SoCal EcoServe: an online mapping tool to estimate wildfire impacts in southern California

Introduction

Wildfire is a characteristic of the world’s Mediterranean-type climate regions. In California, the extent of shrubland burned and the frequency of wildfires are of ecological and socio-economic concern, especially given the proximity of shrublands to heavily populated coastal regions (Keeley and Safford 2016; Safford et al. 2022). The environmental impacts of wildfires vary depending on fire characteristics such as intensity, size and frequency (Roces-Díaz et al. 2022). The combustion of shrublands leads to both negative and positive impacts. Most notably, combustion reduces biomass and fuels and the ability of landscapes to sequester carbon from greenhouse gases (Pausas and Keeley 2019). With the removal of vegetation, soil erosion, water runoff and groundwater recharge increase, especially in the first year or two post-fire (Wohlgemuth and Lilley 2018; Flint et al. 2019). In addition, other services such as recreational opportunities on public lands or habitat for wildlife and biodiversity can be temporarily lost.

Understanding and quantifying fire impacts on ecosystem services is increasingly important given potential impacts of future climates on wildfire. Shrublands in southern California experience burning under two scenarios: spring/summer fires where higher fuel moisture and moderate onshore winds restrict the area burned, whereas autumn/winter fires occur when ignitions are influenced by hot strong winds, like the Santa Ana, and the areas burned are extensive (Safford et al. 2022). Warming climates are increasing drought and, as a consequence, are contributing to the expansion of the fire season year-round in California (Diffenbaugh et al. 2015). More drought is increasing shrub mortality and fuel loads, while increasing short-interval fire results in vegetation type conversion and expansion of annual (largely non-native) grasslands providing highly combustible fuels (Safford et al. 2021; Pratt 2022). Moreover, post-fire drought negatively influences seedling survival and growth of resprouting shrubs (Pratt et al. 2014) and thus shrubland recovery.

For resource managers of shrubland-dominated national forests, being able to quantify the impacts of wildfire is important for numerous reasons, including tracking carbon storage and estimating the impacts of wildfire on carbon storage, understanding fuel loads to inform fire models and prescriptive fuel treatments, and identifying sites for post-fire restoration. Furthermore, generating data on wildfire impacts is necessary for post-fire Natural Resource Damage Assessments frequently conducted for fires on public lands, which are used to estimate and acquire cost-recovery funds to complete much needed post-fire restoration.

To date, there has been no systematic method for estimating fire impacts on shrubland ecosystem services: impacts have been estimated with little standardisation between fires. In addition, there has been no tool for resource managers to readily access relevant spatial data associated with ecosystem services. To fill this critical need of resource managers, we developed an online mapping tool that allows users to view, query and download pre- and post-fire data on ecosystem services. The SoCal EcoServe tool (https://manzanita.forestry.oregonstate.edu/EcoServeHome/) provides the first transparent, reproducible and scientifically sound tool for resource managers of shrublands in southern California. It provides an important contribution by packaging and summarising data on ecosystem services and wildfires that have been previously published (e.g. Underwood et al. 2018; Schrader-Patton and Underwood 2021; Schrader-Patton et al. 2022, 2023; Underwood and Hollander 2023). The six ecosystem services included in the tool were identified by resource managers at the beginning of the project as those most valuable to understand for management of southern California shrublands, as well as practical to model within the timeframe of the project. The ecosystem services included are: carbon storage, water runoff, groundwater recharge, sediment erosion, recreation services and biodiversity.

Although other studies have compiled and modelled data on ecosystem services in California (e.g. Chan et al. 2006; Chang and Bonnette 2016) as well as estimated fire impacts on services (e.g. Kinoshita and Hogue 2011; Flint et al. 2019), this is the only tool we are aware of for resource managers to access ecosystem services data associated with impacts of specific wildfires across an ecoregion. SoCal EcoServe currently includes large fires since 2017 that have occurred on public lands and have burn severity data available. To date, the tool has been tested on a limited number of fires and has received positive review by resource managers, and we are now rolling it out to USDA Forest Service (USFS) resource managers more widely across southern California.

In this paper, we describe the methods used to quantify wildfire impacts on ecosystem services and illustrate the tool and data outputs using a case study of the Alisal Fire, which burned on the Los Padres National Forest in 2021. We selected the Alisal Fire as it covered a large spatial area, is dominated by shrub vegetation and the majority of the fire burned at high severity.

Methods

SoCal EcoServe tool

The SoCal EcoServe tool consists of two components: a behind-the-scenes desktop tool, currently managed by select USFS staff responsible for maintaining EcoServe, that prepares the ecosystem services raster data, and an online mapping tool that allows users to query these outputs. The goal of the desktop tool is to process all fires on USFS National Forest lands within the southern California ecoregion (Bailey 2016) (Fig. 1). It requires two inputs: the ecosystem services data in pre-fire condition and burn severity data. More specifically, it uses the fire-induced canopy cover loss produced by the USDA Forest Service’s Rapid Assessment of Vegetation Condition after Wildfire (RAVG), which is produced for all fires over 1000 acres (405 ha) on public lands. In a small number of instances, RAVG data may not be available, e.g. in fires <405 ha (1000 acres) or if running the tool on non-USDA Forest Service lands. As an alternative, the canopy cover loss data can be generated through the use of a Google Earth Engine application (Parks et al. 2018).

Fig. 1.

Map of southern California displaying the area within which the EcoServe tool can currently be applied. The boundary consists of the Level 4 ecoregions surrounding the four National Forests in southern California.

Using a Python script, the desktop tool clips the input ecosystem services raster files to the fire boundary and also to the portion of the fire that is within USFS lands, performs the raster calculations and prepares the tabular outputs as described for each ecosystem service below. Outputs from the desktop tool are compiled into a zip file that is then uploaded by Forest Service staff to the server hosting the online mapping tool.

The EcoServe online mapping tool provides the interface for users to query the ecosystem services data prepared in the desktop tool, view outputs and download data as tables and rasters to include in reports or to conduct further spatial analysis. The tool accounts for all vegetation types within the fire perimeter; however, additional data relating to carbon storage are available for shrublands. The online mapping tool is an ASP.NET Core Model-View-Controller (MVC) application and built using Microsoft Visual Studio 2019 software using C# (C-sharp) programming language along with hypertext markup language (HTML), Cascading Style Sheets (CSS) and Javascript. It is hosted on a Dell PowerEdge R440 server running Windows 2016 Standard, build 10.0.14393. Below, we summarise the ecosystem services data developed for southern California that are used in the tool and provide citations to previous uses of these data.

We illustrate the EcoServe tool using the Alisal Fire, 6868 ha (16 970 acres). The fire was started by lightning on 11 October 2021, and occurred on the Los Padres National Forest in Santa Barbara County.

Results

We obtained the Alisal Fire perimeter and burn severity raster data from the Burn Severity Portal (https://burnseverity.cr.usgs.gov/) and created the pre- and post-fire ecosystem services data using the desktop tool, and uploaded outputs to the EcoServe server. A resource manager accesses spatial data and estimates of fire impacts by opening the online mapping EcoServe tool (https://manzanita.forestry.oregonstate.edu/EcoServeHome/). On starting the tool, the user is presented with a map of the southern California ecoregion and options to add viewable layers, e.g. National Forest boundaries, hydrological unit boundaries (Fig. 2). A user then selects the year, fire (either entire perimeter or portion of fire on National Forest land), and one of six ecosystem services to quantify or ‘hotspots of change’ (Fig. 2). In the map display for carbon storage, surface runoff, groundwater recharge and sediment erosion, the user can toggle between viewing spatial data in pre-fire and post-fire condition, RAVG burn severity data and (for carbon storage) pixels identified with low regeneration potential.

Fig. 2.

The initial page of the SoCal EcoServe tool, showing three steps for selecting year, fire and service to query, along with options for changing map background layers. Red boundary shows the Alisal Fire (2021).

The outputs for carbon storage services in the EcoServe tool include a table summarising data of pre- and post-fire carbon storage for five different biomass pools (Table 1) and corresponding bar chart; summary of major shrub communities in the fire; and the amount of carbon ‘permanently lost’ post-fire. The majority of the Alisal Fire consisted of shrub (69%), with mixed chaparral (54%) dominating, followed by coastal sage scrub (14%).

The carbon stored in vegetative carbon pools in the Alisal Fire pre-fire totalled 102 944 metric tons, which was slightly higher than the 7-year average (99 130 metric tons). The aboveground live biomass and standing dead pools contributed relatively more than the belowground and litter (Table 1, Fig. 3). Using biomass data for the post-fire year (2022), carbon storage decreased by 25% across all vegetative pools to 76 941 metric tons (Table 1). Recognising that shrubland that experiences high-frequency short-interval fire is unlikely to recover and is vulnerable to conversion to annual grassland (Zedler et al. 1983; Syphard et al. 2019), we estimated 13 831 metric tons or 20% of the pre-fire total vegetative biomass could be permanently lost based on the identification of pixels with low regeneration potential according to fire history. The Alisal Fire had 1344 ha (3322 acres or 20% of the fire) with low regeneration pixels owing to the occurrence of the Sherpa Fire (2016) that occurred in the eastern portion of the Alisal burn perimeter.

Fig. 3.

Graphical outputs of EcoServe tool for the Alisal Fire showing different vegetative carbon pools pre- and post-fire.

For runoff, recharge and sediment erosion services that integrated the RAVG burn severity data, the tabular outputs reported in the EcoServe tool include the area within each canopy cover loss class. For the Alisal Fire, 57% experienced 75–100% canopy cover loss; 11% of the fire experienced 1–25% canopy cover loss; and 6% of the fire experienced 25–50 and 50–75% loss each (Fig. 4).

Fig. 4.

EcoServe tool showing outputs for water runoff pre- and post-fire for the Alisal Fire and reported in table by different canopy cover loss classes in the RAVG burn severity data. Map image also shows display options including pre-fire, post-fire water runoff and burn severity data.

The difference between pre-fire and post-fire (normalised) change in water runoff for the Alisal Fire (calculated based on the 12-year average) showed an increase of 21%, from 1154 to 1401 ha metres year⁻¹ (Table 2, note the tool reports runoff and recharge in acre feet per year but here we report in SI units of ha metres a year). The increase in groundwater recharge was greater, increasing seven times from 73 to 560 ha metres year⁻¹ (Table 2). Sediment erosion increased by 24-fold from an average pre-fire estimate of 5499 m³ year⁻¹ to an average post-fire estimate of 131 407 m³ year⁻¹ (Table 2). The difference in hydrological and sediment erosion services can also be viewed in graphical format in the tool (Fig. 5). In the case of the Alisal Fire, there were no recreation sites contained in the two data sources integrated into the EcoServe tool. However, in other shrubland fires currently integrated in the tool, the number of recreation sites ranges from 1 to 17.

Fig. 5.

Graphical outputs of EcoServe tool for the Alisal Fire showing water runoff pre- and post-fire including average pre-fire for comparison.

The mean Marxan biodiversity score for the Alisal Fire was 27 out of a maximum irreplaceability score of 100. Highest biodiversity scores were mapped along the southern edge of the fire perimeter (Fig. 6). The table output compares biodiversity information for the Alisal Fire with the four southern National Forests (Fig. 6): the mean Marxan score ranked third after the San Bernardino and Cleveland National Forests. The Alisal Fire had rare and native species richness levels similar to the four National Forests (California Department of Fish and Wildlife’s Areas of Conservation Emphasis Map data 2010, https://map.dfg.ca.gov/ace/) (table in Fig. 6).

Fig. 6.

Biodiversity output from Marxan showing irreplaceability of planning units within the Alisal Fire. Tabular results compare the Alisal Fire with the four southern National Forests.

The ‘hotspots of change’ calculation showed the majority of the Alisal Fire was impacted by high change (defined as >80% change) in groundwater recharge and sediment erosion (78% and 74% respectively, Fig. 7), whereas only 1% of the fire experienced high change in water runoff. As reported previously under post-fire carbon storage, the EcoServe tool identified 20% of the fire (1 344 ha) with low regeneration potential. These data appear in tabular format in the tool and the change rasters can be viewed in the map window (Fig. 7).

Fig. 7.

Table (right) reporting change in ecosystem services in Alisal Fire. Map (left) showsoptions for map display for the hotspots of change selection.

Finally, a range of data presented in the EcoServe online mapping tool can be downloaded for reports or conducting additional spatial analyses. Data include: pre- and post-fire raster spatial data and change data for carbon storage, runoff, recharge and sediment erosion, and chart and tabular data can be downloaded for each service. For biodiversity, the downloadable table includes additional information such as a list of rare species by taxonomic groups (California Natural Diversity Database, https://www.wildlife.ca.gov/Data/CNDDB).

Discussion and conclusion

We have described the input data on ecosystem services and the methodology used in the SoCal EcoServe tool to quantify the impacts of fire in shrublands in southern California and report outputs using the Alisal Fire as a case study. The outputs generated by EcoServe provide important data for resource managers to quantify fire impacts for assessing damages to natural resources on National Forest lands. This was identified as a need by resource managers and will become increasingly important given that future climates will result in higher fuel loads from drought-related shrub mortality and vegetation type conversion associated with drought, fire and nitrogen deposition (Allen et al. 2018; Safford et al. 2021; Pratt 2022).

Outputs from the EcoServe tool showed the Alisal Fire reduced carbon storage by 25% across all vegetative biomass pools (Table 1) and approximately one-fifth of this (1344 metric tons) is estimated not to recover and be permanently lost. When interpreting the change in carbon storage post-fire, it is important to note the time between the fire occurrence (in this case 11 October 2021; note ‘Fire start date’ in Fig. 3) and the date of the post-fire AGLBM data (July-August 2021). For the Alisal Fire, there were 9 months for shrub vegetation to recover post-fire. This should be noted when comparing change in carbon storage between different fires; for example, a spring fire will have fewer months for post-fire recovery before the July-August date of the post-fire biomass data.

In applying the EcoServe tool to the Alisal Fire, there are a number of discussion points associated with outputs of the different ecosystem services. First, the Alisal Fire had a notable increase in water runoff of 21%, which is expected with the reduction in evaporation and transpiration as vegetation is removed. To provide context for this output, other shrub-dominated fires in southern California have experienced water runoff increases between 8 and 46% post-fire (E. Underwood, unpubl. data). However, for fires occurring in years with severe or extreme drought conditions, runoff (and recharge) have been found to decline (versus increase) post-fire. This was a key motivation for reporting the average runoff and recharge over 12 years, which capture a range of climatic conditions. For groundwater recharge, the 12-year pre-fire average for the Alisal Fire is similar on a per hectare basis to other shrub-dominated fires in the tool such as the Willow (2021), Thomas (2017), Whitter (2017) and Valley (2020) fires. Groundwater recharge increased seven times post-fire which, in part, is explained by it occurring on excessively drained soils (Maymen series) that will have relatively more recharge when the transpiring plant cover is removed.

Second, the substantial increase in sediment erosion after the Alisal Fire is a pattern found in other shrub-dominated fires both in 2021 and in other years. One caveat associated with the InVEST model is that it only considers annual precipitation and does not account for the intensity of individual winter storm events that characterise southern California and cause substantial erosion. Furthermore, the loss of soil also reduces ecosystem productivity upslope as well as impacts infrastructure downstream, which are not reflected in this estimate of impact.

Third, there were no recreation sites reported for the Alisal Fire using the two datasets currently in the EcoServe tool which, in part, reflects the fairly remote location of the fire. However, there may well be recreation facilities affected but just not contained in these datasets. Fourth, in interpreting the summary of the biodiversity data, it is important to note that the output is dependent on the individual biodiversity layers entered into the Marxan software. Different conservation targets and associated goals would result in different solutions. Furthermore, the input biodiversity data originated in different years.

Finally, in relation to areas identified as ‘hotspots of change’ post-fire, there is variation in the amount of area identified for each ecosystem service. This is, in part, due to the services being generated by different models and approaches (i.e. water runoff and groundwater recharge from the Basin Characterisation Model and sediment erosion retention from InVEST). In addition, it is important to note that this is only one way to define hotspots; other approaches could include integrating ecosystem services data with fire frequency, climatic water deficit data, or non-native species, which could be incorporated into future versions of the tool.

Currently, the outputs from the EcoServe tool report post-fire impacts over the short term, which is often sufficient for quantifying some immediate environmental damages to resources. However, we also recognise that understanding longer-term impacts is equally important for managing shrublands. We have taken steps to do this for carbon storage in two ways. First, we calculate the permanent loss of carbon storage over the long-term given fire history data and second, we include in the tool a link to a graph showing the trajectory of biomass recovery for mixed chaparral shrubland. One challenge in projecting recovery is the difference in temporal scales for each service. For example, in the spring following shrubland fires, a flush of wildflowers can attract visitors (Safford et al. 2018), even though recreation facilities and infrastructure may take much longer to be usable; similarly, hydrological services are estimated to recover 2–4 years post-fire however have been found to be more sensitive to climate (Flint et al. 2019).

Climate, particularly drought conditions, is a major challenge for developing pre- and post-fire data relating to carbon storage, hydrology and sediment erosion. We address the year-to-year variations in these services by reporting the pre-and-post-fire values for these services using a 12-year average climatic condition. This approach has the advantage of facilitating comparisons of fires occurring in different years, increasing the influence of burn severity (RAVG) data over climate on determining fire impacts, and allowing fire impacts to be calculated as soon as burn severity data are available.

SoCal EcoServe is primarily a compilation and aggregation tool for existing ecosystem services data that have been previously published. The use of burn severity data to capture fire impacts on biomass, hydrology and sediment erosion has appeared in previous studies (e.g. Neary et al. 2011; Sankey et al. 2015). Ideally, a quantitative landscape-level assessment of the ecosystem services in pre- and post- fire condition generated by the EcoServe tool is required to conduct a full validation of outputs; however, this would be challenging in practice (e.g. it would require a priori knowledge of a fire perimeters or fires located in the identical geographies burning at different severities to measure). However, we welcome the opportunity to fully validate EcoServe results should appropriate data become available. Development of spatial ecosystem services data is an active area of research and we are open to future datasets and methods that could contribute to capturing pre- and post-fire conditions. That said, we believe the data and methods used in EcoServe are currently the best available for generating data on fire impacts to shrub-dominated ecosystems in southern California.

The spatial data on ecosystem services within the current EcoServe tool are limited to the southern California ecoregion (Fig. 1). However, the same tool framework can be utilised if data on ecosystem services inputs are generated for shrublands in different areas of California or the western US, including non-National Forest lands if burn severity data are generated through alternative means (e.g., Park). The EcoServe tool outputs can be used to determine the loss of ecosystem services to inform natural resource damage assessments, help prioritise restoration activities post-fire (e.g. focus on high fire impactareas or hotspots of change), identify areas vulnerable to type-conversion to non-native grasses (using the data on low biomass regeneration potential), and, ultimately, contribute to supporting resource managers to ensure the provision of ecosystem services over the long term.

Data availability

The SoCal EcoServe tool and associated ecosystem services data are available for selected fires at: https://manzanita.forestry.oregonstate.edu/EcoServeHome/, along with methods and caveats for the approach.

Conflicts of interest

The authors declare no conflicts of interest.

Declaration of funding

Funding for this work was provided by the USDA Forest Service Pacific Southwest Region, the USDA Western Wildlands Environmental Threat Assessment Center, the National Fish and Wildlife Foundation, and the California Department of Forestry and Fire Protection.