A UAV-based remote-sensing modeling framework for recognizing coarse woody debris carbon-storage strata in structurally complex subtropical forests

Abstract

Under climate change, increased inputs of coarse woody debris (CWD) have heightened the need for reliable estimates of CWD carbon storage to support forest carbon management and accounting. Conventional field-based approaches are often constrained by high labor demands and substantial time costs, while standard remote sensing faces challenges from canopy occlusion, limiting accurate CWD quantification in forests with complex stand structures. Here, we developed a stratified representation framework for identifying spatial patterns of CWD carbon storage and evaluated it in representative subtropical forests of Mount Wuyi, China. Within this framework, operational CWD carbon strata were first delineated from a feature-selected subset and then recognized using spatially available UAV-derived and topographic predictors. The results showed that (1) remote-sensing-only schemes performed poorly for direct continuous estimation in these complex stands, yielding negative cross-validated R² values (–0.134 to –0.049). By contrast, the feature-selected subset (S8), driven primarily by CWD physical attributes and elevation, best represented the observed gradient in CWD carbon storage (R² = 0.713). (2) The strata-based recognition achieved moderate overall performance (mean accuracy = 67.80%) but showed low sensitivity for the high CWD carbon stratum (recall = 25.57%–37.50%), whereas the medium and low strata were classified more consistently. (3) Reframing CWD carbon storage estimation as a strata-based recognition task provided a more feasible indirect pathway for identifying relative carbon storage gradients under structurally complex forest conditions, although discrimination of the high CWD carbon stratum remained limited. Overall, this framework provides a strata-based screening workflow for identifying spatial patterns of CWD carbon storage under structurally complex forest conditions.

Keywords

Woody debris / Carbon pools / Machine learning / Remote sensing / Mount Wuyi

References

[1] Agee JK, Skinner CN. Basic principles of forest fuel reduction treatments. For Ecol Manag, 2005, 211: 83-96

[2] Ali I, Greifeneder F, Stamenkovic J, Neumann M, Notarnicola C. Review of machine learning approaches for biomass and soil moisture retrievals from remote sensing data. Remote Sens, 2015, 7(1216398-16421

[3] Caliński T, Harabasz J. A dendrite method for cluster analysis. Commun Stat, 1974, 3(1): 1-27

[4] Cao Y, You WB, Wang FY, Wu LY, He DJ. Research progress on carbon storage of coarse woody debris in forest ecosystems. Acta Ecol Sin, 2021, 41(20): 7913-7927

[5] Chen S, Feng Z, Chen P, Khan TU, Lian Y. Nondestructive estimation of the above-ground biomass of multiple tree species in boreal forests of China using terrestrial laser scanning. Forests, 2019, 10(11): 936

[6] Cheng Z, Aakala T, Ji C, Larjavaara M. Disturbance dynamics and its effects on carbon in human-impacted mountain forests in northwestern Yunnan, China. Ecol Evol, 2025, 15(9 e72165

[7] de Conto T, Armston J, Dubayah R. Characterizing the structural complexity of the Earth’s forests with spaceborne lidar. Nat Commun, 2024, 15 8116

[8] Dietenberger S, Mueller MM, Stöcker B, Dubois C, Arlaud H, Adam M, Hese S, Meyer H, Thiel C. Accurate mapping of downed deadwood in a dense deciduous forest using UAV-SfM data and deep learning. Remote Sens, 2025, 17(9 1610

[9] Dixon P. VEGAN, a package of R functions for community ecology. J Veg Sci, 2003, 14(6): 927-930

[10] Duan Z, Zhao D, Zeng Y, Zhao Y, Wu B, Zhu J. Assessing and correcting topographic effects on forest canopy height retrieval using airborne LiDAR data. Sensors (Basel), 2015, 15(6): 12133-12155

[11] Fan X, Zang Z, Tang J, Zhao L, Xu W, Ouyang Z. Ecological integrity assessment system for Wuyishan national park. Ecol Indic, 2025, 178 113910

[12] Feng ZM, Li WJ, Li P, Xiao CW. Relief degree of land surface and its geographical meanings in the Qinghai-Tibet Plateau. China. Acta Geogr Sin, 2020, 75(7): 1359-1372

[13] Filho AF, Machado SA, Carneiro MRA. Testing accuracy of log volume calculation procedures against water displacement techniques (xylometer). Can J for Res, 2000, 30(6): 990-997

[14] Francazi E, Baity-Jesi M, Lucchi A. 2023 A theoretical analysis of the learning dynamics under class imbalance. In: Proceedings of the 40th international conference on machine learning. proceedings of machine learning research, vol. 202, pp. 10285–10322. https://proceedings.mlr.press/v202/francazi23a.html

[15] Gale N. The aftermath of tree death: coarse woody debris and the topography in four tropical rain forests. Can J for Res, 2000, 30(9): 1489-1493

[16] Guo ZD, Hu HF, Li P, Li NY, Fang JY. Spatio-temporal changes in biomass carbon sinks in China's forests during 1977–2008. Sci China Life Sci, 2013, 56: 661-671

[17] Hamraz H, Contreras MA, Zhang J. Forest understory trees can be segmented accurately within sufficiently dense airborne laser scanning point clouds. Sci Rep, 2017, 7 6770

[18] Harmon ME, Franklin JF, Swanson FJ, Sollins P, Gregory SV, Lattin JD, Anderson NH, Cline SP, Aumen NG, Sedell JR, Lienkaemper GW, Cromack KJr, Cummins KW. Ecology of coarse woody debris in temperate ecosystems. Adv Ecol Res, 1986, 15: 133-302

[19] Harmon ME, Woodall CW, Fasth B, Sexton J. 2008 Woody detritus density and density reduction factors for tree species in the United States: a synthesis. Gen Tech Rep NRS-29. U.S. Department of Agriculture, Forest Service, Northern Research Station, Newtown Square, PA, p. 84. https://doi.org/10.2737/NRS-GTR-29

[20] Hartmann H, Bastos A, Das AJ, Esquivel-Muelbert A, Hammond WM, Martínez-Vilalta J, McDowell NG, Powers JS, Pugh TAM, Ruthrof KX, Allen CD. Climate change risks to global forest health: emergence of unexpected events of elevated tree mortality worldwide. Annu Rev Plant Biol, 2022, 73: 673-702

[21] He H, Garcia EA. Learning from imbalanced data. IEEE Trans Knowl Data Eng, 2009, 21(9): 1263-1284

[22] Helbach J, Frey J, Messier C, Mörsdorf M, Scherer-Lorenzen M. Light heterogeneity affects understory plant species richness in temperate forests supporting the heterogeneity-diversity hypothesis. Ecol Evol, 2022, 12(2 e8534

[23] Hirschmugl M, Lippl F, Sobe C. Assessing the vertical structure of forests using airborne and spaceborne LiDAR data in the Austrian Alps. Remote Sens, 2023, 15(3 664

[24] Hoffman CM, Ziegler JP, Tinkham WT, Hiers JK, Hudak AT. A comparison of four spatial interpolation methods for modeling fine-scale surface fuel load in a mixed conifer forest with complex terrain. Fire, 2023, 6(6): 216

[25] Hopkinson C, Chasmer LE. 2007 Modelling canopy gap fraction from lidar intensity. In: ISPRS Workshop on Laser Scanning 2007 and SilviLaser 2007, Espoo, Finland, September 12–14, pp. 190–194.

[26] Hu TY, Liu XQ, Wu XY, Niu CY, Su YJ. Advance in forest canopy structural complexity research. Natl Remote Sens Bull, 2025, 29(1): 83-101

[27] Huang XF, Kang XG, Yang H, Zhao HY, Jiang F, Wang JZ, Wan M. Comparison of 5 tree competition index models. J Northwest a&f Univ (Nat Sci Ed), 2012, 40(7): 127-134

[28] Husch B, Beers TW, Kershaw JA. Forest mensuration, 2002, Hoboken, New Jersey, John Wiley & Sons

[29] Ikotun AM, Habyarimana F, Ezugwu AE. Cluster validity indices for automatic clustering: a comprehensive review. Heliyon, 2025, 11(2 e41953

[30] Jiang H, Wang XQ, Wu B, Chen YZ. A topography-adjusted vegetation index (TAVI) and its application in vegetation fraction monitoring. J Fuzhou Univ (Nat Sci Ed), 2010, 38(4): 527-532

[31] Johnson JM, Khoshgoftaar TM. Survey on deep learning with class imbalance. J Big Data, 2019, 6: 27

[32] Jonckheere I, Fleck S, Nackaerts K, Muys B, Coppin P, Weiss M, Baret F. Review of methods for in situ leaf area index determination: Part I. Theories, sensors and hemispherical photography. Agric for Meteorol, 2004, 121(1–2): 19-35

[33] Joshi N, Mitchard ETA, Brolly M, Schumacher J, Fernández-Landa A, Johannsen VK, Marselis M, Fensholt R. Understanding 'saturation' of radar signals over forests. Sci Rep, 2017, 7: 3505

[34] Komposch A, Ensslin A, Fischer M, Hemp A. Aboveground deadwood biomass and composition along elevation and land-use gradients at Mount Kilimanjaro. Front Ecol Evol, 2022, 9 732092

[35] Korhonen L, Korpela I, Heiskanen J, Maltamo M. Airborne discrete-return lidar data in the estimation of vertical canopy cover, angular canopy closure and leaf area index. Remote Sens Environ, 2011, 115(4): 1065-1080

[36] Laliberté E, Legendre P. A distance-based framework for measuring functional diversity from multiple traits. Ecology, 2010, 91: 299-305

[37] Li JX. The study on characteristics of coarse woody debris of Larix gmelinii in the Great Hing'an Mountains, 2019, Harbin, Northeast For Univ

[38] Li MY, Huang WQ, Liu M. Construction of forest biodiversity monitoring network based on continuous forest inventory system: a case study in Hunan Province. For Resour Manage, 2012, 5: 51-58

[39] Li Z, Lin W, Guan Z, Zhang J, Chen S, You W. Tree growth and mortality of secondary evergreen broadleaved and temperate coniferous forests and their drivers along elevation gradients in subtropical mountain of China. J for Res, 2025, 36: 17

[40] Lin CG. The problem of stem form control in timber volume estimation techniques. Sci Silvae Sin, 1964, 9(4): 1-9

[41] Lin C, Ma SE, Huang LP, Chen CI, Lin PT, Yang ZK, Lin KT. Generating a baseline map of surface fuel loading using stratified random sampling inventory data through cokriging and multiple linear regression methods. Remote Sens, 2021, 13(8): 1561

[42] Liu WY, Xie SC, Xie KJ, Yang GP. Preliminary studies on the litterfall and coarse woody debris in mid-mountain humid evergreen broad-leaved forest in Ailao Mountains. Acta Bot Sin, 1995, 37(10): 807-814

[43] Lopes Queiroz G, McDermid GJ, Linke J, Hopkinson C, Kariyeva J. Estimating coarse woody debris volume using image analysis and multispectral LiDAR. Forests, 2020, 11(2): 141

[44] Lu D, Chen Q, Wang G, Liu L, Li G, Moran E. A survey of remote sensing-based aboveground biomass estimation methods in forest ecosystems. Int J Digit Earth, 2016, 9(1): 63-105

[45] Ni WL, Chang BB, Zhang JL, Li ZR, Lin WJ, Zou XF, You WB. Spatial point pattern of coarse woody debris in temperate coniferous forest along an elevation gradient in Mount Wuyi. J Cent South Univ for Technol, 2025, 45(11): 66-77

[46] Paletto A, Tosi V. Deadwood density variation with decay class in seven tree species of the Italian Alps. Scand J for Res, 2010, 25(2): 164-173

[47] Panagiotidis D, Abdollahnejad A, Surový P, Kuželka K. Detection of fallen logs from high-resolution UAV images. N Z J for Sci, 2019, 49: 1-11

[48] Pirotti F. Analysis of full-waveform LiDAR data for forestry applications: a review of investigations and methods. iForest, 2011, 4(3): 100-106

[49] Richardson JJ, Moskal LM. An integrated approach for monitoring contemporary and recruitable large woody debris. Remote Sens, 2016, 8(9): 778

[50] Rousseeuw PJ. Silhouettes: a graphical aid to the interpretation and validation of cluster analysis. J Comput Appl Math, 1987, 20: 53-65

[51] Roussel JR, Auty D, Coops NC, Tompalski P, Goodbody TRH, Meador AS, Bourdon JF, Boissieu FD, Achim A. LidR: an R package for analysis of airborne laser scanning (ALS) data. Remote Sens Environ, 2020, 251 112061

[52] Rubino DL, McCarthy BC. Evaluation of coarse woody debris and forest vegetation across topographic gradients in a southern Ohio forest. For Ecol Manage, 2003, 183(1–3): 221-238

[53] Sheng C, Wang Y, Ye S. Discussion on the national park management mode under China's National Park System: a case of Wuyishan National Park System Pilot Area. Int J Geoheritage Parks, 2020, 8(4): 230-234

[54] Shokirov S, Schaefer M, Levick SR, Jucker T, Borevitz J, Abdurahmanov I, Youngentob K. Multi-platform lidar approach for detecting coarse woody debris in a landscape with varied ground cover. Int J Remote Sens, 2021, 42(24): 9324-9350

[55] Smith WH, Mullins CZ. Light detection and ranging (LiDAR)-assisted detection of rock outcrops in Appalachian hardwood forests. J Fish Wildl Manage, 2022, 13(1): 250-261

[56] Sollins P. Input and decay of coarse woody debris in coniferous stands in western Oregon and Washington. Can J for Res, 1982, 12(1): 18-28

[57] Spies TA, Franklin JF, Thomas TB. Coarse woody debris in Douglas-fir forests of western Oregon and Washington. Ecology, 1988, 69(6): 1689-1702

[58] Stephens SL, Moghaddas JJ. Fuel treatment effects on snags and coarse woody debris in a Sierra Nevada mixed conifer forest. For Ecol Manag, 2005, 214: 53-64

[59] Stephens SL, McIver JD, Boerner RE, Fettig CJ, Fontaine JB, Hartsough BR, Kennedy PL, Schwilk DW. The effects of forest fuel-reduction treatments in the United States. Bioscience, 2012, 62: 549-560

[60] Stitt JM, Hudak AT, Silva CA, Vierling LA, Vierling KT. Evaluating the use of lidar to discern snag characteristics important for wildlife. Remote Sens, 2022, 14(3): 720

[61] Sullivan AL, Surawski NC, Crawford D, Hurley RJ, Volkova L, Weston CJ, Meyer CP. Effect of woody debris on the rate of spread of surface fires in forest fuels in a combustion wind tunnel. For Ecol Manag, 2018, 424: 236-245

[62] Tang XL, Zhou GY, Zhou X, Wen DZ, Zhang QM, Yin GC. Coarse woody debris in monsoon evergreen broad-leaved forests of Dinghushan Nature Reserve. Acta Phytoecol Sin, 2003, 27(4): 484-489

[63] Thiel C, Mueller MM, Epple L, Thau C, Hese S, Voltersen M, Henkel A. UAS imagery-based mapping of coarse wood debris in a natural deciduous forest in central Germany (Hainich National Park). Remote Sens, 2020, 12(20 3293

[64] Waddell KL. Sampling coarse woody debris for multiple attributes in extensive resource inventories. Ecol Indic, 2002, 1(3): 139-153

[65] Wang XQ, Wang MM, Wang SQ, Wu DY. Extraction of vegetation information from visible unmanned aerial vehicle images. Trans Chin Soc Agric Eng, 2015, 31(5): 152-159

[66] Wang N, Ding B, Zhang R, Chen H, Xie T, Bai S, Chen H, Pan X. The six-year decomposition of coarse woody debris drives shifts in soil fungal communities in subtropical forests. Front Microbiol, 2025, 16 1544163

[67] Weedon JT, Cornwell WK, Cornelissen JHC, Zanne AE, Wirth C, Coomes DA. Global meta-analysis of wood decomposition rates: A role for trait variation among tree species?. Ecol Lett, 2009, 12(1): 45-56

[68] Weiss M, Baret F, Smith GJ, Jonckheere I, Coppin P. Review of methods for in situ leaf area index (LAI) determination: Part II. Estimation of LAI, errors and sampling. Agric for Meteorol, 2004, 121(1–2): 37-53

[69] Wijas BJ, Allison SD, Austin AT, Cornwell WK, Cornelissen JHC, Eggleton P, Fraver S, Ooi MKJ, Powell JR, Woodall CW, Zanne AE. The role of deadwood in the carbon cycle: implications for models, forest management, and future climates. Annu Rev Ecol Evol Syst, 2024, 55: 133-155

[70] Woodall CW, Evans DM, Fraver S, Green MB, Lutz DA, D'Amato AW. Real-time monitoring of deadwood moisture in forests: lessons learned from an intensive case study. Can J for Res, 2020, 50(11): 1244-1252

[71] Wu DH. Mechanisms of community change impact on the fine wood decomposition in subtropical forests, 2021, Hangzhou, Zhejiang University

[72] Wu Y, Liu L, Guo C, Zhang Z, Hu G, Ni J. Low carbon storage of woody debris in a karst forest in southwestern China. Acta Geochim, 2019, 38: 576-586

[73] Yan X, Li J, Smith AR, Yang D, Ma T, Su Y, Shao J. Evaluation of machine learning methods and multi-source remote sensing data combinations to construct forest above-ground biomass models. Int J Digit Earth, 2023, 16(2): 4471-4491

[74] Yang L, Chen YH, Yue DP, Feng ZK. Information extraction of urban green space based on UAV remote sensing image. Sci Surv Mapp, 2017, 42(2): 59-64

[75] Ye ZX, Guo Q, Zhang J, Zhang HX, Deng H. Extraction of urban impervious surface based on the visible images of UAV and OBIA-RF algorithm. Trans Chin Soc Agric Eng, 2022, 38(4): 225-234

[76] Zhang YC, Zhang HQ, Chen YF, Li YL, Ma LY. Study of tree competition index based on crown feature. For Res, 2016, 29(1): 80-84

[77] Zhang JL, Guan ZJ, You WB, Hu HL, He DJ. Advances in the impact of forest woody debris on species diversity. Acta Ecol Sin, 2025, 45(7): 3555-3567

[78] Zhao J, Yang HB, Lan YB, Lu LQ, Jia P, Li ZM. Extraction method of summer corn vegetation coverage based on visible light image of unmanned aerial vehicle. Trans Chin Soc Agric Mach, 2019, 50(5): 232-240

[79] Zhu NN, Yang BS, Dong Z. A novel framework for forest vegetation carbon stock estimation based on remote sensing. Natl Remote Sens Bull, 2025, 29(1): 134-146

[80] Zhu M, You WB, Ni SJ. Deadwood-driven improvements in soil health are amplified by forest gaps and elevation. CATENA, 2026, 272 110349

Funding

National Natural Science Foundation of China(32271872)

Rights & permissions

Northeast Forestry University