Skip to content

Author

D. Jakimov

1 paper indexed here

We haven’t gathered this author’s papers yet. Follow them and we’ll fetch their work.

Not the right person? Other researchers publish under this name.

Open access Aug 2026

Long-Term Multivariate Screening of a Recirculating Landfill Leachate Circuit: Pollutant Dynamics, Statistical Structure and Associated Risk to Biota

Landfill leachate circuits that operate without discharge, by recirculating aerated leachate onto the waste mass, are widespread in South-East Europe, yet their long-term behaviour is rarely documented with sample-level data. This study reports a six-year (2020–2025) seasonal monitoring campaign at a sanitary landfill in northern Serbia (alluvial aquifer of the Sava River, transboundary Danube basin) and re-examines it with a transparent multivariate protocol. Seventy-two leachate samples (collection well, aeration lagoon, sedimentation lagoon; n = 24 each, 30 parameters), 28 realised surface-water campaigns, and six years of groundwater summaries were evaluated by principal component analysis/factor analysis (PCA/FA, Varimax normalized), hierarchical cluster analysis, PERMANOVA, non-parametric paired tests and, for benchmarking, supervised machine learning. The pooled leachate model (n = 72; 21 variables; KMO = 0.700; Bartlett χ2 = 956, p < 0.001) retained four factors by parallel analysis, explaining 61.6% of total variance; after rotation the factors accounted for 27.7%, 14.3%, 10.4%, and 9.3%. Factor 1 grouped organic load with particle-reactive metals (COD, BOD5, Fe, Ni, Cr, As, Zn), Factor 2 a reduced sulfur–fluoride–BTEX signature, Factor 3 temperature-driven nitritation, and Factor 4 a nitrate–manganese redox contrast. Crucially, paired campaign-by-campaign comparison showed no removal of the dominant pollutants along the circuit. Median COD, BOD5 and NH4-N were not lower in the sedimentation lagoon than in the collection well, while pH rose from 8.08 to 8.75 (p < 0.001); only Cu, Pb, NO3-N, and NO2-N decreased significantly. The circuit therefore homogenises and concentrates dissolved load rather than removing it. Downstream surface water was significantly enriched in electrical conductivity (+110 µS/cm), total dissolved solids, NH4-N, and NO2-N relative to upstream (Wilcoxon, p < 0.05), and groundwater showed episodic conductivity up to 12,760 µS/cm and NH4-N up to 102 mg/L. Cytotoxicity (MTT) confirmed biological relevance, with MRC-5 viability falling to 37% after 24 h exposure to 50 vol.% groundwater (Pw3) versus 60% in A549 cells. A random-forest classifier separated circuit units far better than PCA-based discrimination (76.4% versus 54.2% cross-validated accuracy) and distinguished the 2020–2021 pandemic period from 2022–2025 with 94.2% accuracy, a period effect also confirmed by PERMANOVA (R2 = 7.2%, p < 0.001). The results indicate that closed-loop recirculation without an engineered discharge barrier transfers, rather than eliminates, contaminant load, and that after-care of such systems requires mass-balance monitoring and polishing treatment.

N. Grba, V. Mihajlović, Goran Benedeković et al. · 0 citations