Wildfires are increasing in frequency, extent, and severity globally due to climate change. These fires alter watershed characteristics, reducing vegetation cover and changing soil properties, leading to changes in surface runoff (increased low and high flows, extreme floods). Post-fire runoff mobilizes burned terrestrial material, creating aquatic disturbances that cascade through fluvial networks, impacting environmental conditions and ecosystem services. The magnitude, extent, and duration of these aquatic disturbances depend on burn severity, watershed characteristics, precipitation patterns, and ecosystem recovery rates. Altered water quality is a primary post-fire disturbance, with increased runoff transporting sediment, nutrients, and metals into streams. While most water quality disturbances decline within 5 years, some effects persist for decades. Current fire models predict continued increases in wildfire size and damage, necessitating quantification and prediction of disturbance propagation. This study addresses this knowledge gap by deploying water quality sensors along watersheds affected by the largest wildfire in New Mexico's history (Hermit's Peak-Calf Canyon Fire, burning 1382 km²). Data collection began before rainfall mobilized burned material and covered the first post-fire rainy season. The study quantifies the fire's impacts on terrestrial vegetation and surface runoff, the longitudinal propagation of rainfall-runoff events and their impacts on water quality, and the sub-year impacts of wildfire disturbances on water quality dynamics and ecosystem processes along fluvial networks.

Existing research highlights the significant impacts of wildfires on water quality in streams draining burned watersheds. Studies have documented increased sediment export, changes in water chemistry (increased ions, nutrients, metals, and episodic declines in dissolved oxygen), and the persistence of these effects for years, even decades. However, the longitudinal propagation of these disturbances throughout fluvial networks is poorly understood. Limited data suggests impacts extend hundreds of kilometers downstream, affecting reservoirs and water supplies. The coupling between post-fire rainfall-runoff events and the propagation of aquatic disturbances is not well-established. This lack of understanding necessitates further research into the network-scale propagation of these disturbances.

The study focused on the Gallinas and Sapello watersheds in New Mexico, affected by the Hermit's Peak-Calf Canyon (HPCC) wildfire. Approximately 87% of the Gallinas watershed and 47% of the Sapello watershed above the monitoring sites were burned. The study used a combination of data: 1. **Discharge Data:** Discharge and stage data were collected at 5- and 15-minute intervals at stream gages maintained by the USGS. Discharge was estimated at some sites using lagged relationships between gages or historical comparisons. 2. **Meteorological Data:** Rainfall data were collected at 5-minute intervals at USGS gages. Air pressure and shortwave radiation were obtained from NASA's NLDAS, and meteorological data from MesoWest stations. 3. **Water Quality Data:** Water quality sondes were deployed to collect continuous data on parameters such as dissolved oxygen (DO), fluorescent dissolved organic matter (fDOM), pH, specific conductivity (SpCond), and turbidity. Data quality control involved outlier removal, sensor drift correction, and visual inspection. fDOM was corrected for temperature and turbidity. Photosynthetically active radiation (PAR) was derived from shortwave radiation. Stream depth and velocity were estimated using power-law relationships. 4. **Vegetation Index:** Landsat enhanced vegetation index (EVI) data (2007–2022) from MODIS Terra and Aqua were used to assess the impact of the fire on vegetation. The 15-year pre-fire mean and standard deviation were compared with 2022 pre- and post-fire values. 5. **HEC-HMS Model:** A Hydrologic Engineering Center-Hydrologic Modeling Software (HEC-HMS) model was used to simulate surface runoff flows, comparing pre- and post-fire conditions using hydrologic impact factors based on burn severity. 6. **Historical Water Quality Data:** Historical discrete water quality data (suspended sediment discharge (SSD), turbidity, and nitrate) at P226 were compared with post-fire data using the Mann-Whitney U test. 7. **Stream Metabolism:** Daily averages of stream metabolism were estimated with the USGS *streamMetabolizer* model. Stream metabolic fingerprints were used to compare metabolism patterns. 8. **Principal Component Analysis (PCA):** PCA was used to analyze relationships between water quality parameters and stream metabolism. 9. **Wavelet Coherence Analysis:** Wavelet coherence was used to analyze the propagation of water quality signals along the fluvial networks.

The HPCC wildfire severely decreased plant biomass, as indicated by reduced EVI values. Post-fire, even average rainfall events generated significantly larger runoff events compared to pre-fire conditions, as modeled by HEC-HMS. The model with post-fire parameters better predicted the observed runoff, indicating a drastic change in watershed hydrologic response. Runoff volume decreased downstream, with the median volume decreasing from 312,000 m³ at P226 to 47,200 m³ at P5167. Significant differences were found between historical pre-fire and post-fire data for SSD, turbidity, and nitrate at P226, near the burn perimeter. PCA of daily water quality data showed clear separation between pre- and post-connectivity periods, indicating consistent impacts to water quality parameters and ecosystem processes at downstream locations. Diel DO variation decreased post-fire, and stream metabolism (GPP and ER) also shifted. Wavelet coherence analysis revealed propagation of pH, temperature, and other water quality signals along the fluvial network, although the strength of coherence decreased with distance downstream in some cases. However, propagation was detectable to the furthest downstream sites. The study also found that the simple model of stream length impacted (SLLE) from Ball et al. (2021) provides a reliable estimate of the actual longitudinal extent of the impact.

The results demonstrate the profound and far-reaching impacts of the HPCC wildfire on both the terrestrial and aquatic systems. The increased runoff generation and mobilization of sediment and nutrients highlight the significant alterations to watershed hydrology and water quality. The longitudinal propagation of disturbances, even over long distances and across multiple months, emphasizes the interconnectedness of atmospheric, terrestrial, and aquatic systems. The attenuation of runoff and water quality signals downstream reflects the influence of losing stream conditions and the presence of large reservoirs. While the model SLLE presented in Ball et al. (2021) performed well in this context, adjustments may be necessary for arid watersheds with losing conditions. This study improves our understanding of wildfire disturbance propagation and emphasizes the need for spatially and temporally comprehensive monitoring strategies.

This study provides crucial insights into the longitudinal propagation of wildfire disturbances in fluvial networks. The HPCC wildfire caused significant changes to watershed hydrology and water quality, with detectable impacts extending >160 km downstream. The findings emphasize the need for spatially resolved longitudinal sampling designs and highlight the interconnectedness of atmospheric, terrestrial, and aquatic processes. Future research should focus on improving predictive models of wildfire impacts and developing effective mitigation strategies.

The study had limited pre-fire water quality data for most locations. The HEC-HMS model, while useful, is a simplified representation of complex hydrologic processes and may not capture all nuances of runoff generation and flow dynamics. The temporal coverage of the study (first post-fire rainy season) limits the assessment of long-term recovery. The study focused on two watersheds, and findings may not be generalizable to all contexts.