发表机构
Tsinghua Shenzhen International Graduate School, Tsinghua University; National Supercomputing Center in Shenzhen; School of Artificial Intelligence, Sun Yat-Sen University（清华大学深圳国际研究生院; 深圳国家超级计算中心; 中山大学人工智能学院）
机构由 AI 辅助整理，请以论文原文为准。AI 中文总结
提出GeoPrior-Mamba，利用语言模型生成结构化过程先验并注入多方向Mamba网络，实现高分辨率XCO2重建，显著降低RMSE并加速收敛。
AI 中文摘要
从稀疏卫星观测中重建高分辨率柱平均干空气CO2（XCO2）场，要求模型推断仅受直接测量弱约束的空间结构。现有基于学习的方法通常将环境协变量视为普通数值输入，因此必须在很大程度上依赖稀疏监督来学习异质的源-汇关系。我们提出了GeoPrior-Mamba，一种多方向Mamba框架，并辅以离线语言模型诱导的结构化过程先验。我们并非使用语言模型来预测XCO2，而是在训练前利用其组织生物圈吸收、生态系统呼吸和人为排放的相对过程知识，形成确定性先验表。这些先验通过地理、生态、排放相关和季节信息在空间上实例化，并通过轻量级知识适配器自适应地注入重建主干网络。使用2018-2020年的OCO-2观测数据，GeoPrior-Mamba在留出观测上实现了0.81 ppm的RMSE和0.93的R2，相对于CAMS背景插值，RMSE降低了48.2%；在相同评估协议下，相对于Trans-XCO2，RMSE降低了3.1%。消融实验表明，知识先验分支具有可量化的贡献，且收敛速度显著快于无知识的Mamba主干。独立的TCCON评估进一步支持重建场与地基柱CO2测量的一致性。这些结果表明，当难以直接获得全局一致的过程-响应表示时，语言模型可以提供一种构建结构化过程先验的实用机制，同时保持在数值预测循环之外。
英文摘要
Reconstructing fine-resolution column-averaged dry-air CO2 (XCO2) fields from sparse satellite observations requires models to infer spatial structure that is only weakly constrained by direct measurements. Existing learning-based methods typically treat environmental covariates as ordinary numerical inputs and must therefore learn heterogeneous source-sink relationships largely from sparse supervision. We introduce GeoPrior-Mamba, a multi-directional Mamba framework augmented with offline language-model-induced structured process priors. Rather than using a language model to predict XCO2, we use it before training to organize relative process knowledge for biospheric uptake, ecosystem respiration, and anthropogenic emissions into deterministic prior tables. These priors are spatially instantiated using geographic, ecological, emission-related, and seasonal information and are adaptively injected into the reconstruction backbone through a lightweight knowledge adapter. Using OCO-2 observations from 2018-2020, GeoPrior-Mamba achieves an RMSE of 0.81 ppm and an R2 of 0.93 on held-out observations, reducing RMSE by 48.2% relative to CAMS background interpolation and by 3.1% relative to Trans-XCO2 under the same evaluation protocol. Ablation experiments show a measurable contribution from the knowledge-prior branch and substantially faster convergence than the knowledge-free Mamba backbone. Independent TCCON evaluation further supports the consistency of the reconstructed fields with ground-based column CO2 measurements. These results suggest that language models can provide a practical mechanism for constructing structured process priors when globally consistent process-response representations are difficult to obtain directly, while remaining outside the numerical prediction loop.
