Mihir M, Guzzetti F (2018) A review of statistically-based landslide susceptibility models. Earth-Sci Rev 180:60–91. https://doi.org/10.1016/j.earscirev.2018.03.001
[35]	Ren H, Zhang L, Yan M, Chen B, Yang Z, Ruan L (2022) Spatiotemporal assessment of forest fire vulnerability in China using automated machine learning. Remote Sens 14(23):5965. https://doi.org/10.3390/rs14235965
[36]	Renza D, Cárdenas EA, Jaramillo CM, Weber SS, Martinez E (2021) Landslide susceptibility model by means of remote sensing images and AutoML. In: Figueroa-García JC, Díaz-Gutierrez Y, Gaona-García EE, Orjuela-Ca?ón AD (eds) Applied computer sciences in engineering: 8th workshop on engineering applications, WEA 2021, Medellín, Colombia, October 6–8, 2021, Proceedings. Springer International Publishing, Cham, pp 25–37. https://doi.org/10.1007/978-3-030-86702-7_3
[37]	Rihan M, Bindajam AA, Talukdar S et al (2023) Forest fire susceptibility mapping with sensitivity and uncertainty analysis using machine learning and deep learning algorithms. Adv Space Res 72(2):426–443. https://doi.org/10.1016/j.asr.2023.03.026
[38]	Rusk J, Maharjan A, Tiwari P, Chen TK, Shneiderman S, Turin M, Seto KC (2022) Multi-hazard susceptibility and exposure assessment of the Hindu Kush Himalaya. Sci Total Environ 804:150039. https://doi.org/10.1016/j.scitotenv.2021.150039
[39]	Sahana M, Hong H, Sajjad H, Liu J, Zhu AX (2018) Assessing deforestation susceptibility to forest ecosystem in Rudraprayag district, India using fragmentation approach and frequency ratio model. Sci Total Environ 627:1264–1275
[40]	Sar?kaya O, Ibis HM, Toprak ? (2013) The flight activity and population density of Orthotomicus erosus (Wollaston, 1857) in the Brutian pine (Pinus brutia Ten.) forests of Izmir Province, Turkey. Int J Sci Basic Appl 12:208–221
[41]	Senf C, Seidl R (2021) Mapping the forest disturbance regimes of Europe. Nat Sustain 4(1):63–70. https://doi.org/10.1016/j.scitotenv.2018.01.290
[42]	Shabani S, Pourghasemi HR, Blaschke T (2020) Forest stand susceptibility mapping during harvesting using logistic regression and boosted regression tree machine learning models. Glob Ecol Conserv 22:e00974. https://doi.org/10.1016/j.gecco.2020.e00974
[43]	Sivrikaya F, ?zcan GE, Enez K (2023) Predicting the susceptibility to Pityokteines curvidens using GIS with analytical hierarchy process and, maximum entropy models in fir forests. In: Analytic hierarchy process-models, methods, concepts, and applications. IntechOpen. https://doi.org/10.5992/intechopen.1001074
[44]	Stritih A, Senf C, Seidl R, Grêt-Regamey A, Bebi P (2021) The impact of land-use legacies and recent management on natural disturbance susceptibility in mountain forests. For Ecol Manag 484:118950. https://doi.org/10.1016/j.foreco.2021.118950
[45]	Tien Bui D, Pradhan B, Lofman O, Revhaug I (2012) Landslide susceptibility assessment in Vietnam using support vector machines, decision tree, and Naive Bayes models. Math Probl Eng 2012:974638. https://doi.org/10.1155/2012/974638
[46]	Turner MG (2010) Disturbance and landscape dynamics in a changing world. Ecology 91(10):2833–2849. https://doi.org/10.1890/10-0097.1
[47]	Ullah K, Wang Y, Fang Z, Wang L, Rahman M (2022) Multi-hazard susceptibility mapping based on convolutional neural networks. Geosci Front 13(5):101425. https://doi.org/10.1016/j.gsf.2022.101425
[48]	URL (2023) https://www.automl.org/automl/. Accessed 7 July 2023
[49]	Vaccaro L, Sansonetti G, Micarelli A (2021) An empirical review of automated machine learning. Computers 10(1):11. https://doi.org/10.3390/computers10010011
[50]	Wang H, Zhang L, Yin K, Luo H, Li J (2021) Landslide identification using machine learning. Geosci Front 12(1):351–364. https://doi.org/10.1016/j.gsf.2020.02.012
[51]	Waring J, Lindvall C, Umeton R (2020) Automated machine learning: review of the state-of-the-art and opportunities for healthcare. Artif Intell Med 104:101822. https://doi.org/10.1016/j.artmed.2020.101822
[52]	Yanar T, Kocaman S, Gokceoglu C (2020) Use of Mamdani fuzzy algorithm for multi-hazard susceptibility assessment in a developing urban settlement (Mamak, Ankara, Turkey). ISPRS Int J Geoinf 9(2):114. https://doi.org/10.3390/ijgi9020114

Funding

Izmir Katip Celebi University