Data and code from: Western larch regeneration more sensitive to wildfire-related factors than seasonal climate variability
Data files
Sep 16, 2024 version files 2.38 GB
-
GIS_data.zip
2.38 GB
-
README.md
12.08 KB
-
Vieira_et_al_climate_timeseries.csv
230.04 KB
-
Vieira_et_al_plot_data.csv
170.65 KB
Abstract
To understand the impacts of changing climate and wildfire activity on conifer forests, we studied how wildfire and seasonal post-fire climate conditions influence western larch (Larix occidentalis) regeneration across its range in the northwestern US. We destructively sampled 1651 seedlings from 57 sites within 32 fires that burned at moderate or high severity between 2000 and 2015, and which were within 100 m of reproductively mature western larch. Using dendrochronological methods, we estimated germination years of seedlings to calculate annual recruitment rates. We used boosted regression trees to model the annual probability of recruitment as a function of (i) “wildfire-related factors” including distance-to-seed-source, satellite-derived fire severity, and time-since-fire, and (ii) seasonal climate conditions, including variables reflecting temperature and water availability. Most recruitment occurred within five years after wildfire, at sites within 25 m of reproductively mature western larch. Wildfire-related factors had the highest relative influence on post-fire recruitment (87%), whereas post-fire seasonal climate had less influence (13%). Annual recruitment probability increased with growing season actual evapotranspiration, to a maximum c. 275 mm, and then decreased. Annual recruitment probability decreased as growing season climatic water deficit increased. Our results suggest that recent climate trends – increased growing season water deficit and decreased actual evapotranspiration – have had variable, yet net-neutral, impacts on the climate suitability for post-fire western larch regeneration across its range. Climate suitability increased modestly at ‘cooler-and-wetter’ sites and decreasing modestly at ‘warmer-and-drier’ sites. The strong influence of wildfire-related factors highlights the potential for management decisions to promote western larch. For instance, facilitating prescribed or managed wildfire with moderate- to high-severity patches will generate conditions suitable for natural regeneration, provided sufficient seed sources. Additionally, our findings support monitoring for natural regeneration or directing outcomes by planting within the first five years after fire, consistent with current management practices.
README: Impacts of climate and wildfire on western larch regeneration
https://doi.org/10.5061/dryad.0k6djhb6r
Authors: Spencer T. Vieira, Kimberley T. Davis, Zachary A. Holden, Andrew J. Larson, and Philip E. Higuera
Journal: Forest Ecology and Management
FILES' NAMES AND CONTENTS
Vieira_et_al_climate_timeseries.csv
- For each year: summer minimum soil moisture, recruitment season (April to May) mean soil moisture, recruitment season mean temperature, summer maximum surface temperature, growing degree days above 5 degrees C, growing season (May to Sept.) climatic water deficit, and growing season actual evapotranspiration (climatic water deficit and actual evapotranspiration are included at both 30m and 250m pixel resolution).
- 30-yr mean annual climatic water deficit (1991 to 2020)
Vieira_et_al__plot_data.csv
- Annual recruitment and site information for each site.
GIS_Data
[Folder]
- Rehfeldt_laoc_distro_complete.shp
- Total western larch distribution, modeled by Rehfeldt et al. (2010), including Canada.
- citation: Rehfeldt, G. E., & Jaquish, B. C. (2010). Ecological impacts and management strategies for western larch in the face of climate-change. Mitigation and Adaptation Strategies for Global Change, 15(3), 283–306. https://doi.org/10.1007/s11027-010-9217-2
- Rehfeldt_laoc_distro_US_1km_buffer.shp
- Clipped (of above) to northwestern states (WA,OR, ID, MT) including a 1 kiliometer buffer to account for potential seed dispersal.
- laoc_fireperim_19862020.shp
- Fire perimeters from 1986 to 2020 from Monitoring Trends in Burn Severity (MTBS) records that overlap with the the 'Rehefeldt_laoc_distro_US_1km_buffer' polygons.
- topofire_aet_gs_1991_2020_laoc_distro.tif AND topofire_def_gs_1991_2020_laoc_distro.tif
- Rasters of 30-yr mean, actual evapotranspiration ('aet') and climatic water deficit ('def') at 250m pixel resolution from Topofire model.
- citation: Holden, Z. A., Swanson, A., Klene, A. E., Abatzoglou, J. T., Dobrowski, S. Z., Cushman, S. A., Squires, J., Moisen, G. G., & Oyler, J. W. (2016). Development of high-resolution (250 m) historical daily gridded air temperature data using reanalysis and distributed sensor networks for the US Northern Rocky Mountains. International Journal of Climatology, 36(10), 3620–3632.
- Model_Prediction [folder]:
- Rasterised estimations of recruitment probability of western larch after fire given climatic water deficit and actual evapotranspiration for time since fire years 1 though 5 and 6 though 10, for each year from 1991 to 2021 and the 30-yr mean (1991 to 2020). The probability of recruitment was estimated while holding distance to seed source and fire severity constant at their median values (i.e., 24.5 m and dNBR of 545).
- laoc_dnbr_19862020.csv
- pixel values of fire severity (dNBR) at 30-m pixel resolution of all fires that occurred within the the distribution of western (w/ 1 km buffer; 'Rehfeldt_laoc_distro_US_1km_buffer')
- laoc_fires_19862020.csv
- Describes area of each fire severity class (i.e., low, moderate, and high) of all fires that overlap and where they overlap within the the distribution of western larch (w/ 1 km buffer; 'Rehfeldt_laoc_distro_US_1km_buffer'). Also describes the mean fire severity of the total burned area that overlaps within the distribution of western larch.
- laoc_fires_19862020_DSS_summary.csv
- Dataset used for analysing the the total proportion of area within 100 meter of seed source (i.e. core cells) after fires by year within the the distribution of western larch.
FILES' STRUCTURES AND UNITS OF MEASUREMENTS
Vieira_et_al_climate_timeseries.csv
The file is a comma delimited file.
Column headings are in the following order and format:
plot
- Name of the site.
Long_WGS84
- longitude in decimal degrees and WGS84 coordinate reference system
Lat_WGS84
- latitude in decimal degrees and WGS84 coordinate reference system
regen_year
- year of the climate conditions that follow.
tmeanC_mean_spring_30m
- Mean air temperature (degrees Celsius) occuring in the recruitment season (April to May) each year.
tskinC_cc10_max_summer_ECH20
- Maximum land surface temperature (degrees Celsius) occuring in the summer(June - August) each year.
soilmoist_10cm_cc10_mean_spring_ECH20
- Mean soil moisture (volumetric water content) in the recruitment season (April-May) in the top 10 cm of the soil.
soilmoist_10cm_cc10_min_summer_ECH20
- Minimum soil moisture (volumetric water content) in the summer (June - August) in the top 10 cm of the soil.
GDD_5_30m
- Cumulative growing degree-days over 5 degrees C of each year
topofire_def_mm_sum_gs_30m
- Growing season (May to Sept.) climatic water deficit (mm) at 30m pixel resolution
topofire_aet_mm_sum_gs_30m
- Growing season (May to Sept.) actual evapotranspiration (mm) at 30m pixel resolution
topofire_def_mm_sum_an_mean_19912020_30m
- 30-yr mean (1991-2020) growing season (May to Sept.) climatic water deficit (mm) at 30m pixel resolution
topofire_aet_mm_sum_gs_250m
- Growing season (May to Sept.) actual evapotranspiration (mm) at 250m pixel resolution
topofire_def_mm_sum_gs_250m
- Growing season (May to Sept.) climatic water deficit (mm) at 250m pixel resolution
[further details on climate data in journal article, Table 1)
Vieira_et_al_plot_data.csv
The file is a comma delimited file.
Column headings are in the following order and format:
plot
- Name of the site.
plot_size_m2
- Area (m^2) of the belt transect that was used to destructively sample wester larch juveniles for aging.
Fire_name
- Name of fire according to Monitorig Trends in Burn Severity (MTBS) records
Fire_ID
- ID# of fire according to Monitorig Trends in Burn Severity (MTBS) records
Fire_year
- Year fire at site occurred
year_sampled
- year of field sampling (i.e., 2021 or 2022)
Long_WGS84
- longitude in decimal degrees and WGS84 coordinate reference system
Lat_WGS84
- latitude in decimal degrees and WGS84 coordinate reference system
elevation
- elevation at site recorded in the field with GPS device
slope_percent
- - Slope of the site (%) recorded in the field with a clinometer.
aspect_field_deg
- Aspect of the site (degrees) recorded in the field with a compass.
aspect_dem_deg
- Aspect of the site (degrees) from digital elevation model.
severity_ground
- estimation of fire severity class recorded in the field.
dnbr
- Fire severity (differenced normalized burn ratio) extracted using protocol from Parks et al. (2018). Where the transect crossed two cells with different dNBR values they were averaged.
dss_m
- Averaged distance (m) to the nearest reproductively mature (cone-bearing) western larch individual at 0, 30, and 60 m along transect, measured with a laser range finder.
tsf_yrs
- Time between the fire year and the year of recruitment ("pith_year").
regen_yr
- year of recruitment following the fire.
laoc_regen_count
- count of juveniles established at a site and associated recruitment year.
laoc_regen_ha
- Annual recruitment density (juveniles/m^2) of western larch at the given site that occurred in the given recruitment year. This is the count divided by the transect area for the given species.
PA
- Presence (1) or absence (0) of recruitment at the given site during the given recruitment year.
PERC25
- Success/failure (1/0) of reaching the 25th percentile of recruitment at the given site during the given recruitment. Calculated from recruitment density of all years with recruitment across plots
PERC50
- Success/failure (1/0) of reaching the 50th percentile of recruitment at the given site during the given recruitment. Calculated from recruitment density of all years with recruitment across plots
cv.num
- Unique numeral characters used in place of names of the fires, used to group folds for cross validation in gbm.step (boosted regression tree models). In cases where none of the plots within a fire had recruitment above our annual recruitment thresholds that fire was included with the most similar (climatically and geographically) fire to it for the purposes of cross validation.
tmeanC_mean_spring_30m
- Mean air temperature (degrees Celsius) occuring in the recruitment season (April to May) each year.
tskinC_cc10_max_summer_ECH20
- Maximum land surface temperature (degrees Celsius) occuring in the summer(June - August) each year.
soilmoist_10cm_cc10_mean_spring_ECH20
- Mean soil moisture (volumetric water content) in the recruitment season (April-May) in the top 10 cm of the soil.
soilmoist_10cm_cc10_min_summer_ECH20
- Minimum soil moisture (volumetric water content) in the summer (June - August) in the top 10 cm of the soil.
GDD_5_30m
- Cumulative growing degree-days over 5 degrees C of each year
topofire_def_mm_sum_gs_30m
- Growing season (May to Sept.) climatic water deficit (mm) at 30m pixel resolution
topofire_aet_mm_sum_gs_30m
- Growing season (May to Sept.) actual evapotranspiration (mm) at 30m pixel resolution
topofire_def_mm_sum_an_mean_19912020_30m
- 30-yr mean (1991-2020) growing season (May to Sept.) climatic water deficit (mm) at 30m pixel resolution
topofire_aet_mm_sum_gs_250m
- Growing season (May to Sept.) actual evapotranspiration (mm) at 250m pixel resolution
topofire_def_mm_sum_gs_250m
- Growing season (May to Sept.) climatic water deficit (mm) at 250m pixel resolution
[further details on climate data in journal article, Table 1)
laoc_fires_19862020.csv
The file is a comma delimited file.
Column headings are in the following order and format:
Incid_Name
- Name given to wildfire and archived in MTBS dataset.
Fire_ID
- Unique ID given to wildfire and archived in MTBS dataset.
hectare
- hectares of fire overlapping with distribution of western larch in the US (i.e., 'Rehfeldt_laoc_distro_US.shp').
year
- year of wildfire ignition.
Low_Severi
- hectares of low fire severity portion of wildfire --extracted using protocol from Parks et al. (2018) -- that overlaps with distribution of western larch in the US (i.e., 'Rehfeldt_laoc_distro_US.shp').
Mod_Severi
- hectares of moderate fire severity portion of wildfire --extracted using protocol from Parks et al. (2018) -- that overlaps with distribution of western larch in the US (i.e., 'Rehfeldt_laoc_distro_US.shp').
High_Sever
- hectares of high fire severity portion of wildfire --extracted using protocol from Parks et al. (2018) -- that overlaps with distribution of western larch in the US (i.e., 'Rehfeldt_laoc_distro_US.shp').
Mean_Sever
- mean fire severity (dNBR) across overlap with distribution of western larch in the US (i.e., 'Rehfeldt_laoc_distro_US.shp').
laoc_fires_19862020_DSS_summary.csv
The file is a comma delimited file.
Column headings are in the following order and format:
fire_year
- year in which the fires ignited
core_cells
- the number of pixels burned at high severity in the given fire year that were >100 m from the nearest unburned pixel or pixel burned at low or moderate severity and that fell within the distribution of western larch in the US (i.e., 'Rehfeldt_laoc_distro_US.shp').
total_cells
- the number of pixels within fire perimeters that burned in the given fire year and fell within the distribution of western larch in the US (i.e., 'Rehfeldt_laoc_distro_US.shp')..
number_fires
- the number of fires in the given fire year that overlapped with western larch’s distribution and were used for analysis (only part of some fires were used based on western larch’s distribution within each fire).
total_prop_dss_100
- the proportion of all cells burned at high severity in western larch’s distribution in the given fire year that were >100 m from an unburned pixel or a pixel burned at low or moderate severity.
DATA QUESTIONS
Questions about data can be addressed to Spencer Vieira (stvieira@alaska.edu).