Global amphibian populations are declining rapidly and the development of effective cryopreservation protocols for germ cells has become a critical tool in ex situ conservation programs. Post-thaw membrane-integrity viability in four endangered Korean amphibians (Dryophytes suweonensis, Pelophylax chosenicus, Kaloula borealis, and Hynobius yangi) were evaluated. Sperm were cryopreserved using dimethyl sulfoxide (DMSO) or N,N-dimethylformamide (DMF) at 10–30% (v/v) in combination with 0.6 M sucrose, and were frozen at two suspension heights (5 cm vs. 10 cm) above liquid nitrogen. Post-thaw membrane-integrity viability was assessed using a SYBR-14/propidium iodide membrane-integrity assay (LIVE/DEAD kit). Low concentrations of permeating cryoprotectants (CPs) improved membrane-integrity viability, whereas high concentrations led to high toxicity, particularly with DMSO. Across species, DMF produced the highest membrane-integrity viability and the most consistent performance. The cooling rate influenced membrane-integrity viability, with samples frozen at 10 cm exhibiting greater viability, reflecting the balance between intracellular ice formation during rapid cooling and solution effects during slow cooling. Optimal conditions for D. suweonensis were 15% DMSO at 10 cm (86.5% membrane-integrity viability); for P. chosenicus, 10% DMF at 10 cm (75.5%); and for K. borealis, 10% DMSO at 5 cm (81.6% membrane-integrity viability). Hynobius yangi showed modest improvement under 15% DMF at 5 cm (19.7%), although overall membrane-integrity viability was low. ED₅₀ modeling indicated species-specific thresholds requiring low CP concentrations. Sperm cryopreservation outcomes in amphibians are strongly influenced by CP type, concentration, cooling regime, and species physiology. GLM and ED₅₀ modeling provide a methodological framework for refining cryopreservation strategies for non-model, endangered species.

Species distribution models often assume that species occupy all environmentally suitable space, but low‐dispersal amphibians may track landscape structure that has persisted over long periods as well as current local conditions. We examined how broad‐scale landscape structure and fine‐scale microhabitat jointly shape the distribution of the Jeju salamander, Hynobius quelpaertensis , on Jeju Island, Republic of Korea. Using 85 spatially thinned occurrence records and 10,000 background points sampled on land, we compared a broad‐scale predictor set (terrain at ≥ 1 km, drainage position and long‐term climate, used as a proxy for a long‐term topographic‐hydrological template) with a fine‐scale set (microtopography, wetness, insolation and canopy). Weighted elastic‐net models were evaluated by spatial block cross‐validation with block‐bootstrap confidence intervals, and a generalized additive model (GAM) with broad‐scale suitability as an offset tested whether microhabitat refined occupancy within broadly suitable space. The broad‐scale and fine‐scale models discriminated occurrences equally well (ROC AUC = 0.79 and 0.78), but the broad‐scale model was better calibrated (Tjur's R ² = 0.16 vs. 0.12), and the combined model performed best (ROC AUC = 0.81; Tjur's R ² = 0.20). Both sets identified drainage position: occurrence increased with upstream drainage area and local wetness and was higher in hollows and valley bottoms. Within the broad‐scale core, fine‐scale smooth terms doubled Tjur's R ² (0.14 to 0.28): occurrence rose with canopy shade and valley depth and declined on exposed south‐facing slopes, although the gain in discrimination was uncertain. Areas suitable only at the fine scale covered 24% of the island but had one‐sixth the presence rate of areas suitable at both scales. Valley drainage structure thus sets the template for the distribution of H. quelpaertensis , and shaded, moist microhabitat refines occupancy within it. Conservation should maintain drainage networks together with their forest cover, and candidate unoccupied areas should be surveyed in the field.