Multiple stressors in multiple species: Effects of different RDX soil concentrations and differential water-resourcing on RDX fate, plant health, and plant survival

Response to simultaneous stressors is an important facet of plant ecology and land management. In a greenhouse trial, we studied how eight plant species responded to single and combined effects of three soil concentrations of the phytotoxic munitions constituent RDX and two levels of water-resourcing. In an outdoor trial, we studied the effects of high RDX soil concentration and two levels of water-resourcing in three plant species. Multiple endpoints related to RDX fate, plant health, and plant survival were evaluated in both trials. Starting RDX concentration was the most frequent factor influencing all endpoints. Water-resourcing also had significant impacts, but in fewer cases. For most endpoints, significant interaction effects between RDX concentration and water-resourcing were observed for some species and treatments. Main and interaction effects were typically variable (significant in one treatment, but not in another; associated with increasing endpoint values for one treatment and/or with decreasing endpoint values in another). This complexity has implications for understanding how RDX and water-availability combine to impact plants, as well as for applications like phytoremediation. As an additional product of these greenhouse and outdoor trials, three plants native or naturalized within the southeastern United States were identified as promising species for further study as in situ phytoremediation resources. Plumbago auriculata exhibited relatively strong and markedly consistent among-treatment mean proportional reductions in soil RDX concentrations (112% and 2.5% of the means of corresponding values observed within other species). Likewise, across all treatments, Salvia coccinea exhibited distinctively low variance in mean leaf chlorophyll content index levels (6.5% of the means of corresponding values observed within other species). Both species also exhibited mean wilting and chlorosis levels that were 66% and 35%, and 67% and 84%, of corresponding values observed in all other plants, respectively. Ruellia caroliniensis exhibited at least 43% higher mean survival across all treatments than any other test species in outdoor trials, despite exhibiting similar RDX uptake and bioconcentration levels.

5 86 to target concentrations of 50 ppm and 100 ppm in a 10:3 silica quartz sand:loess soil (loess local 87 to study area). 88 89 Table 1. Plant species and sample sizes per treatment (n t ) for greenhouse studies.

Species
Common Name n t Plants (n = 3-5) in each of the 0, 50, and 100 ppm soil concentration groups were further 92 assigned to either a 1X or 0.5X water-resourcing treatment, for a total of six treatment groups 93 (Table 2). Each plant received either approximately 1 L (1X) or 0.5 L (0.5X) of municipal tap 94 water every other day. We assumed that at least one of the two levels of water-resourcing (1X or 95 0.5X) would constitute a stressor for each species, being comparatively further from the optimal 96 water-resourcing level for that species under greenhouse conditions. Miracle-Gro ® Water  (Table 3). The former two species were included in this trial because they had 142 provided a relatively consistent and productive bloom set in the greenhouse, and the outdoor 143 trials were also part of a concurrent pollination study. The latter species, C. canescens, had 144 thrived on the outdoor plot in an earlier pilot study. This species was thus included in this trial, 145 despite being excluded from the earlier greenhouse trials due to rapid mortality across all 146 treatments.  Table).   204 We were able to maintain all eight plants species (Table 1) (Figs. 2-3). In many cases, PRC values for the 100 ppm RDX treatments were higher 238 than with the 50 ppm RDX treatments, but pairwise comparisons were not always statistically 12 239 significant (Figs. 2). In terms of bioaccumulation, the 50 ppm RDX treatments generally 240 exhibited higher root BCF values than treatments with 100 ppm RDX, though, again, pairwise 241 comparisons were not always statistically significant (Fig. 3). As a single factor, water-242 resourcing had little effect on PRC across species and treatments (Fig. 2), but was more 243 commonly a factor in root BCF (Fig. 3). In three plant species, reduced water-resourcing was 244 associated with lower root BCFs (Fig. 3). Of particular interest, significant interaction effects 245 between the main factors (RDX soil concentrations and water-resourcing levels) were apparent 246 for both PRCs and BCFs within several species. The 0.5X|100 ppm treatment resulted in the 247 highest levels of PRC for most species, though PRCs for this treatment group was not always 248 statistically different from other treatment groups (Fig. 2). In terms of root RDX accumulation, 249 the 01X |100 ppm treatment group often exhibited the highest BCFs (Fig. 3). Significant 250 interaction effects were observed for both PRCs and BCFS, in three and two species 251 respectively (Figs. 2-3). The strength (significant or not significant) and direction (increasing or 252 decreasing endpoint values) of those effects varied from case to case (Table 4).

Interaction
Patterns soil RDX reduction (proportional reduction concentration; PRC), root RDX accumulation (bioconcentration factor; BCF;), 256 and plant health metrics including leaf wilting, leaf chlorosis, and leaf chlorophyll content (CCI) across multiple plant species exposed 257 to different combinations of two main factors (50 ppm and 100 ppm soil RDX concentration, 0.5X water-resourcing level, and 258 interaction effects). Effect directions for factors are provided for those cases in which at least one factor, and in some cases, different 259 treatments, were found to be associated with significant differences in endpoint levels (Figs. 2-6). Blank cells indicate that no 260 statistically significant effect was observed. Indicators include ↑ (increase in endpoint value or metric), ↓ (decrease), ↕ (increase and 261 decrease), ↥ (increase and little to no effect depending on treatment groups), ↧ (decrease and little to no effect depending on treatment 262 groups), and -(no discernible effect on a metric).
14 263 a Relative to effect observed for 500 ppm RDX treatment. 264 b Relative to effect observed with 1X water-resourcing treatment.
265 c Relative to endpoint values observed in treatments without RDX. The effects of different water-resourcing levels, soil RDX concentrations, and the 267 interaction of these factors on wilting levels were varied and complex for nearly all of the eight 268 plant species within the greenhouse trials (Fig. 4). One general trend was that different RDX 269 concentrations were most common factor influencing plant health responses to different 270 treatments (Fig. 4-6; Table 4). Most plants across treatments and species exhibited only mild wilt 271 (i.e., 0-25%), or in several cases, moderate wilt (i.e., 25-50%; Fig. 3). Higher degrees of wilt (> 272 50% or greater of leaf surface) were relatively rare, and were observed in only a few treatment 273 groups within Dianthus and P. lanceolata. In seven of eight species, there was a statistically 274 significant association between starting RDX soil concentrations and observed differences in 275 mean wilting category. In some cases, higher RDX concentrations appeared to be associated with 276 increasing levels of wilt, yet conversely, in some cases, higher soil RDX concentrations appeared 277 to benefit plants (reduced levels of wilt; Fig. 4; Table 4). Differences in water-resourcing levels 278 were associated with significant differences in mean wilting category in three species, with 279 varied effect strengths and directions ( Fig. 4; Table 4). Significant interaction effects between 280 the main factors were observed in four species, with the strength and direction of those effects 281 being mixed, even within species ( Fig. 4; Table 4).  (Fig. 7). PRCs were not widely divergent