🍇 Vineyard Wildfire Impacts and Exposure: The 2019 Kincade Fire

Parastoo Farajpoor, Mahla Ardebili Pour, Mohammad Bagher Ghiasi, and Mohammadreza Narimani* (corresponding author)

Frontiers in Sustainability · Volume 7 · Article 1997548 · 27 October 2026 · Open access CC BY

Department of Biological and Agricultural Engineering · University of California, Davis

Read the open-access paper Cite on Frontiers

Why Vineyard Wildfire Impacts Need Their Own Framework

Wildfire risk in California is often framed through the wildland–urban interface. That framing is essential, but it leaves working agricultural landscapes—especially North Coast vineyards—as background. Winegrape systems can experience direct canopy impact, smoke exposure, disrupted road access, and delayed greenness recovery even when their spectral burn response looks milder than neighboring wildland.

For people searching Kincade Fire vineyard, Sonoma County wildfire agriculture, vineyard smoke exposure, or Sentinel-2 burn severity winegrape, this page summarizes an open-access Frontiers study led with corresponding author Mohammadreza Narimani at the University of California, Davis. Coauthors are Parastoo Farajpoor, Mahla Ardebili Pour, and Mohammad Bagher Ghiasi.

The electrically initiated 2019 Kincade Fire burned about 77,758 acres, destroyed 374 structures, and prompted Sonoma County’s largest evacuation. The paper analyzes 4,581 vineyard fields (8,813.2 ha) with Sentinel-2, OpenET, gridMET fire weather, soils, terrain, NOAA/NESDIS smoke polygons, an ignition-date OpenStreetMap road network, and surveyed structures.

Core story: lower vineyard dNBR than wildland does not prove a universal firebreak effect. Smoke, accessibility, and recovery must be read alongside spectral severity—and readers should cite the Frontiers article, not only this summary.
4,581Vineyard fields
0.130 / 0.337Vineyard vs wildland dNBR
7.78Mean smoke-days / field
34.2%Dead-end road nodes
0.815Inside recovery ratio (2021)
CC BYOpen-access license
Prefer the publisher version for reading and citing

Event-Anchored Geospatial Evidence Pathway

The study organizes evidence in four stages: hazard drivers (fire extent, spectral response, weather, electrical context, terrain, soils), exposed systems (vineyards, structures, roads, overhead smoke), multisystem analyses (pre-fire water status, boundary contrasts, accessibility, smoke, recovery), and a qualitative synthesis for planning attention. This template is designed so agricultural wildfire assessments do not collapse to a single dNBR map.

Mohammadreza Narimani UC Davis multidisciplinary workflow for Kincade Fire vineyard wildfire impacts from hazard drivers to evidence synthesis
Figure 1. Four-stage evidence pathway from hazard drivers through exposed systems to synthesis.
Mohammadreza Narimani Kincade Fire Sonoma County study area with vineyard fields fire perimeter and CAL FIRE destroyed-structure inspections
Figure 2. Kincade Fire study area: vineyard inventory, perimeter, and destroyed-structure inspections.
Mohammadreza Narimani Sentinel-2 true-color and shortwave-infrared chronology of the 2019 Kincade Fire for vineyard wildfire remote sensing
Figure 3. Sentinel-2 event chronology in true color and SWIR false color.

Pre-Fire Vineyard Canopy, Water Use, and Fire Weather

Field-mean pre-fire NDMI and NDVI were related to April–October OpenET evapotranspiration across the vineyard inventory (NDMI–ET Pearson r ≈ 0.567). OpenET is interpreted as seasonal crop water use—not metered irrigation. These baselines matter because autumn phenology and water status can confound raw pre–post spectral change.

gridMET humidity, vapor-pressure deficit, and wind place the event in a severe fire-weather window, and the documented electrical-transmission ignition timing frames Kincade as infrastructure and weather acting inside an agricultural–community system. Lidar, USGS 3DEP, and GEDI inventories show where three-dimensional structure data exist to support future fuel and canopy analyses.

Mohammadreza Narimani scatterplots of pre-fire vineyard NDMI NDVI versus OpenET seasonal evapotranspiration for Kincade Fire winegrape fields
Figure 4. Pre-fire canopy moisture/greenness versus seasonal OpenET water use.
Mohammadreza Narimani maps of airborne lidar USGS 3DEP and GEDI coverage around the Kincade Fire vineyard landscape
Figure 5. Three-dimensional structure data availability across the study landscape.
Mohammadreza Narimani Kincade Fire weather time series of humidity vapor-pressure deficit and wind with electrical ignition timing
Figure 6. Fire-weather context and documented electrical-ignition timing.

Sentinel-2 dNBR Contrast Is Not a Universal Firebreak Proof

Inside the perimeter, vineyard pixels showed substantially lower descriptive dNBR than wildland (means 0.130 vs 0.337). That landscape contrast is real—and easy to over-interpret. A segment-clustered vineyard–wildland boundary contrast at 100 m was small (τ = −0.0166) and fragile across bandwidths, slope continuity, donut specifications, and placebo checks.

A 250 m spatial generalized additive model, after conditioning on location, terrain, and water use, showed a negative vineyard-fraction association (joint-smooth coefficient −0.111) but left strong residual spatial autocorrelation (Moran’s I = 0.611). Lower spectral change therefore cannot be marketed as a proven, universal vineyard firebreak without those diagnostics.

Mohammadreza Narimani scale-dependent vineyard-wildland dNBR boundary contrast analysis for the 2019 Kincade Fire
Figure 7. Scale-dependent vineyard–wildland boundary contrast in dNBR.
Mohammadreza Narimani vineyard dNBR versus OpenET and vineyard versus wildland spectral severity distributions for Kincade Fire
Figure 8. Associations and confounding in the spectral severity response.
Read the full paper: boundary tests, GAM specifications, placebo checks, and residual diagnostics are in Frontiers. Cite DOI 10.3389/frsus.2026.1997548.

Vineyard Smoke Exposure, Road Access, and Post-Fire Recovery

All mapped vineyards intersected NOAA/NESDIS overhead smoke on at least one day (mean 7.78 potential smoke-days per field). Potential smoke exposure is not a PM2.5 dose or smoke-taint diagnosis, but it is a planning-relevant exposure indicator that spectral severity alone misses.

On the ignition-date road network, about 34.2% of nodes were dead ends, highlighting access fragility during evacuation and response. Through 2021, fields inside the perimeter showed a larger greenness deficit than fields outside (recovery ratios 0.815 inside vs 0.854 outside). Lower immediate spectral change therefore did not establish vine survival or recovered agricultural function.

Mohammadreza Narimani vineyard road-access distributions and OpenStreetMap network topology for the Kincade Fire study
Figure 9. Vineyard road access and ignition-date network topology.
Mohammadreza Narimani NOAA smoke exposure and post-fire vineyard NDVI recovery trajectories inside versus outside the Kincade perimeter
Figure 10. Potential smoke exposure and post-fire vineyard greenness recovery.
Mohammadreza Narimani summary of complementary vineyard wildfire indicators for severity smoke access and recovery after Kincade Fire
Figure 11. Complementary vineyard wildfire indicators for planning attention.

Spectral severity

Sentinel-2 dNBR and boundary tests describe immediate optical impact—not vine mortality by themselves.

Exposure & access

Smoke-day counts and dead-end network topology capture disruption pathways beyond burn maps.

Recovery evidence

Multi-year NDVI trajectories separate short-term spectral change from longer greenness deficits.

Paper and Citation

Please read and cite the publisher version when you use this work. This research card is a visual summary only—it is not a substitute for the article, figures, or methods in Frontiers.

Farajpoor, P., Pour, M. A., Ghiasi, M. B., & Narimani, M. (2026). Multisource remote sensing and geospatial analysis of vineyard wildfire impacts and exposure: the 2019 Kincade Fire. Frontiers in Sustainability, 7, 1997548.

DOI: 10.3389/frsus.2026.1997548 · Open access under CC BY 4.0

Read & cite the open-access paper Open on Frontiers

Contact

Mohammadreza Narimani
PhD Candidate, University of California, Davis

📧 mnarimani@ucdavis.edu  |  🎓 Google Scholar