High photovoltaic (PV) penetration in renewable energy bases located in desert and Gobi regions presents significant challenges. These include source-load mismatches and multi-timescale fluctuations driven by uncertainties, which threaten both reliability and economic performance of the power system. To address these issues, this paper proposes a two-stage robust optimization (TSRO) strategy. To enhance PV utilization and improve operational economics by coordinating long-duration hydrogen energy storage systems (HESs), shortduration electrochemical energy storage systems (EESs), and demand response (DR). Specifically, in the first stage, a polarity-consistent variational mode decomposition (VMD) method is applied to extract low-frequency components from the net load curve, laying the foundation for long-duration HES optimization. Then, in the second stage, EES and DR are co-optimized under generated uncertainty scenarios, which incorporate system operation costs, EES degradation cost, load heterogeneity, and Time-of-Using (TOU) price-driven DR. Furthermore, a column-and-constraint generation (C&CG) algorithm is employed to expand the uncertainty set based on worst-case feedback iteratively. Finally, case studies show that the proposed strategy reduces system operating costs by 21.4 % and PV curtailment by 15.9 %, while mitigating EES degradation-related costs. Overall, the approach significantly enhances both the economic performance and adaptability of the power system under uncertainty.