Abstract: In water-limited ecosystems, photodegradation is a dominant pathway of carbon (C) turnover. However, the combined effects of spectral composition and litter traits in hyper-arid deserts remain poorly constrained, limiting the accuracy of C flux predictions. In a 637-day field experiment with three representative desert species (Populus euphratica, Alhagi sparsifolia, and Karelinia caspia), we applied natural light filters to establish six spectral treatments. Full-spectrum exposure increased litter decomposition by 70%; of the five individual wave-bands, only UV-B (280–315 nm) and blue light (400–500 nm) significantly accelerated mass loss, accounting for 43% and 29% of the full-spectrum effect, respectively. These two wave-bands accelerated C and cellulose loss, whereas green light (500–580 nm) selectively promoted hemicellulose and lignin degradation without affecting total mass. UV-B and blue light also increased specific leaf area (SLA) by 12.9% and 7.0% and elevated litter microbial respiration rate (LMR, 24-h incubation at 25 °C, 60% relative humidity) by 54.6% and 40.6%, respectively. Across four spectral regions, initial C content, lignin: N ratio, SLA, and LMR were significantly correlated with photodegradation rate, with LMR and lignin content per unit surface area being the strongest predictors of susceptibility. These results highlight the co-regulation of photodegradation by spectral composition and plant traits, advancing mechanistic understanding of C cycling in hyper-arid ecosystems and providing refined parameters for Earth system models changing solar regimes.

Key words: Photodegradation, Litter decomposition, Spectral regions, Plant traits, Hyper-arid desert, Carbon cycle