Summary This dataset contains 10,000 samples of synthetic bio-optical data generated specifically for the optical characteristics of Danjiangkou Reservoir (China). It was created using the Danjiangkou Bio-Optical Model (DBOM) forward simulation framework. The dataset links biogeochemical concentrations (Chlorophyll-a, Inorganic Suspended Matter, CDOM) and phytoplankton community structure (Cyanobacteria, Green algae, Brown algae) to Inherent Optical Properties (IOPs) and hyperspectral Remote Sensing Reflectance ($R_{rs}$). The dataset contains the stochastic perturbation of bio-optical parametrization coefficients (e.g., spectral slopes, specific absorption coefficients), which were randomized by $\pm 30\%$ around their local default values to simulate natural variability and support algorithm uncertainty analysis. Data Generation Methodology Concentration Generation: Input concentrations were generated using a multivariate log-normal distribution to mimic natural covariance. Phytoplankton community fractions were generated using a Dirichlet distribution. Forward Modeling: IOP Model: Localized bio-optical models for Danjiangkou Reservoir were used to compute absorption ($a$) and backscattering ($b_b$) coefficients. Parameter Perturbation: Key bio-optical parameters (including $S_{cdom}$, $S_{nap}$, $a^*_{ph}$, etc.) were independently perturbed using a uniform distribution $U(0.7, 1.3)$ for each sample to introduce realistic optical complexity. Radiative Transfer: Remote sensing reflectance ($R_{rs}$) was derived from IOPs using the semi-analytical model by Lee et al. (2011). Spectral Resolution: Hyperspectral data covering 350 nm to 900 nm with a 2 nm interval. File Structure The data is stored in a single NetCDF file (.nc) compatible with Xarray and other multidimensional array libraries. Dimensions: sample_id: 10,000 (Number of samples) wavelen: 276 (Spectral bands from 350 to 900 nm) Data Variables (Outputs): Rrs: Remote Sensing Reflectance [$sr^{-1}$] aph: Absorption coefficient of phytoplankton [$m^{-1}$] ad: Absorption coefficient of detritus/non-algal particles [$m^{-1}$] acdom: Absorption coefficient of CDOM [$m^{-1}$] bbp: Particulate backscattering coefficient [$m^{-1}$] bp: Particulate total scattering coefficient [$m^{-1}$] Coordinates (Inputs & Truth Data): Chla: Chlorophyll-a concentration [$mg/m^3$] ISM: Inorganic Suspended Matter concentration [$g/m^3$] ag440: CDOM absorption at 440 nm [$m^{-1}$] frac_cyano, frac_green, frac_brown: Fractional composition of phytoplankton groups (0-1). Perturbed Parameters: The file also records the specific randomized parameters used for each sample (e.g., S_cdom, A_ph, gamma_nap) to allow for sensitivity analysis. Potential Applications Development and validation of semi-analytical algorithms for inland waters. Training machine learning models for water quality retrieval. Phytoplankton functional type (PFT) discrimination studies. Sensitivity analysis of bio-optical inversion algorithms. Technical Info Format: NetCDF4 Compression: zlib (level 9) Code Author: Shun Bi (bishun@niglas.ac.cn)