Diane M Thomson1, Wallace M Meyer2, Isobel F Whitcomb1. 1. W.M. Keck Science Department, Claremont, California, United States of America. 2. Department of Biology, Pomona College, Claremont, California, United States of America.
Abstract
Non-native plant invasions, changes in fire regime, and increasing drought stress all pose important threats to biodiverse mediterranean-climate shrublands. These factors can also interact, with fire and drought potentially creating opportunities for non-native species to establish dominance before native shrubs recover. We carried out post-fire demographic monitoring of the common native shrub Artemisia californica in a southern California sage scrub fragment for 7 years, including several with very low rainfall. Experimental removals of non-native plants were included for the first 4 years. We quantified A. californica post-fire crown resprouting and seedling emergence, and tested effects of precipitation, non-native plants, and their interactions on seedling and adult survival. Only 7 A. californica were confirmed as resprouts; almost all individuals established after the fire from seedlings, with 90% of emergence occurring in the second growing year after fire (spring 2015). Higher spring precipitation increased both adult and seedling survival. Non-native grasses and forbs rapidly recolonized control plots, but the removal treatment reduced non-native cover by nearly 60%. For seedlings, non-native removal reduced the probability of dropping leaves by start of summer drought and increased survival both directly and through positive interactions with rainfall. Non-native removal also reduced mortality in smaller adult plants. By 2020, mean A. californica canopy area was nearly four times greater in non-native removal plots. These findings reinforce the high vulnerability of sage scrub habitat to post-fire loss of shrub cover and potential type conversion, particularly with increasing drought frequency and in stands and species with limited crown resprouting. Yet they also illustrate the potential for targeted management of non-natives immediately after fire to promote recovery of native shrubs in this increasingly endangered community.
Non-native plant invasions, changes in fire regime, and increasing drought stress all pose important threats to biodiverse mediterranean-climate shrublands. These factors can also interact, with fire and drought potentially creating opportunities for non-native species to establish dominance before native shrubs recover. We carried out post-fire demographic monitoring of the common native shrub Artemisia californica in a southern California sage scrub fragment for 7 years, including several with very low rainfall. Experimental removals of non-native plants were included for the first 4 years. We quantified A. californica post-fire crown resprouting and seedling emergence, and tested effects of precipitation, non-native plants, and their interactions on seedling and adult survival. Only 7 A. californica were confirmed as resprouts; almost all individuals established after the fire from seedlings, with 90% of emergence occurring in the second growing year after fire (spring 2015). Higher spring precipitation increased both adult and seedling survival. Non-native grasses and forbs rapidly recolonized control plots, but the removal treatment reduced non-native cover by nearly 60%. For seedlings, non-native removal reduced the probability of dropping leaves by start of summer drought and increased survival both directly and through positive interactions with rainfall. Non-native removal also reduced mortality in smaller adult plants. By 2020, mean A. californica canopy area was nearly four times greater in non-native removal plots. These findings reinforce the high vulnerability of sage scrub habitat to post-fire loss of shrub cover and potential type conversion, particularly with increasing drought frequency and in stands and species with limited crown resprouting. Yet they also illustrate the potential for targeted management of non-natives immediately after fire to promote recovery of native shrubs in this increasingly endangered community.
Mediterranean biome regions are important global biodiversity hotspots, harboring high species richness and under intense pressure from human impacts [1]. Mediterranean-climate ecosystems have undergone extensive losses to urban and agricultural development, and negative effects of fragmentation threaten the habitat that remains [2]. Growth in the wildlands-human interface facilitates the spread of non-native invasive species that can reduce native plant richness [3]. Many mediterranean-climate communities evolved with and depend upon fire, but whether fires promote persistence of native diversity depends on the frequency and intensity [4]. Anthropogenic changes in fire regime pose another major threat to these habitats [2, 5, 6].Interactions between fire and non-native plant invasion can create positive feedback loops that further promote loss of native cover and diversity [7, 8]. Fires open up opportunities for non-natives to invade, particularly annual grasses and forbs that respond to disturbance more rapidly than native perennials [9]. Non-native grass and forb invasion in turn magnify fine fuel loads, potentially increasing ignition risk, fire frequency, or fire intensity [10, 11]. These changes in fire regime can lead to habitat degradation and even type conversion, meaning a major and persistent shift in community structure such as replacement of native shrubland by non-native grassland [12].Southern California supports two native shrubland communities heavily impacted by ongoing type conversion, chaparral and sage scrub [13]. By 1994, sage scrub had declined to an average of 36% shrub cover compared to the 60–90% observed 60 years earlier [14]. Similarly, between 1953 and 2016 nearly 30% of study plots in chaparral shifted from shrub to herbaceous dominance of cover [15]. Annual grasses such as Bromus diandrus, Bromus madritensis, and Avena fatua are the most abundant non-native species in these type-converted communities, along with annual forbs such as Brassica spp. and Hirschfeldia spp. Increased fire frequency is associated with higher non-native grass and forb cover in both chaparral and sage scrub [14, 16–18]. Yet the directions of any causal relationships remain unclear; these correlations could reflect the effects of fire disturbance on non-native plants, the effects of non-native plants on ignition risk and fuel loads, or both.The first few years after fire in southern California shrublands may be a critical window determining whether native communities recover or non-native invasion takes hold [19]. Once annual grasses are established they reduce native seedling germination and survival, through suppressive effects of litter [20, 21] and alteration of soil water availability [22] or soil microbial communities [23]. In contrast, high shrub cover helps control non-native grasses and forbs through shading, nitrogen depletion, and herbivory [24-27]. As a result, rapid shrub recovery after fire is critical to prevent non-native grasses from establishing dominance [12, 28]. Reintroducing shrubs to restore type-converted, non-native grasslands has proved very difficult [5, 29].This key role of the years immediately after fire potentially also amplifies effects of increased drought and climate warming on native shrub communities. Low rainfall years can reduce shrub survival and recruitment [24, 30], and more arid regions have proved at greater risk of non-native invasion after fire [12, 19]. Between 2011 and 2018, southern California experienced an intense drought that included the driest conditions of the last 1200 years [31]. This drought followed a period of longer-term drying trends over the last several decades, increasing the urgency of understanding how precipitation affects post-fire recovery and non-native invasion [32].Both fire ecology and non-native invasions in many respects have been extensively investigated in southern California shrublands. Yet post-fire monitoring [33] and experimental tests of non-native plant effects [21] generally are not integrated into the same studies. Moreover, most previous southern California post-fire research followed large-scale burns in 1978 and 1993, before the onset of recent intense drought conditions (but see [34]). While a few studies document the importance of water availability to native shrubs in sage scrub [35], none have yet developed statistical models that link rainfall with shrub demographic rates.We tracked recovery of the common shrub Artemisia californica over the first 7 years after a small-scale fire in a sage scrub fragment, combined with experimental removal of non-native plants. We quantified A. californica post-fire resprouting and seedling emergence, as well as survival and growth of both seedlings and established plants. Experimental removals of non-native grasses and forbs were carried out during the first four years to evaluate their effects on post-fire demography and recovery of A. californica. We tested effects of precipitation, non-native plants, and their interactions on seedling and adult survival.
Materials and methods
Study system
As is typical in other mediterranean climate regions, southern California shrublands include an evergreen, sclerophyllous community type (chaparral) and a drought-deciduous, softer-leaved one (sage scrub) [36]. Sage scrub concentrates along the Pacific coast from San Francisco, CA (latitude 37.3) to El Rosario, Mexico (latitude 30.06), with this distribution in some areas extending eastward towards the Mojave Desert [37]. Sage scrub supports high total plant diversity, but local within-patch species richness tends to be low [36]. Species composition varies widely over short distances, and several subtype classifications have been proposed based on specific plant associations [38]. Estimates of historic habitat loss vary between 40% and 90%, with much of the remaining range considered degraded [5, 39]. Sage scrub supports more than 60 plant and more than 30 animal taxa classified as rare, threatened, or endangered [40].Historic fire return intervals in sage scrub are uncertain but have been estimated at about 30 years [41]. In some areas with higher non-native grass cover, fire intervals have shortened to less than 8 years [14]. Post-fire recovery in both sage scrub and chaparral is mostly driven by species present at the time of fire, which regenerate either through resprouting or germination from the seed bank [42]. The majority of woody shrubs are facultative seeders that can both resprout and regenerate from seed, with the balance of these two processes varying across species and habitats [42]. Chaparral species resprout at higher rates after fire than sage scrub dominants, but may experience little to no recruitment without fire [41]. In contrast, sage scrub shrubs can germinate in gaps without fire disturbance [24]. Several studies suggest that the ratio of resprouting to recruitment from seed in sage scrub declines across a moisture gradient, from more mesic coastal to drier inland habitat [33, 43].Artemisia californica is a suffrutescent sub-shrub that occurs in chaparral but is more common in sage scrub. Among the most widely distributed sage scrub species, A. californica is co-dominant at many sites [38, 39]. This species is both facultatively drought-deciduous and seasonally dimorphic, producing smaller leaves during summer [37]. Artemisia californica appears as a community dominant primarily in south coastal areas, replaced by species such as Encelia farinosa in interior regions [39]. Across sage scrub sites, A. californica reaches peak cover with intermediate temperatures and low litter [38]. Artemisia californica has been classified as a facultative seeder and can crown resprout after fire, but at lower rates than some other common co-occurring species such as Salvia apiana and Eriodictyon trichocalyx var. trichocalyx [37, 42, 44].We conducted this study at the Robert J. Bernard Field Station, which supports fragments of intact sage scrub on 34 hectares embedded in a suburban landscape (34.8 ha; 34°6’ N, 117°42’ W; 348 m elevation; Claremont, CA, U.S.A.). Stands of sage scrub are bordered by roads and a matrix of anthropogenically altered habitats, including type-converted annual grassland. The exact time of last previous fire is unknown but extends back at least 60 years. The climate is mediterranean, with cool, wet winters and warm, dry summers. Winter rains typically start in October and end in April or May with onset of summer drought. Growing year rainfall (September to August) averages 415.5 mm per year (n = 95 years with complete records; 1896–1978 NOAA Claremont Station; 2000–2020 Western Regional Climate Center data for Claremont, CA). Over the study period, growing year rainfall varied from 55.4% below to 79.3% above this mean; the first growing season after fire (2013–2014) experienced almost exactly mean precipitation (418 mm), while four of the six subsequent years were at least 24% below average (S1 Table).Late in the 2013 dry season (September), an accidental fire burned 6.9 hectares containing two patches of sage scrub separated by a road (Fig 1). The western burned patch included approximately 1000 m2 of sage scrub bordered by non-native grassland on two sides. The eastern patch encompassed about 4000 m2 of sage scrub bordered by non-native grassland on all sides. Artemisia californica dominated pre-fire vegetation in both patches (mean ± one standard error 2012–2013 foliar cover, west: 47.4 ± 4.8%, east: 30.1 ± 3.7%), with Eriodictyon trichocalyx var. trichocalyx (west: 7.6 ± 1.2%, east: 16.1 ± 2.8%), and Eriogonum fasciculatum var. foliolosum (west: 5.9 ± 1.6%, east: 4.1 ± 1.5%) the next most abundant native shrubs. Both areas also supported a high cover of non-native grasses in the understory (west: 27.6 ± 4.4%, east: 53.3 ± 5.3%). Bromus diandrus and B. madritensis L. subsp. rubens were equally represented in the west patch, and B. diandrus about twice as abundant as B. madritensis in the east patch. Non-native forbs constituted less than 5% of pre-fire cover in both patches, with bare ground common (22.4%).
Fig 1
Map of the study area at the Robert J. Bernard Field Station in Claremont, CA.
The filled squares represent the 12 study plots (100 m2 each, red for non-native removal treatment and blue for control).
Map of the study area at the Robert J. Bernard Field Station in Claremont, CA.
The filled squares represent the 12 study plots (100 m2 each, red for non-native removal treatment and blue for control).
Experiment and data collection
We established 12 plots spanning the west to east axis of the burned sage scrub area, each 10 m by 10 m. Plots were separated by a 5 m wide buffer. Eight plots were in the western and four plots in the eastern patch (Fig 1). Treatments were paired to control for the west to east pre-fire gradient in shrub and non-native grass cover. We randomly assigned one plot to each treatment within neighboring pairs, starting with the westmost boundary.For the first four years after fire (spring 2014–2017), non-native grasses and forbs were hand weeded from plots assigned to the removal treatment. Removal began each year in mid-January and continued until collection of cover data started in the last week of March. Once per week during that time period, a team of two to three volunteers with experience in basic identification of local weedy species spent 20 minutes per treated plot removing non-natives. All non-native species were included, except that Erodium spp. proved difficult to control and were therefore targeted less.We measured vegetative cover in all plots during years when treatments were applied (2014–2017), using point-intercept sampling on a grid. Nine transects were established from west to east in each plot, spaced at 1 m intervals from south to north. Along each transect, we sampled points at 0.5 m intervals (N = 162 per plot) between the last week in March and the first week in May. The identities of all species touching a straight edge held up and down from the soil surface were recorded for each sample point (foliar cover); as a result, total cover values can exceed 100%.Demographic data for A. californica were collected from 2014 to 2020 at an annual census between June 14 and June 30, early in summer drought. In the first census after fire (2014), we individually tagged and mapped the locations of all A. californica in 10 plots. For one control and one removal plot adjacent to each other (6 and 7), high numbers of plants precluded tagging all individuals. All plants in the western halves of these two plots were tagged in June 2014, and any surviving plants in the eastern halves in November 2014. Plants were treated as distinct individuals if their stem bases were separate where entering the soil. At every annual census, leaf condition of each plant was classified as either good, if leaves were still present and appeared to be photosynthetically active, or deciduous, if the plant had begun to shed leaves. We quantified size for all individuals as canopy volume, by recording plant height, plant diameter on the longest axis, and a second diameter perpendicular to the first.All seedlings also were assigned individual tag numbers, mapped, and measured in most years after 2014 (2016–2018, 2020). In 2019, we counted all seedlings and mapped, tagged, and measured up to 6 per plot. Subsampling was used to assess seedling density and sizes in 2015, because of high emergence rates. We searched for seedlings in 2015 using a grid of points spaced at 2 m intervals within each plot (N = 16). At each sample point, seedlings within a 0.5 m radius were counted, scored for condition, and marked with a twist tie loop help in place by a plastic fruit fork. The seedling closest to the sampling point was measured. If no seedlings were found within 0.5 m, we repeated the same protocol at a 1 m radius. We then systematically searched any plot where no seedlings were detected at the sampling points. If fewer than five seedlings were measured from the sampling grid, additional sizes were recorded until reaching at least N = 5 per plot, or until all seedlings were measured. All seedling density data were collected in the same June census used for adult plants.In 2015, we also marked seedlings in a subset of plots earlier in the spring to assess survivorship from germination to the June census. The two westernmost plots (one control, one removal) were surveyed between February 20 and March 1, and the next pair of plots between March 12 and April 3. A second control plot was added to the early April tagging, to increase sample sizes for control seedlings.In some cases, seedlings were missed and appeared as untagged individuals in the following year. We used size records from known seedlings and first year plants to create criteria for assigning untagged individuals to a recruitment cohort. Untagged plants that were either less than 20 cm in height or smaller than 3000 cm3 in canopy volume were assumed to be seedlings. This classification system correctly predicted 98.4% of records for confirmed seedlings (N = 128), and 97.0% of those for confirmed one-year old plants (N = 34).Artemisia californica likely resprout from aboveground organs [36], but definitively identifying resprouting individuals in the first year after fire would have required excavating roots [41]. We classified plants tagged in June 2014 as likely seedlings unless they exceeded the size thresholds used in other years or could be confirmed as resprouts by the presence of dead stems. Keeley and Keely (1984) report May sizes for sage scrub shrubs that resprouted after fire well over our criteria (e.g., mean heights of 50–52 cm), although these values were combined across species.
Data analysis
All statistical analyses were carried out in R (version 3.6.1). Treatment effects on non-native cover from 2014–2017 were quantified by first aggregating data for individual plots within each year. We determined the number of non-native species observed per sampling point, then calculated plot means for those values. A linear mixed effects regression in the package lme4 [45] was used to test for effects of treatment, time (year) since fire and their interaction, with plot as a random effect.Records of A. californica emergence and survival included two years from the time period after removal treatments ended (2018–2020). We treated all plots in these two years as controls. To make sure this decision did not drive any of our findings, we also ran the adult and seedling survival analyses with data only from the time period when experimental manipulations were carried out (2014–2017). None of the qualitative results for effects of precipitation, non-native removal, and plant size change when data from 2018–2019 are excluded.We aggregated seedling emergence data by plot and year, calculating seedling density as the total number of seedlings counted divided by the total area sampled. Transformation did not normalize these data, so effects of treatment and year were assessed with a linear model permutation test using the package lmPerm [46]. We compared the final distribution of A. californica plants in 2020 between plots that received control and removal treatments from 2014–2017, using a chi square goodness of fit test against the null hypothesis of a 50:50 ratio. Total A. californica canopy volume per plot in 2020 was then compared between removal and control treatments, using a t test after square root transformation.We also quantified effects of non-native removal and precipitation on survival of both adults and seedlings, with binomial generalized linear mixed effects models in lme4. We used total precipitation from January through June as the rainfall predictor because this window captures the winter and spring months when A. californica is most visibly leafed out and growing (National Climate Data Center records, Claremont, CA). Spring precipitation values were normalized as a proportional deviation from the mean for 2000–2020. Both seedling and adult models included plot as a random effect to account for repeated measures. Adult survival models also included individual plant identity as a random effect and log transformed canopy volume as a fixed effect. We selected a final best fit model for both adults and seedlings by comparing models including different combinations of fixed effects and interactions with likelihood ratio tests (LRT) (S2 Table). Significance of fixed effects was assessed with bootstrapped LRT in the R package pbkrtest, comparing the best fit model selected by AIC with one that either added or dropped the individual parameter [47]. In total, analyses included N = 1006 records for marked seedling survival and N = 2785 annual transitions for 461 adult plants.For the large spring 2015 cohort specifically, we compared June leaf condition between control and removal plants with a mixed binomial model including plot as a random effect. Estimates of early seeding survival (from March or April to June) drew on records from only one to two plots per treatment, so the frequency of survival was compared between control and removal with a Fisher’s Exact Test (March: N = 47 control, N = 26 removal.; April: N = 99 control, N = 47 removal).
Results
The removal treatment reduced cover of non-native grasses and forbs other than Erodium by nearly 60% relative to controls (t = -3.44, df = 42.8, p = 0.001), although with a marginal trend towards smaller effects over time (t = 1.63, df = 34.0, p = 0.11). Erodium cover did not differ between control and removal plots (t = 0.26, df = 9.99, p = 0.80), but increased after the first year until peaking at 15.3 ± 4.8% in 2016 (Table 1). Non-natives rapidly recolonized control plots, reaching mean foliar cover values of 49–60% by the third year after fire (Table 1). The Mediterranean annual grass Bromus madritensis dominated non-native cover, present at 17.4% to 42.3% of all sample points in control plots and 9.6% to 32.4% in removal plots from 2015–2017. Bromus diandrus also reached cover values over 10% in control plots during some years (S3 Table). The most common non-native forb after Erodium was Brassica nigra, observed at up to 11.4% of sampling points in control plots (2015).
Table 1
Mean and standard error (in parentheses) values for total foliar cover of non-native grasses and forbs in control compared to removal plots over the four treatment years.
Year
Control (SE)
Removal (SE)
Erodium (SE)
2014
18.0 (8.0)
1.9 (0.8)
3.2 (1.4)
2015
49.4 (11.1)
13.5 (3.6)
8.4 (3.3)
2016
60.2 (6.8)
32.2 (5.9)
15.3 (4.8)
2017
48.7 (6.7)
35.6 (7.1)
9.5 (2.5)
Cover for Erodium spp. is shown separately, averaged for both control and removal plots; removal treatments had no effect on Erodium cover.
Cover for Erodium spp. is shown separately, averaged for both control and removal plots; removal treatments had no effect on Erodium cover.Seven A. californica were confirmed as resprouts in June 2014, 3 in control and 4 in removal plots. The other 454 adult plants identified and tagged through 2019 recruited from seed (98.5% of total). Seedling emergence varied strongly spatially, but densities were far higher in the second growing year after fire (spring 2015) than any other year (Fig 2, p < 0.0001). In 2020, 54.3% of all adults were from the 2015 seedling cohort, compared to 19.1% from the 2019 cohort and 12.4% from the 2014 cohort. Non-native removal did not significantly change seedling emergence (Fig 2, p = 0.2).
Fig 2
Seedling densities as measured in June of each year, for control (black circles) and non-native removal (green triangles) treatments.
The lines represent annual mean values for control (black dotted) and removal (green solid). Seedling densities are shown on a square-root- transformed scale to reduce the distortion by outliers, but our statistical analysis used non-parametric permutation tests on the untransformed data. Control and non-native removal plots were sampled on equivalent dates each year; the series are offset to facilitate visual comparison.
Seedling densities as measured in June of each year, for control (black circles) and non-native removal (green triangles) treatments.
The lines represent annual mean values for control (black dotted) and removal (green solid). Seedling densities are shown on a square-root- transformed scale to reduce the distortion by outliers, but our statistical analysis used non-parametric permutation tests on the untransformed data. Control and non-native removal plots were sampled on equivalent dates each year; the series are offset to facilitate visual comparison.Both higher spring rainfall and larger plant size improved adult survival (Table 2, Fig 3). Non-native removal significantly increased survival of the smallest adults, but this effect disappeared with increasing plant size (Table 2). Removal treatments also strongly improved seedling survival (Table 2, Fig 4; S1 Table). Higher spring rainfall both benefitted seedlings directly and enhanced the effects of non-native removal (Table 2). For the largest cohort in 2015, early seedling survival did not differ between control and removal plots for plants tagged either in March (control: 70.2%, removal: 76.9%, p = 0.59) or in April (control: 93.9%, removal: 87.2%, p = 0.20). However, non-native removal significantly reduced the likelihood that seedlings had begun to drop leaves by June (control: 26.3%, removal 9.0%, LR = 6.2, p = 0.013).
Table 2
Results of best-fit generalized linear mixed effects models for the effects of January to June rainfall and non-native removal on adult (top) and seedling (bottom) survival.
Life stage
Factor
Estimate
SE
LR
p
Adult
Spring rainfall
1.63
0.86
6.08
0.015
Exotic removal
0.82
1.23
16.9
0.001
Log canopy volume
1.04
0.29
51.6
0.001
Removal x Log canopy volume
-0.65
0.31
8.4
0.008
Seedling
Spring rainfall
0.87
0.9
18.61
0.001
Exotic removal
1.78
0.74
12.51
0.006
Removal x rainfall
2.08
1.2
5.62
0.023
The adult survival model included plant size, as measured by log canopy volume. Columns give the coefficient estimate and standard error determined from 1000 bootstrap replicates, as well as the likelihood ratio test statistic (LR) and p value. Bolded p values indicate results significant at a threshold of < 0.05.
Fig 3
Effects of non-native plant removal on the probability of survival in adults of different sizes.
The three panels show predictions for spring rainfall at (A) the study minimum (60% below the 1999–2020 average), (B) the study median (28% below average), and (C) the study maximum (70% above average). Center lines for each group show best-fit predictions, and the filled areas 95% confidence bounds based on bootstrap replicates.
Fig 4
Effects of spring rainfall (mm) on the probability of seedling survival, by treatment (black dashed line indicates control, green solid line non-native grass removal).
Center lines for each group represent best-fit predictions, and the filled areas 95% confidence bounds based on bootstrap replicates. Seedling survival could only be compared between control and removal treatments in years with less than 300 mm of rainfall (S1 Table). So, we did not extrapolate the model predictions for non-native removal beyond 300 mm of precipitation”.
Effects of non-native plant removal on the probability of survival in adults of different sizes.
The three panels show predictions for spring rainfall at (A) the study minimum (60% below the 1999–2020 average), (B) the study median (28% below average), and (C) the study maximum (70% above average). Center lines for each group show best-fit predictions, and the filled areas 95% confidence bounds based on bootstrap replicates.
Effects of spring rainfall (mm) on the probability of seedling survival, by treatment (black dashed line indicates control, green solid line non-native grass removal).
Center lines for each group represent best-fit predictions, and the filled areas 95% confidence bounds based on bootstrap replicates. Seedling survival could only be compared between control and removal treatments in years with less than 300 mm of rainfall (S1 Table). So, we did not extrapolate the model predictions for non-native removal beyond 300 mm of precipitation”.The adult survival model included plant size, as measured by log canopy volume. Columns give the coefficient estimate and standard error determined from 1000 bootstrap replicates, as well as the likelihood ratio test statistic (LR) and p value. Bolded p values indicate results significant at a threshold of < 0.05.Adult plants were almost exactly divided between control and removal plots from 2014–2015 (S4 Table). This ratio shifted sharply after the 2015 seedling cohort established, with 79.6% of adults found in removal plots by 2016. The same pattern held through 2020, three years after the removal treatment ended (81.9% adults in removal plots, Χ = 42.8, df = 1, p < 0.0001). By 2020, mean A. californica canopy area per plot was nearly four times greater where non-natives had been removed from 2014–2017 (t = 2.77, df = 10, p = 0.02; control: 20.2 ± 13.0 m2; removal: 79.7 ±26.6 m2).
Discussion
Preventing type conversion after disturbance is an important management goal in southern California, particularly given substantial barriers to restoration once habitats are dominated by non-native plants [5, 21, 29, 48]. We found that non-native grass and forb removal in the first four years after a sage scrub fire facilitated recovery of the dominant native shrub by increasing survival of seedlings and small adults. We also observed strong effects of precipitation, with higher rainfall directly benefitting seedling and adult survival as well as strengthening positive responses to non-native removal by seedlings. Our results illustrate how an increasing probability of drought in the critical first few years after fire could create additional obstacles to post-fire shrub recovery in sage scrub.Both observational and experimental studies support negative effects of non-native grasses and forbs on native seedling recruitment in sage scrub. Artemisia californica seedlings are largely absent from non-native annual grasslands in southern California, occurring primarily in vegetation gaps within intact sage scrub [13, 24, 25, 49]. Suppression of native seed germination by non-native grass thatch from the previous growing season is one potential cause. This effect has been shown experimentally for native forbs in competition with non-native annual grasses such as B. diandrus [20, 50], but not tested directly for A. californica. Germination of A. californica is inhibited in the dark, suggesting that heavy grass thatch could suppress seedling emergence [51]. Nevertheless, we did not find significant effects of non-native removal on seedling emergence. One important caveat is that our data provided low power to test for such effects because of high variation in seedling emergence among a modest number of plots and years (Fig 2). Non-native grass and forb foliar cover totaled only 17% in control plots during the first growing year after fire, so thatch levels were in any case low in fall of 2014 when most A. californica seedlings emerged (Table 1).Post-germination competition with non-native annuals can also limit A. californica recruitment, a mechanism our results support. One previous study found that increasing grass density reduced A. californica seedling survival [21], while another observed complete seedling mortality unless non-native grasses were removed [52]. Our results document significant benefits of non-native removal for over-summer seedling survival. Seedlings in control plots were more likely to show evidence of water stress by onset of summer drought, consistent with previous findings that non-native grasses suppress native shrub seedlings by changing soil water availability [21]. Similarly, small adults benefitted from non-native removal, although as in other studies this effect disappeared for larger individuals [21]. At the largest sizes, survival estimates for plants in control conditions appear somewhat higher than for plants in removal conditions (Fig 3). However, many records for large plants were from individuals in removal plots observed after 2017, when removal treatments stopped (70.8% of adult survival data for plants with a log canopy volume greater than 5 log [m3]).Wet years likewise increased both seedling and adult survival. Previous studies similarly document the importance of water limitation to native shrub communities. Drought years have been associated with reduced post-fire survival of resprouts in chaparral [30] and low native shrub seedling recruitment in sage scrub [24]. Experimentally lowering water availability depressed growth rates and increased allocation to below-ground biomass in A. californica seedlings [53]. Detrimental effects of water stress on adult shrubs in chaparral and sage scrub can reduce canopy cover and increase vulnerability to invasion [26, 54], as well as slow post-fire regeneration of cover [34]. Still, no previous study has to our knowledge quantified demographic effects of drought in southern California across shrub life history stages, a critical step for modeling post-fire recovery. Two patterns from our data seem particularly noteworthy. First, rainfall effects on seedling survival interacted strongly with non-native removal, suggesting little benefit of wet years for recruitment unless non-native competitors are controlled. Second, larger plants survived better than smaller, more recently established ones but still experienced substantial mortality in dry years (Table 2, Fig 1).The balance between resprouting and recruitment from seed plays a critical role in shaping post-fire recovery of chaparral and sage scrub. Sage scrub is considered resilient and can quickly recover to pre-fire composition, so long as enough individuals crown resprout [5]. Yet resprouting rates vary dramatically across habitats and individual shrub species [43]. We observed a surprisingly small number of resprouts given the high pre-fire cover of A. californica (N = 7 across 0.12 hectares). While A. californica resprouts after fire less than other sage scrub shrubs such as Salvia apiana and Eriodictyon trichocalyx var. trichocalyx, rates documented in other studies still range from 13% to 25% [42, 43]. Using size to classify resprouts rather than excavating around roots may have led to underestimation of resprouts in our data. Still, individuals present in the first year after fire represented only a small fraction of those remaining in 2020 (12.4%). This 2013–2014 cohort experienced poor survival and contributed little to A. californica cover in our plots.Several factors could explain low resprouting rates at this site. First, shrubs may resprout less at inland locations compared to coastal ones, potentially because of past selection due to lower historical fire frequencies in drier habitats [55]. This pattern is not consistently supported for A. californica by previous studies, however [42]. Stand age is another potential explanation, as older individuals likely lose their capacity to resprout [33, 43]. We do not know the age distribution of A. californica before the 2013 fire, but no other major disturbance events likely to generate stand replacement had occurred in at least 60 years. Previous work argues that most populations of A. californica are relatively even-aged, although assigning ages to individual plants is complicated by their tendency to resprout even in the absence of fire [56]. Regardless of the cause, our data reinforce that crown resprouting is highly variable across sites.When resprouting rates are low, native shrub recovery hinges on recruitment from seed. This may make communities more vulnerable to increased fire frequency if time between fires is insufficient for seedbanks to replenish [48]. Since seedlings are particularly vulnerable to effects of non-native species, experience high mortality (Fig 3), and contribute less cover, communities with low resprouting rates are likely more vulnerable to type conversion after fire [57]. Pulse seedling recruitment is common for obligate seeding shrubs in both chaparral and sage scrub, typically in the first year after fire [42]. Strong reproduction by resprouting shrubs immediately after fire can also lead to a large peak in germination during the second year [41, 58]. Given limited resprouting, high seedling emergence in the second post-fire year at our site likely came from the seed bank. Dispersal beyond the vicinity of a parent shrub is thought unusual in most sage scrub species [5]. In contrast to well-studied, larger-scale fires, unburned habitat remained within a few hundred meters of all our plots (Fig 1). Still, the large drop in seedling emergence after 2014–2015 suggests depletion of the seed source; 90% of the seedling density in our plots concentrated into this single growing year. Keeley [40] found that more than 80% of A. californica seedlings recorded within 5 years after fire emerged in the first two years.Delay of seedling emergence into the second year at our site may have been caused by low rainfall immediately after the fire occurred, in the window between November and January when most germination takes place. Precipitation during these months in 2013–2014 was 72% below the mean for 1999–2020 (2013–2014: 58.41 mm; 1999–2020 mean: 207.2 ± 38.9 mm). During the following growing year when A. californica emerged in large numbers, early rainfall increased by more than three times (177.5 mm). In general, early rain likely to stimulate germination (November through January) correlates with January to June precipitation linked to seedling and adult survival (1999–2020, r = 0.75, df = 20, p < 0.0001). This creates potential for germination cuing to reduce the risk of emergence into a low survival year. Still, our data show that early precipitation is not always a reliable cue of spring conditions. The small number of seedlings germinating in 2013–2014 benefitted from much higher January to June rainfall (277.8 mm; S3 Table) than the large number of seedlings germinating in 2014–2015 (114.7 mm).In summary, our findings reinforce the high vulnerability of sage scrub to post-fire loss of shrub cover and potential type conversion, particularly with increasing drought frequency and in stands with low rates of crown resprouting. Yet they also illustrate the potential for targeted management of non-natives immediately after fire to promote recovery of native shrubs. Caution in extrapolating these results is important, given the small scale of both the study area and fire. At the same time, most remaining intact sage scrub habitat in southern California consists of small fragments bordering on urban and suburban development [39, 40]. Our work helps fill spatial gaps in previous research, given other studies have mostly concentrated in coastal areas or further east near the boundary of sage scrub distributions. Additional studies in similar small fragments will help improve our understanding of the diverse successional pathways in these threatened shrub communities.
Survival of seedlings from when they were marked in June of their first year to the following June.
Columns show the year of germination (spring); total spring rainfall (January to June) for that year (mm); treatment (control or non-native removal); the number of marked seedlings that survived; the number of marked seedlings that died; and the percent survival. Percent survival values were only calculated for years in which at least 20 seedlings were tagged for each treatment. Non-native removal treatments were applied in 2014–2017 but not 2018–2019. For 2018–2019, we denoted treatment as “None”, with the original treatment assignment in parentheses (C = control, R = removal).(DOCX)Click here for additional data file.
Summary of all mixed effects survivorship models tested, for both seedlings and adults.
The model specifications are given, with fixed effect predictors abbreviated as Size (log transformed canopy volume in m3), Treat (removal or control treatment), and Rain (rainfall from January through June). Random effects for Plot were included in all models, and random effects for Tag (individual plant identity) in the adult models. Terms that changed from the best-fit model are shown (- for removed, + for added), along with the AIC values.(DOCX)Click here for additional data file.
Percent cover of the most common non-native species in control and removal plots for the first four spring surveys (late March to April) after the October 2013 fire.
The most common non-native forb Erodium spp. was unaffected by the removal treatment and is not included; cover values for Erodium can be found in Table 1. Species are annotated by life history (G = non-native annual grasses, F = non-native annual forbs, SS = non-native subshrubs).(DOCX)Click here for additional data file.
Distribution of established (greater than one year old) Artemisia californica between control and removal plots over the 7 post-fire study years.
Numbers for each treatment represent the total plants across all plots in that treatment (N = 6). The percent of plants in removal plots is also shown for each year.(DOCX)Click here for additional data file.27 Apr 2021PONE-D-21-06311Non-native plant removal and high rainfall years promote post-fire recovery of Artemisia californicain southern California sage scrubPLOS ONEDear Dr. Thomson,Thank you for submitting your manuscript to PLOS ONE. After careful consideration, we feel that it has merit but does not fully meet PLOS ONE’s publication criteria as it currently stands. Therefore, we invite you to submit a revised version of the manuscript that addresses the points raised during the review process (see below).Please submit your revised manuscript by Jun 11 2021 11:59PM. If you will need more time than this to complete your revisions, please reply to this message or contact the journal office at plosone@plos.org. When you're ready to submit your revision, log on to https://www.editorialmanager.com/pone/ and select the 'Submissions Needing Revision' folder to locate your manuscript file.Please include the following items when submitting your revised manuscript:A rebuttal letter that responds to each point raised by the academic editor and reviewer(s). You should upload this letter as a separate file labeled 'Response to Reviewers'.A marked-up copy of your manuscript that highlights changes made to the original version. You should upload this as a separate file labeled 'Revised Manuscript with Track Changes'.An unmarked version of your revised paper without tracked changes. You should upload this as a separate file labeled 'Manuscript'.If you would like to make changes to your financial disclosure, please include your updated statement in your cover letter. Guidelines for resubmitting your figure files are available below the reviewer comments at the end of this letter.If applicable, we recommend that you deposit your laboratory protocols in protocols.io to enhance the reproducibility of your results. Protocols.io assigns your protocol its own identifier (DOI) so that it can be cited independently in the future. For instructions see: http://journals.plos.org/plosone/s/submission-guidelines#loc-laboratory-protocols. Additionally, PLOS ONE offers an option for publishing peer-reviewed Lab Protocol articles, which describe protocols hosted on protocols.io. Read more information on sharing protocols at https://plos.org/protocols?utm_medium=editorial-email&utm_source=authorletters&utm_campaign=protocols.We look forward to receiving your revised manuscript.Kind regards,Cristina ArmasAcademic EditorPLOS ONEJournal Requirements:When submitting your revision, we need you to address these additional requirements.Please ensure that your manuscript meets PLOS ONE's style requirements, including those for file naming. The PLOS ONE style templates can be found athttps://journals.plos.org/plosone/s/file?id=wjVg/PLOSOne_formatting_sample_main_body.pdf andhttps://journals.plos.org/plosone/s/file?id=ba62/PLOSOne_formatting_sample_title_authors_affiliations.pdfWe note that you have stated that you will provide repository information for your data at acceptance. Should your manuscript be accepted for publication, we will hold it until you provide the relevant accession numbers or DOIs necessary to access your data. If you wish to make changes to your Data Availability statement, please describe these changes in your cover letter and we will update your Data Availability statement to reflect the information you provide.We note that Figure 1 in your submission contain satellite images which may be copyrighted. All PLOS content is published under the Creative Commons Attribution License (CC BY 4.0), which means that the manuscript, images, and Supporting Information files will be freely available online, and any third party is permitted to access, download, copy, distribute, and use these materials in any way, even commercially, with proper attribution. For these reasons, we cannot publish previously copyrighted maps or satellite images created using proprietary data, such as Google software (Google Maps, Street View, and Earth). For more information, see our copyright guidelines: http://journals.plos.org/plosone/s/licenses-and-copyright.We require you to either (1) present written permission from the copyright holder to publish these figures specifically under the CC BY 4.0 license, or (2) remove the figures from your submission:1. You may seek permission from the original copyright holder of Figure 1 to publish the content specifically under the CC BY 4.0 license.We recommend that you contact the original copyright holder with the Content Permission Form (http://journals.plos.org/plosone/s/file?id=7c09/content-permission-form.pdf) and the following text:“I request permission for the open-access journal PLOS ONE to publish XXX under the Creative Commons Attribution License (CCAL) CC BY 4.0 (http://creativecommons.org/licenses/by/4.0/). Please be aware that this license allows unrestricted use and distribution, even commercially, by third parties. Please reply and provide explicit written permission to publish XXX under a CC BY license and complete the attached form.”Please upload the completed Content Permission Form or other proof of granted permissions as an "Other" file with your submission.In the figure caption of the copyrighted figure, please include the following text: “Reprinted from [ref] under a CC BY license, with permission from [name of publisher], original copyright [original copyright year].”2. If you are unable to obtain permission from the original copyright holder to publish these figures under the CC BY 4.0 license or if the copyright holder’s requirements are incompatible with the CC BY 4.0 license, please either i) remove the figure or ii) supply a replacement figure that complies with the CC BY 4.0 license. Please check copyright information on all replacement figures and update the figure caption with source information. If applicable, please specify in the figure caption text when a figure is similar but not identical to the original image and is therefore for illustrative purposes only.The following resources for replacing copyrighted map figures may be helpful:USGS National Map Viewer (public domain): http://viewer.nationalmap.gov/viewer/The Gateway to Astronaut Photography of Earth (public domain): http://eol.jsc.nasa.gov/sseop/clickmap/Maps at the CIA (public domain): https://www.cia.gov/library/publications/the-world-factbook/index.html and https://www.cia.gov/library/publications/cia-maps-publications/index.htmlNASA Earth Observatory (public domain): http://earthobservatory.nasa.gov/Landsat: http://landsat.visibleearth.nasa.gov/USGS EROS (Earth Resources Observatory and Science (EROS) Center) (public domain): http://eros.usgs.gov/#Natural Earth (public domain): http://www.naturalearthdata.com/4. Please review your reference list to ensure that it is complete and correct. If you have cited papers that have been retracted, please include the rationale for doing so in the manuscript text, or remove these references and replace them with relevant current references. Any changes to the reference list should be mentioned in the rebuttal letter that accompanies your revised manuscript. If you need to cite a retracted article, indicate the article’s retracted status in the References list and also include a citation and full reference for the retraction notice.Additional Editor Comments :Comments besides those from reviewers:Please add line numbers in your next manuscript versionPage 4. “non-native effects on fire”. Too vague. Something is missing here. On what fire characteristic do non-native plant species had an effect?“Once type conversion occurs, subsequently restoring shrubs through management has proved very difficult” Please simplify this sentence. Moreover, “type conversion” or “type-converted” are terms/concepts/processes used in many instances in the introduction and in the concluding paragraph in the discussion. Thus, please be precise and define it.I agree with one of the reviewers, the International System of Units (SI) suggests using mm instead of cm for precipitationPage 17. Please, rephrase this sentence in a more formal tone. There are no control or removal plants. “At the largest sizes, survival estimates for control plants appear somewhat higher than for removal plants (Fig 3).” The same occurred for the next sentence in that paragraph.FiguresFig. 2 displays the square-root transformed densities of seedlings across years and exotic-removal treatments. What is the purpose of this transformation of the data? In page 11 (data analysis) authors state that transformation of the data did not normalize the distribution of errors.Legend of Fig. 4. Please rephrase the last sentence, unclear “Removal treatments were not observed under conditions of greater than 300 mm spring rainfall, so model results are not extrapolated to higher values.”[Note: HTML markup is below. Please do not edit.]Reviewers' comments:Reviewer's Responses to QuestionsComments to the Author1. Is the manuscript technically sound, and do the data support the conclusions?The manuscript must describe a technically sound piece of scientific research with data that supports the conclusions. Experiments must have been conducted rigorously, with appropriate controls, replication, and sample sizes. The conclusions must be drawn appropriately based on the data presented.Reviewer #1: YesReviewer #2: Partly**********2. Has the statistical analysis been performed appropriately and rigorously?Reviewer #1: YesReviewer #2: Yes**********3. Have the authors made all data underlying the findings in their manuscript fully available?The PLOS Data policy requires authors to make all data underlying the findings described in their manuscript fully available without restriction, with rare exception (please refer to the Data Availability Statement in the manuscript PDF file). The data should be provided as part of the manuscript or its supporting information, or deposited to a public repository. For example, in addition to summary statistics, the data points behind means, medians and variance measures should be available. If there are restrictions on publicly sharing data—e.g. participant privacy or use of data from a third party—those must be specified.Reviewer #1: YesReviewer #2: Yes**********4. Is the manuscript presented in an intelligible fashion and written in standard English?PLOS ONE does not copyedit accepted manuscripts, so the language in submitted articles must be clear, correct, and unambiguous. Any typographical or grammatical errors should be corrected at revision, so please note any specific errors here.Reviewer #1: YesReviewer #2: Yes**********5. Review Comments to the AuthorPlease use the space provided to explain your answers to the questions above. You may also include additional comments for the author, including concerns about dual publication, research ethics, or publication ethics. (Please upload your review as an attachment if it exceeds 20,000 characters)Reviewer #1: In this manuscript, entitled “Non-native plant removal and high rainfall years promote post-fire recovery of Artemisia californica in southern California sage scrub”, authors conducted a long term experimental field study to evaluate the post-fire recovery of the shrub A. californica in a Mediterranean plant community in California, by assessing the interactive effects of non-native plants removal and wet/dry years on A. californica seedling emergence, recruitment, and adult plants survival and growth. The manuscript is well written. The experiment is well designed and the results are sound based on data analysis and support the conclusions. Overall, the manuscript introduces novel evidence that underlines adequate practices (i.e., invasive species removal) in post-fire management to avoid excessive habitat loss by non-native plant invasion under foreseen climatic drought scenarios. Comments below may help to improve the manuscript:In “study system” section, authors provide data on annual rainfall (September to August) in cm (i.e., 41.55 cm per year), while a couple of sentences below they provide data in mm (i.e., 41.8 mm). I guess there is a typo somewhere. Afterwards, they refer to table S1, where they provide data on Spring rainfall as mm. However, data scales to hundreds of mm. Please, revise the data and use always the same units (preferably in mm). Also, please, correct the typo with data period in …complete records; 1986-197? NOOA…I find that Fig 1 would be more informative by directly including plots and treatments distribution.In table 2, does “exotic removal” refers to “non-native removal” treatment? Please clarify terminology.Reviewer #2: The manuscript ref PONE-D-21-06311 describes the monitoring of post-fire seedling emergence and survival as affected by non-native grasses and forbs removal during a 7-yr period. It relates the probability of survival of both seedlings and adults of Artemisia californica with spring rainfall and plant size in a yearly basis, with interesting results. The main problem I see is the lack of replication as the whole study is conducted in 12 plots at the same location which reduces the transferability of results. In fact, authors state that “Seedling emergence varied strongly spatially” (Page 14) and “Caution in extrapolating these results is important, given the small scale of both the study area and fire” (Page 20). However, the large monitoring period provides extra value to the study.Introduction section is well-structured and clearly establishes the objectives and hypotheses.- It is not clear to me whether the ratio resprouting/recruitment of A.californica is higher inland or in coastal areas as I do not know how is the moisture gradient (I can guess but better if it is explicitly stated). Page 6- Table S1, what does ‘none’ mean in the treatment column for years 2018 and 2019? It appears in Page 11 but should be define in the table heading (or in a footnote)- Did you monitor the biomass of the non-natives removed? This could be used as covariable.- Were those two overdense plots that precluded tagging all individuals paired? Page 9- "The two westernmost plots (one control, one removal) were surveyed between February 20 and March 1, and the next pair of plots between March 12 and April 3. A second control plot was added to the early April tagging, to increase sample sizes for control seedlings". Recruitment in the latest sampled control plot was ca. 5 weeks after the earliest which can result in overestimation.- "but resulting drops in sample size and the loss of observations from the very wet spring of 2019 would have reduced statistical significance". This is speculative- Not sure canopy volume can be considered fixed factorTable 1. Remove SE columns and include the numbers in brackets (or after ±) next to the mean value. Include different letters for significant differences between control and removal treatments- "Seven A. californica were confirmed as resprouts in June 2014, 3 in control and 4 in removal plots. Almost all individuals established after the fire from seedlings". What was the percentage of germinated seedlings? These numbers are more precise than ‘almost all’.Figure 2. It seems that density of recruits in removal plots were recorded later than in control plots. Correct it- "absence of non-native grass removal", in page 17, sounds weird these two (three in fact) negations in the sentence**********6. PLOS authors have the option to publish the peer review history of their article (what does this mean?). If published, this will include your full peer review and any attached files.If you choose “no”, your identity will remain anonymous but your review may still be made public.Do you want your identity to be public for this peer review? For information about this choice, including consent withdrawal, please see our Privacy Policy.Reviewer #1: NoReviewer #2: No[NOTE: If reviewer comments were submitted as an attachment file, they will be attached to this email and accessible via the submission site. Please log into your account, locate the manuscript record, and check for the action link "View Attachments". If this link does not appear, there are no attachment files.]While revising your submission, please upload your figure files to the Preflight Analysis and Conversion Engine (PACE) digital diagnostic tool, https://pacev2.apexcovantage.com/. PACE helps ensure that figures meet PLOS requirements. To use PACE, you must first register as a user. Registration is free. Then, login and navigate to the UPLOAD tab, where you will find detailed instructions on how to use the tool. If you encounter any issues or have any questions when using PACE, please email PLOS at figures@plos.org. Please note that Supporting Information files do not need this step.10 Jun 2021Dear Dr. Armas,I am writing to submit a revised version of the MS, “Non-native plant removal and high rainfall years promote post-fire recovery of Artemisia californica in southern California sage scrub”. My co-authors and I greatly appreciated the thoughtful comments from you and the reviewers. I believe the revisions address all of the questions and recommendations, but if the changes did not fully hit the mark in some respects please let me know.I have condensed the original text of the reviews in places to facilitate reading; paragraphs of text with our responses are all denoted with a – mark.Thank you again for your time helping us with this MS,Diane ThomsonJournal Requirements:When submitting your revision, we need you to address these additional requirements.1. Please ensure that your manuscript meets PLOS ONE's style requirements, including those for file naming.- We rechecked that style requirements were met and now use the correct format for file names as well as Figure and Table legends. We also removed the Author Contributions section from the MS body, as it seemed this was not meant to be included there. Please advise if we have misinterpreted any details of the instructions; this is our first submission to PLOS, and we are still learning the formatting.2. We note that you have stated that you will provide repository information for your data at acceptance. Should your manuscript be accepted for publication, we will hold it until you provide the relevant accession numbers or DOIs necessary to access your data. If you wish to make changes to your Data Availability statement, please describe these changes in your cover letter and we will update your Data Availability statement to reflect the information you provide.- We are working on the Dryad repository for this data set, and will complete it immediately if the MS is accepted for publication.3. We note that Figure 1 in your submission contain satellite images which may be copyrighted.We require you to either (1) present written permission from the copyright holder to publish these figures specifically under the CC BY 4.0 license, or (2) remove the figures from your submission:- We have removed the images and remade Figure 1.4. Please review your reference list to ensure that it is complete and correct.- We checked the references carefully. We found and corrected some minor typos and errors in the references, but made no major change. Note that we corrected errors in the references after accepting the other changes to the MS, so they do not appear in the Track Changes version of the revision.Additional Editor Comments:Please add line numbers in your next manuscript version- We added continuous line numbers.Page 4. “non-native effects on fire”. Too vague. Something is missing here. On what fire characteristic do non-native plant species had an effect?- We have revised this sentence to connect more explicitly with the section on page 2 where potential effects of non-native plants on fire are discussed in more detail (revision page 4, lines 66-70).“Increased fire frequency is associated with higher non-native grass and forb cover in both chaparral and sage scrub [14,16-18]. Yet, the directions of any causal relationships remain unclear; these correlations could reflect the effects of fire disturbance on non-native plants, the effects of non-native plants on ignition risk and fuel loads, or both.”“Once type conversion occurs, subsequently restoring shrubs through management has proved very difficult” Please simplify this sentence. Moreover, “type conversion” or “type-converted” are terms/concepts/processes used in many instances in the introduction and in the concluding paragraph in the discussion. Thus, please be precise and define it.- We simplified the sentence as requested. Page 4, lines 78-79.“Reintroducing shrubs to restore type-converted, non-native grasslands has proved very difficult [5,29].”- Yes, the term type conversion is introduced early on in the introduction (the second paragraph) and then used regularly. We expanded on the definition given with this first use of the term. Page 2, lines 56-59.“These changes in fire regime can lead to habitat degradation and even type conversion, meaning a major and persistent shift in community structure such as replacement of native shrubland by non-native grassland [12].”I agree with one of the reviewers, the International System of Units (SI) suggests using mm instead of cm for precipitation- We changed all precipitation measures to mm, as requested.Page 17. Please, rephrase this sentence in a more formal tone. There are no control or removal plants. “At the largest sizes, survival estimates for control plants appear somewhat higher than for removal plants (Fig 3).” The same occurred for the next sentence in that paragraph.- We rephrased as follows (revised MS, page 17 lines 369-373). Please advise if this wording does not fully address the comment.“At the largest sizes, survival estimates for plants in control conditions appear somewhat higher than for plants in removal conditions (Fig 3). However, many records for large plants were from individuals in removal plots observed after 2017, when removal treatments stopped (70.8% of adult survival data for plants with a log canopy volume greater than 5 log [m3]).”FiguresFig. 2 displays the square-root transformed densities of seedlings across years and exotic-removal treatments. What is the purpose of this transformation of the data? In page 11 (data analysis) authors state that transformation of the data did not normalize the distribution of errors.- The purpose of the transformation is to facilitate interpretation of the data. In a plot of the untransformed data, the tail of unusual high values pulls the y axis out at the top and leaves all the other points concentrated near the bottom. The permutation methods we used for the formal statistical analysis have effects similar to transformation, in that they reduce the influence of unusual values. In that sense, the plot of transformed data is a better way of visualizing the analysis we performed than a plot of the untransformed data would be. We have added a sentence to the legend explaining why the data are presented on a square-root- transformed scale, even though the statistical analysis used the untransformed data (revised MS page 14, lines 296-298).“Seedling densities are shown on a square-root-transformed scale to reduce the distortion by outliers, but our statistical analysis used non-parametric permutation tests on the untransformed data.”Legend of Fig. 4. Please rephrase the last sentence, unclear “Removal treatments were not observed under conditions of greater than 300 mm spring rainfall, so model results are not extrapolated to higher values.”- We revised this part of the legend. Please advise if the new wording is not clear enough. Revised MS page 15, lines 326-328.“Seedling survival could only be compared between control and removal treatments in years with less than 300 mm of rainfall (Table S1). So, we did not extrapolate the model predictions for non-native removal beyond 300 mm of precipitation.”Reviewer #1:In “study system” section, authors provide data on annual rainfall (September to August) in cm (i.e., 41.55 cm per year), while a couple of sentences below they provide data in mm (i.e., 41.8 mm). I guess there is a typo somewhere. Afterwards, they refer to table S1, where they provide data on Spring rainfall as mm. However, data scales to hundreds of mm. Please, revise the data and use always the same units (preferably in mm). Also, please, correct the typo with data period in …complete records; 1986-197? NOOA…- Thanks for catching the error; the original MS intended to report both the precipitation values cited above in cm. We have changed all precipitation measures to units of mm. The typo in the year range for complete rainfall records has also been corrected.I find that Fig 1 would be more informative by directly including plots and treatments distribution.- The figure has been modified (also requested by the editor) and now shows individual plots and treatments.In table 2, does “exotic removal” refers to “non-native removal” treatment? Please clarify terminology.- Yes. We changed Table 2 so it consistently uses “non-native removal” throughout.Reviewer #2: The main problem I see is the lack of replication as the whole study is conducted in 12 plots at the same location which reduces the transferability of results. In fact, authors state that “Seedling emergence varied strongly spatially” (Page 14) and “Caution in extrapolating these results is important, given the small scale of both the study area and fire” (Page 20). However, the large monitoring period provides extra value to the study.- We agree that our study is local in scale; that is why we included the caveats cited above. But many if not most ecological studies are limited to a single site, particularly research on uncontrolled disturbance events like fire. Almost all the experimental studies and many of the post-fire studies this MS cites are from a single site. Both the level of replication (plots/treatment) and the spatial dimensions of sampling in our work are just as or even more extensive than in these comparable studies (detailed examples can be found at the end of this response letter). Even the results of multi-site research are not necessarily broadly transferable. Arguably, one of the best ways to build a generalized understanding is through synthesizing many local studies like this one. We have added another sentence noting the value of our work in filling geographic gaps relative to the distribution of previous studies across sage scrub habitats in southern California. (Revised MS page 20 lines 443-445).“Our work helps fill spatial gaps in previous research, given other studies have mostly concentrated in coastal areas or further east near the boundary of sage scrub distributions.”It is not clear to me whether the ratio resprouting/recruitment of A.californica is higher inland or in coastal areas as I do not know how is the moisture gradient (I can guess but better if it is explicitly stated). Page 6- We added some language to clarify this point. Revised MS, page 6 lines 123-125.“Several studies suggest that the ratio of resprouting to recruitment from seed in sage scrub declines across a moisture gradient, from more mesic coastal to drier inland habitat [33,43].”Table S1, what does ‘none’ mean in the treatment column for years 2018 and 2019? It appears in Page 11 but should be define in the table heading (or in a footnote)- The non-native removal treatment was applied in 2014-2017, but not during the final two years of data collection in 2018-2019. We added an explanation to the S1 Table legend. (MS lines 456-458).“Non-native removal treatments were applied in 2014-2017 but not 2018-2019. For 2018-2019, we denoted treatment as “None”, with the original treatment assignment in parentheses (C=control, R=removal).”Did you monitor the biomass of the non-natives removed? This could be used as covariable.- No, we did not measure the biomass of the non-natives removed; this would have been time consuming given the size of our plots (each 100 m2). Our analyses show clear differences in seedling and adult survival between control and removal plots. Adding covariates might have accounted for some additional background variation/noise, but it seems to us very unlikely this would change our core findings.- Were those two overdense plots that precluded tagging all individuals paired?- Yes. We clarified this point in the text. Page 9 lines 188-190.“For one control and one removal plot adjacent to each other (6 and 7)…”"The two westernmost plots (one control, one removal) were surveyed between February 20 and March 1, and the next pair of plots between March 12 and April 3. A second control plot was added to the early April tagging, to increase sample sizes for control seedlings". Recruitment in the latest sampled control -plot was ca. 5 weeks after the earliest which can result in overestimation.- We want to make sure there is no ambiguity about when the recruitment data (seedling densities, as reported in Fig. 2) were collected: at the same time for all plots and years, during the same June census used for adult plants (page 8, lines 188-189). The passage quoted above refers to a different component of our work. We marked a subset of seedlings present in March and April of 2015 to test for treatment effects on early-season survival (between emergence and the June census). To better clarify this distinction, we added a sentence to the section on seedling data collection (page 10, lines 207-208). We also separated the information on early-season seedling tagging in 2015 into a different paragraph, starting on line 209.“All seedling density data were collected in the same June census used for adult plants.”- Returning to the effects of tagging seedlings in two different months: yes, the observed survival was lower for plants we tagged earlier. That is why we reported the results separately for seedlings tagged in March and April (see page 14, lines 305-308, copied in below). The key point is that the control and removal seedlings whose survival we compared were tagged at the same time, for both the March and April cohorts.“For the largest cohort in 2015, early seedling survival did not differ between control and removal plots for plants tagged either in March (control: 70.2%, removal: 76.9%, p = 0.59) or in April (control: 93.9%, removal: 87.2%, p = 0.20).”- "but resulting drops in sample size and the loss of observations from the very wet spring of 2019 would have reduced statistical significance". This is speculative.- We should have been more explicit in the original MS and have revised to clarify. We re-ran both the adult and seedling survival analyses without the 2018-2019 data, to make sure that including those years (after the experimental manipulations had ended) did not drive our findings. The p values do change slightly, of course, but none of the main results are different. The key point is that our findings are robust to classifying the 2018-2019 records as “control” (no non-native removal) data points. We simplified the text to say that and nothing else. Revised MS, page 11, lines 236-239.- “To make sure this decision did not drive any of our findings, we also ran the adult and seedling survival analyses with data only from the time period when experimental manipulations were carried out (2014-2017). None of the qualitative results for effects of precipitation, non-native removal, and plant size change when data from 2018-2019 are excluded.”Not sure canopy volume can be considered fixed factor- We were not quite sure how to interpret this comment; is the idea that canopy volume instead should be a random effect? We do not think it is possible to treat canopy volume (plant size) as a random effect in our models. First, random effects vary across individuals, while fixed effects do not. Treating canopy volume as a random factor would mean the effects of size on survival vary for each individual plant. On the other hand, block or plot identities are regularly treated as random effects. For example, in our analyses each plot (random effect) has a different intercept. Further, random effects in mixed models generate estimates of variance for the estimated distribution of intercepts/slopes, but not specific coefficients. One common way to describe fixed effects is as predictors researchers are “interested in”, meaning they aim to estimate coefficients and test hypotheses for those variables. For example, we could not have made Figures 3 and 4 without treating canopy volume as a fixed effect, one that has a specific coefficient (slope of the linear relationship with a logit link, in this case).- The approach we used is fairly common in demographic modeling for plants. Below we give a small sampling of references from the many studies where authors likewise fit GLMM statistical models for survival and growth, with plant size and sometimes climate variables as fixed effects and individual plot or year as random effects.- Dalgleish et al. 2011. Climate influences the demography of three dominant sagebrush steppe plants. Ecology 92: 75-84.- Miller et al. 2012. Evolutionary demography of iteroparous plants: incorporating non-lethal costs of reproduction into integral projection models. Proc. R. Soc. B. 2792831–2840.- Mandel and Ticktin 2012. Interactions among fire, grazing, harvest and abiotic conditions shape palm demographic responses to disturbance. Journal of Ecology 100 (997-1008).Table 1. Remove SE columns and include the numbers in brackets (or after ±) next to the mean value. Include different letters for significant differences between control and removal treatments.- We changed the columns in Table 1 as requested. It is not possible to put letters indicating significant differences, because the statistical analysis we used to evaluate treatment effects on cover combined all years, while Table 1 shows each year separately. We think Table 1 is a helpful supplement to the statistical results reported in the main text; it shows the changes in both control and non-native removal plots over time, and the non-significant trend towards a time by treatment effect."Seven A. californica were confirmed as resprouts in June 2014, 3 in control and 4 in removal plots. Almost all individuals established after the fire from seedlings". What was the percentage of germinated seedlings? These numbers are more precise than ‘almost all’.- We revised as follows (page 14, lines 286-288):“Seven A. californica were confirmed as resprouts in June 2014, 3 in control and 4 in removal plots. The other 454 adult plants identified and tagged from 2014-2019 recruited from seed (98.5 % of total).”Figure 2. It seems that density of recruits in removal plots were recorded later than in control plots. Correct it.- We offset the points and lines for the two treatments in this plot because otherwise they crowd each other out, making the figure difficult to read. (Because many of the values overlap near 0, we also needed to jitter the individual data points). We have revised the figure to reduce the amount of offset (distance between the two data series), and added a sentence to the legend clarifying that the sampling dates were identical for the two treatments. Note that offsetting in the R ggplot package always puts the series equidistant on either side from the x axis marker. Please advise if these changes do not fully address the concern. Page 14, lines 298-299.“Control and non-native removal plots were sampled on equivalent dates each year; the series are offset to facilitate visual comparison.”- "absence of non-native grass removal", in page 17, sounds weird these two (three in fact) negations in the sentence- We rephrased the sentence (Page 17, lines 362-364).“One previous study found that increasing grass density reduced A. californica seedling survival [21], while another observed complete seedling mortality unless non-native grasses were removed [52].”________________________________________Examples of replication and sampling area from similar researchOur study: 1 site. Experimental and post-fire shrub monitoring. Study area ~ 300 m by 10 m. 6 plots and 600 m2 sampled per treatment. 1006 seedlings and 461 adult plants tagged (7 years).Conlisk 2016 (PLoS One): 1 site. Post-fire monitoring. Study area ~ 200 by ~120 m. Four transects in burned area, four in unburned, each 100 m in length with 5 quadrats. 38 m2 sampled per ”treatment”.Molinari and D’Antonio (Biol Invasions) 2019. 1 site. Experimental. Study area in 8 blocks at least 5 m apart, each 3 m2 . 8 plots and 1.28 m2 sampled per treatment.Minnich 2014 (Global Change Biology). 1 site. Post-fire monitoring. Study area not specified. 416 plants tagged across all shrub species, max 200 per species (6 years).Eliason and Allen 2008 (Restoration Ecology): 1 site. Experimental. Total study area 12 m by 13 m. 3 m2 sampled per treatment.Cox and Allen 2008 (Journal of Applied Ecology): 1 site. Experimental. Study area ~2 hectares. 5 plots and 500 m2 sampled per treatment.DeSimone and Zedler, 1999 (Ecology): 1 site. Experimental. Study area not specified. 8 plots and 8 m2 sampled per treatment.Keeley and Keeley 1984 (American Midland Naturalist). 2 sites, both in Santa Monica Mountains, southern California. Postfire shrub survey., 10 plots and 80 m2 sampled. Total number of plants measured at all sites/for all species=857 (single time point).Malanson and O’Leary 1982 (Oecologia): 6 sites, all at Santa Monica Mountains, southern California. Postfire monitoring. Each site 24 m2 in area. Sampled 16 quadrats and 64 m2 per site.Submitted filename: Thomson PLOS One response to reviews June 10 2021.docxClick here for additional data file.28 Jun 2021Non-native plant removal and high rainfall years promote post-fire recovery of Artemisia californicain southern California sage scrubPONE-D-21-06311R1Dear Dr. Thomson,We’re pleased to inform you that your manuscript has been judged scientifically suitable for publication and will be formally accepted for publication once it meets all outstanding technical requirements.Within one week, you’ll receive an e-mail detailing the required amendments. When these have been addressed, you’ll receive a formal acceptance letter and your manuscript will be scheduled for publication.An invoice for payment will follow shortly after the formal acceptance. To ensure an efficient process, please log into Editorial Manager at http://www.editorialmanager.com/pone/, click the 'Update My Information' link at the top of the page, and double check that your user information is up-to-date. If you have any billing related questions, please contact our Author Billing department directly at authorbilling@plos.org.If your institution or institutions have a press office, please notify them about your upcoming paper to help maximize its impact. If they’ll be preparing press materials, please inform our press team as soon as possible -- no later than 48 hours after receiving the formal acceptance. Your manuscript will remain under strict press embargo until 2 pm Eastern Time on the date of publication. For more information, please contact onepress@plos.org.Kind regards,Cristina ArmasAcademic EditorPLOS ONEAdditional Editor Comments (optional):Reviewers' comments:13 Jul 2021PONE-D-21-06311R1Non-native plant removal and high rainfall years promote post-fire recovery of Artemisia californica in southern California sage scrubDear Dr. Thomson:I'm pleased to inform you that your manuscript has been deemed suitable for publication in PLOS ONE. Congratulations! Your manuscript is now with our production department.If your institution or institutions have a press office, please let them know about your upcoming paper now to help maximize its impact. If they'll be preparing press materials, please inform our press team within the next 48 hours. Your manuscript will remain under strict press embargo until 2 pm Eastern Time on the date of publication. For more information please contact onepress@plos.org.If we can help with anything else, please email us at plosone@plos.org.Thank you for submitting your work to PLOS ONE and supporting open access.Kind regards,PLOS ONE Editorial Office Staffon behalf ofDr. Cristina ArmasAcademic EditorPLOS ONE
Authors: R Brandon Pratt; Anna L Jacobsen; Aaron R Ramirez; Anjel M Helms; Courtney A Traugh; Michael F Tobin; Marcus S Heffner; Stephen D Davis Journal: Glob Chang Biol Date: 2013-12-26 Impact factor: 10.863
Authors: Peter S Coates; Mark A Ricca; Brian G Prochazka; Matthew L Brooks; Kevin E Doherty; Travis Kroger; Erik J Blomberg; Christian A Hagen; Michael L Casazza Journal: Proc Natl Acad Sci U S A Date: 2016-10-25 Impact factor: 11.205
Authors: Emily J Fusco; John T Finn; Jennifer K Balch; R Chelsea Nagy; Bethany A Bradley Journal: Proc Natl Acad Sci U S A Date: 2019-11-04 Impact factor: 11.205
Fig 1
Fig 2
Fig 3
Fig 4