The Effect of Ecophysiological Traits on Live Fuel Moisture Content

Abstract

Live fuel moisture content (LFMC) is an important metric for fire danger ratings. However, there is limited understanding of the physiological control of LFMC or how it varies among co-occurring species. This is a problem for biodiverse yet fire-prone regions such as southern California. We monitored LFMC and water potential for 11 native woody species, and measured ecophysiological traits related to access to water, plant water status, water use regulation, and drought adaptation to answer: (1) What are the physiological mechanisms associated with changes in LFMC? and (2) How do seasonal patterns of LFMC differ among a variety of shrub species? We found that LFMC varied widely among species during the wet winter months, but converged during the dry summer months. Traits associated with LFMC patterns were those related to access to water, such as predawn and minimum seasonal water potentials (Ψ), and water use regulation, such as transpiration. The relationship between LFMC and Ψ displayed a distinct inflection point. For most species, this inflection point was also associated with the turgor loss point, an important drought-adaptation trait. Other systems will benefit from studies that incorporate physiological mechanisms into determining critical LFMC thresholds to expand the discipline of pyro-ecophysiology.

1. Introduction

Live fuel moisture content (LFMC) is a landscape-level management metric that, along with weather and topography, is incorporated into rate-of-spread models and fire danger ratings [1,2,3,4]. LFMC is expressed as the ratio of water content in fresh plant tissue to the dry weight and represents the amount of moisture that needs to evaporate from a fuel source before ignition can occur. Greater fuel moisture means reduced flammability and lower likelihood of ignition [5], hence LFMC has a major effect on combustion, fire spread, and fire consumption [6,7,8]. Yet, the relationship between LFMC and fire risk is not always linear. Instead, there are thresholds. For example, wildfire risk increases as LFMC decreases until a critical threshold is reached, and then the fire risk is constant and high [9]. However, recent studies suggest that these breakpoints need to be considered with caution as field-based and laboratory-based studies often show differing results related to the magnitude of the effect of LFMC on fire rate-of-spread [2,4,10]. Hence, further LFMC studies are needed to resolve these issues.

California’s Mediterranean-type climate shrublands are especially prone to wildfires. For example, the natural fire regime in southern California historically included infrequent yet large, high-intensity fires [11]. However, modern fire frequencies are higher than normal, which threaten chaparral diversity [12,13]. For chaparral shrubs in southern California, LFMC is normally high during the winter and spring and then gradually declines with the onset of the characteristic summer drought (Figure 1) of Mediterranean-type climates (Figure 2a). This leads to a typical fire season about six months long (Figure 1). However, severe or extended drought conditions, which are becoming more common with climate change, lead to fuels drying out sooner, more quickly, to a greater extent, or for longer. In these instances, the predicted fire season starts earlier in the year and/or lasts longer. Extended droughts have been associated with high fire activity or anomalous seasonality. For example, five of nine historic “mega fires” (>50,000 ha) in southern California occurred during a period of anomalous drought [14]. Large, wind driven, spring fires can occur when LFMC is at levels that normally occur in the late summer or fall [15].

In some cases, extreme drought can lead to vegetation mortality [16] and hence increased dead fuel loads, which can also exacerbate fire risk. These drought conditions can lead to increased dead fuels, which are largely controlled by weather conditions and fuel thickness [8]. Alternatively, live fuel moisture covaries with environmental conditions, including temperature, precipitation, and soil moisture, as well as the phenological and physiological characteristics of a plant species [9,17,18]. Despite the widespread use of LFMC as a proxy for flammability and fire risk, only recently have quantitative studies emerged of how these data relate to traits of plant water relations components including water potentials, turgor loss point, relative water content, and hydraulic conductivity [19,20]. This represents a significant gap in our understanding of pyro-ecophysiology [21].

Determining the physiological mechanisms that underpin LFMC, especially in a fire-prone landscape such as California, has important implications for fire-risk management practices [22]. The objective of this research was to address: (1) What are the physiological mechanisms associated with changes in LFMC? and (2) How do seasonal patterns of LFMC differ among a variety of species? To address these questions, we measured LFMC and a variety of ecophysiological traits for 11 native woody chaparral and coastal sage scrub species. These traits included water potentials, gas exchange characteristics, and hydraulic traits. We hypothesized that the following traits would control changes in LFMC: (1a) traits associated with access to water, such as predawn water potentials and minimum seasonal water potentials, (1b) water use regulation, such as transpiration, and (1c) traits associated with drought adaptation, as determined from pressure-volume curves. We also hypothesized that: (2) these trait differences would manifest as differing seasonal LFMC patterns among species.

2. Methods and Materials

2.1. Study Site and Species

The study was conducted from July 2015 to March 2016 at the Stunt Ranch Santa Monica Mountains Reserve, located in the Santa Monica Mountains in California (34°5′27′′ N, 118°39′27′′ W). Stunt Ranch is 125 ha field station dominated by coverage of chaparral, with elevation ranging from 392 to 472 m. With a Mediterranean-type climate, average temperature highs are 32 °C and average lows are 4 °C (Figure 2A). Mean annual precipitation is 610 mm, occurring mostly during the winter months with almost no rainfall during the summer; however, there is very high interannual rainfall variability. Stunt Ranch last burned in 1993, 22–23 years before this study. This study measured 11 common woody species in California shrub land ecosystems (

Table 1. Study species, including family, leaf phenology, post-fire regeneration type, the live fuel moisture content inflection point (LFMC_IP) ± standard error, and the turgor loss point (Ψ_TLP; MPa) ± standard error at the Stunt Ranch Santa Monica Mountains Reserve. Sources of post-fire regeneration type: [31]¹, [32]², [33]³. For Ψ_TLP, different letter superscripts indicate significant differences (p < 0.05).

Species	Family	Leaf Phenology	Post-Fire Regeneration Type	LFMCIP	ΨTLP
Adenostoma fasciculatum	Rosaceae	Evergreen	Facultative sprouter 2	77% ± 5	−3.40 ± 0.17 B
Arctostaphylos glandulosa	Ericaceae	Evergreen	Facultative sprouter 2	86% ± 4	−2.51 ± 0.27 B,C
Ceanothus spinosus	Rhamnaceae	Evergreen	Facultative sprouter 1	121% ± 19	−2.80 ± 0.29 B,C
Cercocarpus betuloides	Rosaceae	Evergreen	Obligate sprouter 2	71% ± 5	−2.02 ± 0.09 A,C
Eriogonum fasciculatum	Polygonaceae	Semi-deciduous	Reseeder 3	67% ± 5	−3.23 ± 0.20 B
Heteromeles arbutifolia	Rosaceae	Evergreen	Obligate sprouter 2	98% ± 14	−2.70 ± 0.21 B,C
Malosma laurina	Anacardiaceae	Evergreen	Facultative sprouter 2	92% ± 10	−2.07 ± 0.14 A,C
Quercus agrifolia	Fagaceae	Evergreen	Obligate sprouter 2	93% ± 12	−2.61 ± 0.27 B,C
Quercus berberidifolia	Fagaceae	Evergreen	Obligate sprouter 2	81% ± 4	−1.97 ± 0.22 A,C
Salvia leucophylla	Lamiaceae	Drought-deciduous	Reseeder 3	85% ± 4	−1.47 ± 0.11 A
Salvia mellifera	Lamiaceae	Drought-deciduous	Reseeder 3	116% ± 8	−2.08 ± 0.14 A,C

).

2.2. Live Fuel Moisture Content

LFMC (%) was measured approximately every three weeks, concurrently with plant water potentials (see below). During sampling, a terminal shoot was collected from each individual, for at least three individuals per species during midday (11:00 to 13:00), approximating the procedures used by the Los Angeles County Fire Department and the Ventura County Fire Department in their monitoring of LFMC. Each sample was sealed in a separate plastic bag (Whirlpak, Nasco, and Fort Atkinson, WI, USA) and transported in a cooler to the lab.

LFMC is the ratio of water weight to dry weight of living plant tissue [23] and was determined by measuring the sample fresh weight and then placing the sample in a coin envelope and drying in an oven at 70 °C for at least 72 h, after which the dry mass was recorded. LFMC (%) was calculated as:

$$LFMC=\frac{Freshweight-Dryweight}{Dryweight}\ast 100$$

(1)

Minimum seasonal LFMC (LFMC_MIN) was determined from the lowest LFMC value a species experienced over the course of a year, and maximum seasonal LFMC (LFMC_MAX) was determined from the highest LFMC value a species experienced over the course of a year.

2.3. Plant Water Potential

At the same time, and from the same branch used for LFMC sample collection (see above), an additional terminal shoot was collected for a paired water potential (Ψ; MPa) measurement, for at least three individuals per species during midday (11:00 to 13:00). Each sample was sealed in a separate plastic bag (Whirlpak, Nasco, and Fort Atkinson, WI, USA) and transported in a cooler to the lab. Water potential was determined with a Scholander-type pressure chamber (PMS Instrument Corp., Corvallis, Oregon). In addition to midday water potentials, predawn water potentials (04:00 to 05:00) were sampled during the dry (August 2015; Ψ_PD;dry) and wet season (March 2016; Ψ_PD;wet) following the same protocol. Minimum seasonal water potential (Ψ_MIN) was determined from the most negative water potential a species experienced over the course of a year.

2.4. Gas Exchange

Photosynthetic gas exchange was measured using a portable photosynthesis system (LI-6400, LI-COR Biosciences, Lincoln, NE, USA) with a red-blue light source. Measurements were frequently taken from September 2014 to April 2016, during the morning and midday hours (9:00 to 14:00) on the most recently mature leaf. We determined the light-saturated rate of photosynthetic CO₂ assimilation per area (A_MAX; µmol·m⁻²·s⁻¹), stomatal conductance to water vapor (g_S; mol·m⁻²·s⁻¹), transpiration (E; mol·m⁻²·s⁻¹), and intrinsic water use efficiency (A/g_S; µmol·mol⁻¹).

2.5. Pressure-Volume Curves

Pressure-volume curves were determined on six terminal shoots per species using the bench-dehydration method [24,25] during August and September, which are dry summer months. Samples were collected from the field, sealed in plastic bags, and placed inside a cooler for transport back to the lab. In the lab, shoots were recut under water and placed in beakers with the distal cut stem end in water, and the entire sample covered in plastic to rehydrate overnight. The next day, samples were weighed and measured for water potential using a Scholander-type pressure chamber, and then dried on a lab bench. Weighing, measuring water potential, and drying samples were repeated until achieving a water potential of about −4 to −6 MPa, depending on the species. Subsequently, samples were put in a paper bag and dried in an oven at 70 °C for at least 48 h to determine dry mass. We calculated saturated water content (SWC; %), water potential at turgor loss point (Ψ_TLP; MPa), relative water content at turgor loss point (RWC_TLP; %), osmotic potential (π_o; MPa), modulus elasticity (ε; MPa), capacitance at full turgor (C_FT; MPa⁻¹), capacitance at turgor loss point (C_TLP; MPa⁻¹), and leaf dry matter content (LDMC; g·g⁻¹) using the Excel spreadsheet available on PrometheusWiki [26].

2.6. Statistical Analyses

All statistical tests were performed in R Studio (ver 0.99.485) [27]. Piecewise regression using the ‘segmented’ package was used to find the inflection point, standard error, and 95% confidence intervals for the relationship between water potential and LFMC (LFMC_IP) for each individual species and for all species combined. We tested for differences in Ψ_TLP among species using ANOVA with the ‘aov’ function from the ‘stats’ package and a Tukey post-hoc test. Pearson correlation was performed with the ‘rcorr’ function from the ‘Hmisc’ package to produce a correlation matrix and evaluate the relationships among all measured traits [28,29]. Before running correlations, traits were tested for normal distribution using the Shapiro-Wilk test of normality [30] and transformed if necessary. Transformed traits included A_MAX, g_S, E, SWC, and LFMC_MAX. The correlogram was constructed using the ‘corrplot’ function.

3. Results

LFMC and water potentials (Ψ) varied among species and seasons (Figure 2B,C). LFMC was highest and most variable among species during the wet months of January through April, while species tended to converge towards low LFMC during the dry season. LFMC_MIN occurred in October and November, while LFMC_MAX occurred in February and March. Ψ were lowest during the dry season, particularly in October, November, and December, right before winter rains began. Variability in Ψ among species was greatest during the dry season. When winter rains started in January, species converged towards less negative Ψ.

Salvia mellifera had the highest LFMC_MAX, while Quercus berberidifolia had the lowest LFMC_MAX. Eriogonum fasciculatum had the lowest LFMC_MIN, while Malosma laurina had the highest LFMC_MIN. Salvia mellifera had the lowest Ψ_MIN, and Malosma laurina had the highest Ψ_MIN. In fact, deeply rooted Malosma laurina did not show a substantial seasonal change in Ψ or LFMC.

In comparing LFMC to Ψ (Figure 3, Supplemental Figure S1), LFMC initially decreased rapidly with only a small decline in Ψ, before reaching an inflection point (LFMC_IP) and leveling off. After LFMC_IP, there were large declines in Ψ with only a small decline in LFMC. In addition, the confidence intervals for LFMC_IP overlapped with the confidence intervals for Ψ_TLP for the compiled data set (Figure 3), which included all study species, as well as for individual species, except for A. fasciculatum, Heteromeles arbutifolia, M. laurina, and S. leucophylla (Supplemental Figure S1). For H. arbutifolia and M. laurina, the relationship between LFMC and Ψ was not tightly constrained, which made it difficult for the model to converge on an inflection point. For A. fasciculatum, Ψ_TLP was more negative than the inflection point, while for S. leucophylla, Ψ_TLP was less negative than the inflection point. For the remaining species, E. fasciculatum had the lowest LFMC_IP at 67%, while C. spinosus had the highest LFMC_IP at 121%.

Species with lower LFMC_MIN also had lower (more negative) Ψ_MIN, lower Ψ_PD;wet, and lower Ψ_PD;dry (Figure 4). For example, A. fasciculatum had the second lowest Ψ_MIN and the lowest Ψ_PD;wet and Ψ_PD;dry. LFMC_MAX was correlated with A_MAX, E, SWC, C_FT, C_TLP, and LDMC. LFMC_IP was correlated with SWC.

Ψ_TLP among species ranged from −1.47 MPa to −3.40 Mpa (Table 1). S. leucophylla had the highest (least negative) Ψ_TLP and A. fasciculatum had the lowest (most negative) Ψ_TLP. While S. leucophylla had the lowest Ψ_TLP, it was not significantly different from Q. berberidifolia, C. betuloides, M. laurina, or S. mellifera. While E. fasciculatum and A. fasciculatum had the two lowest (most negative) Ψ_TLP and were also significantly different from the five previously mentioned species, they were not significantly different from A. glauca, Q. agrifolia, H. arbutifolia, and C. spinosus.

4. Discussion

4.1. Overview

Our study aimed to answer, (1) What are the physiological mechanisms associated with changes in LFMC? We found that minimum seasonal live fuel moisture content (LFMC_MIN) was correlated with water access traits, including wet season predawn water potential (Ψ_PD;wet), dry season predawn water potential (Ψ_PD;dry), and minimum seasonal water potential (Ψ_MIN), confirming hypothesis 1a as well as previous work by [19]. Maximum seasonal live fuel moisture content (LFMC_MAX) was correlated with photosynthesis (A_MAX) and transpiration (E), confirming hypothesis 1b. LFMC_MAX was correlated with drought-adaptation traits, including SWC, relative C_FT, and relative capacitance at zero turgor (C_TLP), LDMC, supporting hypothesis 1c.

We also asked, (2) How do seasonal patterns of LFMC differ among a variety of species? We found that species trait differences were associated with differing seasonal LFMC patterns among species, confirming hypothesis 2. LFMC varied among species to the greatest extent during the wet winter months, but all species converged on low LFMC values during the dry season. LFMC_MAX occurred during the wet season, along with maximum rate of A_MAX and E, while LFMC_MIN and Ψ_MIN occurred during the dry season. The coastal sage scrub species S. mellifera and S. leucophylla, which are drought-deciduous and shallow-rooted, had much higher LFMC_MAX values compared to any other species during the wet season. However, Salvia species displayed the lowest Ψ_MIN. Malosma laurina, which is an evergreen chaparral species characterized by deep roots, maintained the highest LFMC during the dry season, never dropping below 87%.

4.2. Physiological Controls of LFMC

Traits associated with water access and water use regulation were critical determinants of LFMC among species. One of the most important traits that controlled LFMC in this semi-arid plant community was water potential. Predawn water potentials and minimum seasonal water potentials, which are indicative of species access to water [34] and regulation of water loss when stomata are closed (i.e., cuticular conductance) [35], were correlated with LFMC_MIN (Figure 4). Deeply rooted plants have greater access to soil moisture for longer into the dry season than shallow-rooted plants, supporting either the delay of LFMC decline or keeping minimum LFMC high [36]. In addition to access to water, traits associated with water use regulation played a significant role in controlling LFMC. For example, species with the highest LFMC_MAX also had high A_MAX and high E, providing evidence that species with high gas exchange rate rates also have high moisture content in their tissue. Overall, this points to a suite of traits that coordinate water access, water use regulation, and plant water status [37].

By plotting concurrently measured LFMC and Ψ together on the same plot, we examined the relationship between a landscape-level management metric used for assaying fire danger (LFMC) and a well-established physiological measurement of plant water status (Ψ). As plants started to dry out, LFMC rapidly declined with only a small change in Ψ, before reaching an inflection point (LFMC_IP), after which there were large declines in Ψ with only a small change in LFMC. The confidence intervals for LFMC_IP overlapped with turgor loss point (Ψ_TLP) confidence intervals for the compiled data set. Ψ_TLP is an important drought-adaptation trait; species with a lower (more negative) Ψ_TLP can withstand more negative Ψ before losing leaf turgor, or before wilting. At the point that leaves have lost so much water that they lose turgor, they would also be highly flammable as the amount of moisture in the leaves would be low, hence ignition could easily occur. However, this would correspond to a higher critical LFMC threshold than the current 60% used by fire departments in southern California. This finding confirms work by Dennison & Moritz [38], which used a different approach to determine a critical LFMC threshold involving precipitation records, and builds on the recent LFMC versus fire behavior breakpoint work of Pimont [10,39], with our physiological data.

While the LFMC_IP corresponded to the Ψ_TLP for the compiled data set, this was not the case for each individual species (Supplemental Figure S1). There are a few possibilities for this result. One cause could be phenology [40] and, more specifically, growth. Dry matter “accumulates” when plants grow, which would exert a strong seasonal effect on the relationship between LFMC_IP and Ψ_TLP for species with differing phenological patterns [41,42]. Another potential factor is related to water storage and water release dynamics of different plant species, including the release of capillary water versus water released from cavitation events [43,44]. Using a physiologically meaningful LFMC threshold, as supported by flammability and precipitation data, along with an appropriate indicator species is important when monitoring fire-risk conditions.

4.3. Indicator Species Choice Can Impact Fire Danger Rating

As we have shown, LFMC and therefore fire-risk monitoring will vary depending on the indicator species. By comparing 11 dominant species in southern California, a region with high plant biodiversity, we can quantitatively compare LFMC of Adenostoma fasciculatum, the current indicator species, against other co-occurring species. Adenostoma fasciculatum is used as an LFMC indicator species in southern California because of its widespread distribution and abundance. In fact, it is present in 70% of California chaparral communities [45]. In addition, it is viewed as being highly flammable due to its oily leaves (giving rise to its other common name “greasewood”) and multi-stemmed canopy structure that consists of dense clusters of small stems, branches, and flower stalks [46]; about 60% of stems are <1.27 cm in diameter [47]. This makes it a fine fuel with small stems and leaves, especially in comparison to co-occurring large branching or broad-leafed species. Our data show that of the 11 study species examined, A. fasciculatum was one of the first species to reach the 79% critical LFMC threshold proposed by [38] and it stayed below that threshold during the seasonal dry period. In fact, it often also fell below the 60% LFMC threshold used by LACFD when many other species did not. The only other species to consistently do so was C. betuloides. Hence, A. fasciculatum is a good indicator species for measuring LFMC in southern California chaparral mountains. The high seasonal LFMC values of M. laurina, H. arbutifolia, and Q. agrifolia make them less optimal as LFMC indicator species.

5. Conclusions

In their 2018 Perspective paper, Jolly and Johnson [21] argue,

The discipline of ecophysiology is rich and mostly unleveraged in live fuel research, yet it has the potential to link plant flammability traits at both the leaf and plant level to fundamental laws that govern how plants functions.

Our study takes the initial steps to address this and shows that traits associated with water access and water use regulation were critical determinants of seasonal LFMC patterns among species. Other systems would benefit from physiological studies of their chosen indicator species and critical LFMC thresholds. This understanding is crucial as the number and size of large fires in the western United States have increased, especially when coincident with drought [48]. Increasing temperatures and longer, more intense, and/or more frequent droughts will become common with climate change, so accurately forecasting wildfire risk is an immediate issue.