This repository contains MATLAB scripts used for analyzing artificial light at night (ALAN) across the Lake Washington and Lake Sammamish basins. The scripts process in-situ irradiance data and satellite-based VIIRS Day/Night Band (DNB) radiance to explore spatial and temporal patterns in nighttime light.

Jennifer A. Schulien
jennifer@schulcon.com
Schulien Consulting

Processes hyperspectral irradiance profiles collected using a Compact Optical Profiling System (C-OPS). Includes quality control, dark correction, smoothing, interpolation, and calculation of diffuse attenuation coefficients (Kd).

• Multiple *URC.csv files exported from the Biospherical Instruments software

• Smoothed, interpolated irradiance (Ed) profiles by depth
• Wavelength-specific Kd profiles per station
• Quality control plots for pitch/roll and irradiance consistency

• Filters out dark/reference casts
• Removes scans with pitch or roll > ±5°
• Computes dark noise correction
• Applies moving average smoothing
• Interpolates irradiance onto a common depth grid (0.1–40 m)
• Calculates and filters vertical diffuse attenuation coefficients (Kd)

• MATLAB R2022a or later
• CSV reading capability

Loads and subsets monthly VIIRS DNB radiance and cloud-free coverage raster files from 2014–2023. Outputs gridded arrays for time-series analysis of nighttime light in the Pacific Northwest.

• Directory structure containing monthly subfolders (e.g., 202001/, 202002/, ...)
  • *.avg_rade9h.tif — Monthly average radiance (nW·cm⁻²·sr⁻¹)
  • *.cf_cvg.tif — Cloud-free pixel coverage

• lat_cell, lon_cell — Latitude and longitude vectors per month
• rad_cell — Monthly radiance arrays
• cf_cvg_cell — Monthly cloud-free pixel masks

• Longitude: −125° to −116.45°
• Latitude: 45.5° to 50°
• Includes Seattle, Puget Sound, Lake Washington, and surrounding watersheds

• Iterative batch loading of .tif files
• Raster subsetting to defined ROI
• Storage of radiance and coverage arrays in cell format

• MATLAB R2022a or later
• Built-in Tiff class

Performs statistical trend analysis of nighttime light intensity across 36 ecological ROIs using VIIRS DNB composites. Fits GLMs and other models to summer/fall data (June–September).

• viirs_subset_2014_2023_sl.mat — Monthly subsetted VIIRS DNB data
• SVDNB_202208_75N180W.mat — August 2022 cloud-free pixel mask
• SVDNB_202208_75N180W_rad.mat — August 2022 DNB radiance

• sum_rad_out — Summed radiance per ROI/month
• p_vec, s_vec, t_vec, r2_vec — GLM stats
• p_vec_robust, s_vec_robust — Robust regression stats
• p_vec_mk — Mann–Kendall p-values
• ROI-specific trend plots (figures)

Covers 36 custom regions:

• Pelagic zones (e.g., Northern, Middle, Southwest Lake Washington)
• Infrastructure (e.g., I-90, SR-520, Boeing-Renton, Ship Canal)
• Urban nearshore (e.g., Kirkland, Mercer Island)
• Natural/tributary zones (e.g., Sammamish Slough)

1. Load monthly VIIRS radiance and cloud coverage
2. Apply polygon masks to extract ROI-specific values
3. Aggregate radiance for summer/fall months
4. Fit trends using:
   • Generalized Linear Models (GLMs)
   • Robust regression
   • Theil–Sen estimator & Mann–Kendall test
5. Plot GLM trends for significant ROIs (p ≤ 0.05)

• MATLAB R2022a or later
• m_map toolbox (for map projections)
• Mann_Kendall.m and TheilSen.m functions

This MATLAB script compares nighttime in-situ irradiance (C-OPS Ed₀) to VIIRS Day/Night Band (DNB) radiance at coincident spatial locations on Lake Washington. It integrates each C-OPS wavelength channel against the VIIRS spectral response to derive a spectrally weighted in-situ irradiance, then compares this with radiance extracted from VIIRS pixels at each C-OPS station. This script produces Figure 4b–c in the final publication.

• VIIRS nightly radiance (subset .mat)
• C-OPS nighttime irradiance (COPS_night_irr_*.mat)
• C-OPS sensor response curve (SpectralResponse_Ed0_SN*.csv)
• VIIRS spectral response (suomi_NPP_onorbit_datathief.csv)
• Lake Washington shapefile (Waterbodies_with_History_and_Jurisdictional_detail_*.shp)

• Matched and weighted radiance–irradiance scatter plot
• Xvec_norm: Spectrally weighted in-situ Ed₀
• mat_out: Table of VIIRS/C-OPS matched coordinates and values
• ROI-based vectors for Figure 4 panels b–c

1. Load VIIRS/DNB radiance for the target night (viirs_rad_*.mat)
2. Load in-situ C-OPS nighttime Ed₀ data (COPS_night_irr_*.mat)
3. Load and interpolate spectral response curves:
   • C-OPS reference sensor (Ed₀)
   • VIIRS Day/Night Band (DNB)
4. Spectrally weight in-situ Ed₀ data using VIIRS/DNB response
5. Subset VIIRS radiance to Lake Washington using polygon shapefile
6. Extract VIIRS radiance at each in-situ cast location (nearest bin)
7. Assign in-situ stations to predefined ROIs (e.g., Pelagic, Nearshore)
8. Plot spectrally weighted Ed₀ vs. L-VIIRS for matched sites

• MATLAB R2022a or later
• m_map toolbox (for map projections)
• NASA VIIRS/DNB .mat product (from import_nightly_viirs.m)

All scripts are focused on the greater Lake Washington and Lake Sammamish watersheds, including urban and pelagic zones, bridges, tributaries, and natural buffers.