Weed control and environmental risk assessment of herbicides in a Brazilian Atlantic forest railway ecosystem

Abstract

Weeds along railway lines pose serious operational and environmental challenges, particularly in protected natural areas where vegetation management must balance efficacy and ecological safety. The objective of this study was to select herbicides that are effective in controlling weeds and to evaluate the associated environmental risks. The experiment was conducted on a stretch of railway line in the Serra do Mar State Park, São Paulo, Brazil, with the application of single and mixed herbicides (glyphosate, indaziflam, imazapyr and saflufenacil). A phytosociological survey of the weed flora was performed, and the overall weed control and the control of the most common species (Glycine max, Paspalum spp, Commelina spp. and Digitaria spp.) were evaluated at 15, 30, 60, 90, 120, 150 and 180 days after application. For the analysis of possible environmental risks associated with the use of these herbicides, quantification of the active ingredients and their most important metabolites in soil and water was performed. The weed community on the railroad was predominantly composed of exotic species. The combination of pre- and postemergent herbicides provided the highest percentages of control with the longest weed-free period. The herbicide concentrations detected in the soil and water on the railway line and in the immediate vicinity were much lower than the initial concentrations, with no observed effect concentration (NOEC) for the most sensitive organisms found in the literature. The use of these herbicides, under the conditions evaluated, was efficient in the management of weeds and was environmentally safe.

Introduction

Railway corridors are critical linear infrastructures where vegetation management is essential to maintain operational safety and prevent accidents. This challenge is amplified when rail lines traverse biodiversity-rich regions, where weed control must be effective yet compatible with conservation goals. In the Brazilian Atlantic Forest, a global biodiversity hotspot with high endemism and a long history of habitat loss^1–3, the Serra do Mar State Park (Parque Estadual da Serra do Mar, PESM) represents the largest continuous preserved remnant in Brazil and is legally designated as a fully protected conservation unit. Beyond exceptional biodiversity, the PESM safeguards headwaters that supply major coastal and inland regions, stabilizes steep slopes, and contributes to local climate regulation, ecosystem services especially critical where railways intersect rugged terrain and dense forest. Ensuring safe railway operations while safeguarding Atlantic Forest ecosystems therefore requires ecologically informed weed control and robust environmental risk assessment in and around Protected Areas (PAs)⁴.

Against this backdrop, the PESM also intersects strategically with Brazil’s rail logistics. Located in the state of São Paulo and spanning 25 municipalities, it hosts one of the main railway a corridor for agribusiness exports^5,6. The Brazilian rail system has approximately 30 thousand km of rail network^7,8, with cargo transportation focused mostly on products from mining, agricultural production, civil construction and steel mills⁹.

Train traffic and track maintenance can be affected by several factors, including the presence of vegetation along railway lines. The absence of weed control can affect activities related to train traffic and the maintenance of lines, yards and engineering sites that make up the infrastructure and superstructure of the track¹⁰. Therefore, the growth of weeds may be related to visibility problems, wheel slippage, risk of derailment with spillage of the load in the protection areas, fire risks, damage to signaling and electrical systems, and impaired ballast water drainage^11,12, in addition to compromising the safety of employees who work in the maintenance of railway lines¹³. In this sense, studies focused on the management of weeds in railway areas make it possible to know and map weed species, determine the ecological factors that are related to the abundance of the plants, implement effective and sustainable management, and supervise and analyze management techniques to prevent further infestation^14,15.

Chemical control is a widely used technique because it has good results in the control of several species. However, it should be noted that these results should be linked to the absence of side effects, such as risks to the health of employees, damage to equipment and to the environment along the tracks¹⁶. The physicochemical and ecotoxicological properties of herbicides are generally well known, but they are not always applied to the railway environment; thus, to select the appropriate herbicide, in addition to its effects on weeds, it is necessary to understand the dynamics of the products in the environment in which they are being applied¹⁶, especially in Protected Areas. Thus, the use of herbicides with different mechanisms of action and with varied dynamics, such as the herbicides glyphosate, imazapyr, indaziflam and saflufenacil, helps create more efficient management programs.

Glyphosate (N-phosphonomethyl glycine) is an herbicide that acts on plants by inhibiting the 5-enolpyruvylshikimate-3-phosphate synthase (EPSPs) enzyme and is one of the herbicides most used in agricultural and nonagricultural cropping systems^17–20. The physical-chemical and ecotoxicological characteristics of glyphosate (Table 1) make it an environmentally safe product; this product has low volatility, high adsorption in the soil, and a low octanol/water coefficient, which indicates low mobility, low potential for bioaccumulation and high solubility in water²¹. Despite being highly soluble, the risks of groundwater contamination are low as a result of strong sorption and rapid dissipation in soils²⁰. When in contact with the soil, glyphosate is rapidly degraded by microorganisms, and its main degradation product is aminomethylphosphonic acid (AMPA), which is also strongly adsorbed to soil particles, leading to pollution of the environment, especially soil and water resources^17,20.

Table 1.

Physicochemical and ecotoxicological characteristics of the herbicides glyphosate, imazapyr, Indaziflam and saflufenacil.

Characteristics	Glyphosate	Imazapyr	Indaziflam	Saflufenacil
Physical chemistry
Chemical structure
Molecular formula	C3H8NO5P	C13H15N3O3	C16H20FN5	C17H17ClF4N4O5S
Solubility in water (20 °C) (mg/L)	10,500	9740	2,8	2100
pKa	2.34	1.9	3.5	4.41
Kow (log)	−3.2	0.11	2.8	2.6
Koc	1424	-	1000	-
Half-life (days)	15	11	150	20
Vapor pressure (mPa)	1.31 × 10−2	1.3 × 10−2	2.5 × 10−5	4.5 × 10−15
Ecotoxicology*
Terrestrial environment	Earthworms - Chronic NOEC, reproduction	Earthworms - Acute 14-day LC₅₀	Earthworms - Chronic NOEC, reproduction	Earthworms - Acute 14-day LC₅₀
	> 28,8 mg kg⁻¹	133 mg kg⁻¹	34 mg kg⁻¹	> 1000 mg kg⁻¹
Aquatic environment	Fish - Chronic 21-day NOEC	Fish - Chronic 21-day NOEC	Aquatic invertebrates - Chronic 21-day NOEC	Fish - Chronic 21-day NOEC
	1000 ppb	43,100 ppb	340 ppb	997 ppb

^*Ecotoxicological information for each herbicide and the most sensitive test organisms in terrestrial and aquatic environments. Source: PPDB: Pesticide properties database: University of Hertfordshire²¹.

Imazapyr (2-[(RS)-4-isopropyl-4-methyl-5-oxo-2-imidazolin-2-yl]nicotinic acid) is an herbicide of the imidazolinone chemical group that inhibits the enzyme acetolactate synthase (ALS) and controls a wide spectrum of broad- and narrow-leaved weeds in agricultural and nonagricultural areas^22–26. The physicochemical characteristics of imazapyr (Table 1) make it an herbicide with leaching potential due to its persistence and mobility in soils, high water solubility and very low volatility. In addition, low sorption in soils may vary with soil properties^21,23,27,28.

Indaziflam (N-[(1R,2S)-2,3-dihydro-2,6-dimethyl-1H-inden-1-yl]-6-[(1R)-1-fluoroethyl]-1,3,5-triazine-2,4-diamine) is an herbicide that inhibits cellulose biosynthesis and therefore controls mono- and eudicot weeds with pre- or postemergence application. It is characterized by a high residual period in the soil, greater than 150 days²⁹, low water solubility (2.8 mg/L at 20 °C), moderate to little mobility in soil and low bioaccumulation potential. Indaziflam dissipates in the environment mainly through biotic degradation and leaching, and the metabolites formed in the degradation process are more mobile than indaziflam^30–32. The metabolite fluoroethyldiaminotriazine (FDAT) has low sorption and is considered to be mobile to highly mobile in soil (Koc ranging from 10 to 50 mL/g oc) with relatively high leaching potential^32,33.

Saflufenacil (2-chloro-4-fluoro-5-[3-methyl-2,6-dioxo-4-(trifuoromethyl)pyrimidin-1-yl]-N-[methyl(propane-2-yl)sulfamoyl]benzamide) is an herbicide that inhibits the enzyme protoporphyrinogen oxidase (PROTOX) and is used mainly to control broadleaf weeds preemergence and as a preharvest desiccant aid for some crops^34,35. It is an herbicide that has a low sorption coefficient, moderate persistence, high solubility and low volatility (Table 1) and can reach aquatic systems^36–38.

These four herbicides were selected to reflect operational standards for ballast and right-of-way vegetation management in tropical railways, while spanning distinct modes of action and environmental fate, from systemic to contact activity and from short-lived to residual persistence, with differing mobility (Table 1), enabling a comparative assessment under Atlantic Forest conditions. As a program portfolio, they offer complementary spectra of control that sustain efficacy, support resistance management, and may reduce intervention frequency in or adjacent to Protected Areas.

Under labeled rates, the evaluated programs are expected to deliver weed control compatible with railway safety in the PESM, while predicted exposures in soil and surface water remain below effect thresholds for representative non-target endpoints, conditions that indicate acceptability for use within or near Protected Areas. Combining complementary modes of action is anticipated to sustain efficacy without increasing environmental risk relative to single-MOA options. Accordingly, this study characterizes the railway weed flora, identifies herbicide options that meet operational needs, and evaluates their environmental risks under Atlantic Forest conditions to inform evidence-based vegetation management in the PESM.

Materials and methods

Experimental area and phytosociological survey on the railway line

The experiment was carried out in the city of São Vicente, in the Serra do Mar State Park, State of São Paulo (23°55’56.87"S; 46°29’31.64"W), on a stretch of railway line with a length of approximately 1.8 km. During the experiment, the accumulated precipitation was 472 mm and the average minimum and maximum temperatures were 15.4 and 24.6 °C, respectively (Fig. 1).

Fig. 1.

Daily precipitation and average minimum and maximum temperatures during the experiment.

Before the application of the treatments (03/17/2020), water and soil samples were collected for the quantification of the compounds associated with the treatments, characterizing the area prior to herbicide application. In addition to the previous quantification of the compounds, a phytosociological survey was conducted with weed sampling to characterize the species present in the experimental area using the square inventory method³⁹, in which a square (0.25 m²) was dropped at 16 random points in the center and to the sides of the railway line.

The formal identification of these naturally occurring weed species was performed by the authors of this manuscript (Plínio Saulo Simões and Renato Nunes Costa). When necessary, specialized identification manuals were utilized to confirm species identification^40–43. Due to the ephemeral nature of the plant material and the extensive experience of the identifying authors with the local flora, these plants are common and not rare, voucher specimens were not deposited in a public herbarium. For this phytosociological characterization, all plants collected within the sampling areas were dried in a forced-air oven at 70 °C until constant mass and subsequently weighed to determine their dry biomass.

Based on the collection and identification of the plants, the biomass of each species and parameters related to the phytosociology of weeds were determined using the formulas proposed by Mueller-Dombois, D. & Ellemberd (1974)⁴⁴ and Braun-Blanquet (1979)³⁹: relative and absolute frequencies, relative and absolute densities, relative dominance and the importance value index (IVI). For the analysis of similarity between species, the similarity index (SI) was used⁴⁵. We analyzed the similarity of the populations of weed species between the center and the sides of the railway line.

Treatments and herbicide application

For weed management and risk assessment, postemergence and pre-emergence herbicides, applied singly and in mixtures, were adopted as treatments (Table 2). Allnon-agricultural used in the study are registered for use in nonagricultural areas. The herbicide application occurred in 100 m plots for each treatment on 03/18/2020 using a backpack sprayer with a gasoline engine (TEKNA 26 cc, model PC260TK) to which two flat spray nozzles without bars were attached. The nozzles (Teejet brand, 1/4XP10R and 1/4XP10L models) had a working width of 5 m, pressure of 2.5 bar and flow of 200 L ha⁻¹. The conditions at the time of application were clear and sunny, with a relative humidity of 65% and a temperature of 28 °C. All applications occurred within the licensed railway right-of-way and followed the Park’s management plan and applicable permit conditions.

Table 2.

Description of treatments applied for weed management in railways.

Treatment	g i.aa or e.ab ha−1
Control (without herbicide)	-
Glyphosate c	720
	1440
Glyphosate + imazapyr d	720 + 2500
Imazapyr	1500
	2500
Glyphosate + indaziflam e	1440 + 125
	1440 + 200
Imazapyr + indaziflam	1500 + 125
	2500 + 200
Saflufenacil f	180
	390
Imazapyr + saflufenacil	1500 + 180
	1500 + 390

^ag i.a. – grams of active ingredient; ^bgrams of acid equivalent; ^ctrademark Scout^Ⓡ (720 g ea kg ⁻¹), manufactured by Monsanto do Brasil; ^dArsenal trademark^Ⓡ NA (250 g area L ⁻¹), manufactured by Basf; ^eEsplanade trademark^Ⓡ (500 g a.i. L ⁻¹), manufactured by Bayer Environmental Science; ^fHeat trademark^Ⓡ (700 g a.i. kg- ¹), manufactured by Basf.

Evaluations of weed control and water and soil sampling

Visual evaluations of weed control were performed at 15, 30, 60, 90, 120, 150 and 180 days after application (DAA). Control scores were assigned to the general infestation and to the species Glycine max, Paspalum spp., Comellina spp. and Digitaria spp., adopting a percentage scale of scores between “0” and “100”, where “0” corresponded to no control and “100” to total weed control. A control without herbicide application was used as the standard⁴⁶.

For the quantification of herbicides, soil and water sampling was performed at different points throughout the study. Soil samples were taken at a depth of 0–10 cm at 10 points within each plot treated with the herbicides and from the control at 15, 30, 60, 90, 120, 150 and 180 DAA and 10 points outside the control area of each plot at 15, 150 and 180 DAA (Fig. 2). The water samples were collected at 19 points distributed throughout the test, including running and standing water at 15, 30, 60, 90, 120, 150, and 180 DAA (Fig. 2). All the water collection points were georeferenced by GPS (Table 3) so that collections were performed at the same points throughout the study.

Fig. 2.

Experimental design of water and soil collections on a railway track after application of the herbicides glyphosate (Gly), imazapyr (Imaz), indaziflam (Ind) and saflufenacil (Saflu). The blue arrows refer to the water collection points (P1 to P19), and the red dots refer to the soil collection points.

Table 3.

Geographic coordinates of the water collection points and water characteristics (current and standing).

Point	Latitude	Longitude	Water characteristic	Point	Latitude	Longitude	Water characteristic
P1	23°55’56.87"S	46°29’31.64"O	Current	P11	23°55’47.95"S	46°29’11.97"O	Standing
P2	23°55’56.75"S	46°29’31.37"O	Standing	P12	23°55’46.17"S	46°29’8.70"O	Standing
P3	23°55’54.40"S	46°29’28.55"O	Standing	P13	23°55’44.90"S	46°29’6.76"O	Current
P4	23°55’53.50"S	46°29’27.55"O	Standing	P14	23°55’43.73"S	46°29’5.11"O	Standing
P5	23°55’51.53"S	46°29’25.02"O	Standing	P15	23°55’43.00"S	46°28’56.36"O	Current
P6	23°55’50.80"S	46°29’23.40"O	Current	P16	23°55’46.68"S	46°28’51.81"O	Current
P7	23°55’50.50"S	46°29’22.59"O	Standing	P17	23°55’48.29"S	46°28’38.97"O	Standing
P8	23°55’49.86"S	46°29’19.85"O	Current	P18	23°55’48.03"S	46°28’36.39"O	Standing
P9	23°55’49.31"S	46°29’16.11"O	Standing	P19	23°55’48.08"S	46°28’34.01"O	Standing
P10	23°55’48.81"S	46°29’14.38"O	Standing

Quantification of herbicides and metabolites

To quantify the herbicides, after standardization of the soil samples, 7 g of each sample was weighed and placed in plastic cartridges with a total volume of 10 mL, consisting of a porous tablet to retain soil particles and attached to a compartment for the collection of the solution^47,48. The soil samples were saturated with deionized water in amounts ranging from 0.8 to 2.0 ml per cartridge. After being saturated, the cartridges containing the soil were allowed to rest for 24 h at 20 °C in the absence of light.

To extract the soil solution, the cartridges were centrifuged at 3270 G and 20 °C for 5 min (Hettich Zentrifugen centrifuge). The solution in the collector was filtered (0.45 μm PVDF membrane, 13.0 mm diameter), placed in 2.0 mL vials and stored in a freezer for later quantification. The analyzed compounds were glyphosate and its main metabolite, aminomethylphosphonic acid (AMPA); indaziflam and its metabolite, fluoroethyl-diaminotriazine^®6-(1-fluoroethyl)−1,3,5-triazine-2,4-diamine (FDAT); imazapyr; and saflufenacil.

The herbicides and their respective metabolites were analyzed using an LC‒MS/MS system consisting of a High-performance Liquid Chromatograph (HPLC) (Shimadzu, Proeminence UFLC, Kyoto, Japan) equipped with two LC-20AD pumps, SIL autoinjector-20AC, DGU-20A5 degasser, CBM-20 A controller system and CTO-20AC oven. A 4500 Triple Quad mass spectrometer (Applied Biosystems, Foster City, USA) was attached to the HPLC.

The chromatographic analyses of the herbicides imazapyr, saflufenacil, indaziflam and their respective metabolites were conducted with a C18 column (Synergi 2.5 µ Hydro RP 100 Å, Phenomenex) using an injection volume of 20 µl with 0.1% formic acid. (phase A) in water and 0.1% formic acid in methanol (phase B). The flow rate used was 0.6 ml min^−1, and the solvent ratio gradually increased from 20:80 (methanol/water) in the range of 0 to 1 min to 95:5 in the range of 1 to 5 min and returned to the initial condition in the range of 8 to 10 min. The total running time was 12 min. For the glyphosate and AMPA analyses, a C18 column (Gemini 5µ C18RP 110 Å, Phenomemex) was used with an injection volume of 20 µl, with 5 mM ammonium acetate in water (phase A) and 5 mM ammonium acetate in methanol (phase B). The flow rate used was 0.8 mL min^−1, and the gradient mode started with a ratio of 90:10 (water/methanol), increasing to 5:95 at 4 min and returning to the initial condition at 10 min. The run time was 15 min. The electrospray ionization source was used in negative and positive modes. Eight concentrations of the analytical standards of each compound were included in the calibration curve. For each of the compounds, the analytical curve, linearity, limit of detection and quantification, precision (repeatability and intermediate precision) and accuracy were determined^49,50.

Data analysis

The control data were subjected to analysis of variance using the F test (α ≤ 0.05), and when significant, the t test (LSD) (α ≤ 0.05) was applied with Sisvar^® statistical analysis software (Version 5.7 – Build 91, Brazil) to compare the means⁵¹.

For the analysis of the data obtained from the water samples, the international toxicological standards available in the Pesticide Properties Database (PPDB) were used²¹. For each collection time, the highest concentration of each herbicide was selected, added to its metabolites when present, and compared with the concentrations considered safe for the organisms from the most sensitive aquatic environment based on ecotoxicological studies of each herbicide (Table 1)²¹.

Results and discussion

Phytosociological survey and weed control

Approximately 20 species of naturally occurring weeds were identified throughout the experimental area (Table 4). The similarity index between the center of the railway line and the sides was 66.7%. This moderate overlap indicates both uniformity and local variation in species distribution, likely influenced by microenvironmental conditions and disturbance gradients. According to the phytosociological indicators (frequency, density, dominance and importance value index), the species Commelina communis, Paspalum compressum, Glycine max, Commelina diffuse and Talinum paniculatum were predominant in the center of the railway line (Table 4). On the sides of the line, P. compressum, C. communis, Digitaria horizontalis, G. max, and Echinochloa colonum were predominant (Table 4).

Table 4.

Origin of species, relative (RF) and absolute frequency (AF), relative (RD) and absolute (DA) density, relative dominance, importance value index (IVI) and dry mass of weeds collected in the center and along the sides of the railway track.

Species	Origin	Frequency		Density		Relative dominance	IVI	Dry mass (g)
		RF	AF	RD	AD
Center of the railway line
Cyathula prostrata	Exotic	4.2	20	0.7	0.8	1.3	6.2	8.2
Cyperus esculentus	Exotic	4.2	20	0.7	0.8	1.3	6.2	3.4
Eupatorium pauciflorum	Pantropical	4.2	20	1.1	1.2	2.0	7.2	3.5
Galinsoga parviflora	Exotic	4.2	20	1.1	1.2	2.0	7.2	7.5
Borreria palustris	-	4.2	20	1.4	1.6	2.7	8.3	3.9
Digitaria sanguinalis	Exotic	4.2	20	1.8	2.0	3.3	9.3	3.6
Setaria parviflora	Exotic	4.2	20	2.5	2.8	4.7	11.3	4.1
Drymaria cordata	Exotic	4.2	20	2.9	3.2	5.3	12.4	6.3
Alternanthera tenella	Exotic	8.3	40	2.5	2.8	2.3	13.2	5.5
Digitaria horizontalis	Exotic	8.3	40	2.9	3.2	2.7	13.9	4.2
Cenchrus echinatus	Exotic	4.2	20	3.6	4.0	6.7	14.4	7.7
Talinum paniculatum	-	4.2	20	5.4	6.0	10.0	19.5	10.6
Commelina diffusa	-	4.2	20	6.4	7.2	12.0	22.6	6.3
Glycine max	Exotic	16.7	80	8.9	10.0	4.2	29.8	5.4
Paspalum compressum	Exotic	12.5	60	13.6	15.5	8.5	34.5	9.4
Commelina communis	-	8.3	40	16.8	18.8	15.7	40.8	5.8
Side of the railway line
Chamaesyce hyssopifolia	Exotic	4.2	20	0.7	0.4	1.0	5.9	3.3
Cyperus esculentus	Exotic	4.2	20	1.4	0.8	2.0	7.6	3.5
Borreria marshland	-	4.2	20	2.1	2.1	3.0	9.3	10.4
Digitaria sanguinalis	Exotic	4.2	20	2.1	1.2	3.0	9.3	4.0
Eleusine indica	Exotic	4.2	20	2.8	1.6	4.0	11.0	4.5
Eupatorium pauciflorum	Pantropical	8.3	40	3.5	2	2.5	14.3	3.8
Alternanthera tenella	Exotic	4.2	20	4.9	2.8	7.0	16.0	4.0
Commelina diffuse	-	4.2	20	4.9	2.8	7.0	16.0	4.2
Spigelia anthelmia	Pantropical	4.2	20	4.9	2.8	7.0	16.0	5.1
Galinsoga parviflora	Exotic	8.3	40	3.5	2	5.0	16.8	5.1
Kyllinga brevifolia	Exotic	4.2	20	5.6	3.2	8.0	17.7	4.5
Commelina communis	-	4.2	20	8.4	4.8	12.0	24.5	11.2
Echinochloa colonum	Exotic	4.2	20	9.8	5.6	14.0	27.9	3.9
Glycine max	Exotic	16.7	80	9.1	5.2	3.2	29.0	4.8
Digitaria horizontalis	Exotic	8.3	40	17.5	10	12.5	38.3	5.7
Paspalum compressum	Exotic	12.5	60	18.9	1.8	9.0	40.4	9.8

The weed species found in the center and on the sides of the railway line are exotic (Table 4) and pose a risk to the biodiversity of the Atlantic Forest, reinforcing the need for monitoring and control. The transportation of agricultural crops by rail within the forest reserve promotes the dispersion of weed species and seeds of cultivated plants, all of which are exotic and pose a risk to the Atlantic Forest environment. In a study evaluating the influence of railways on the entry of exotic species into Turkey⁵², found a direct effect from species with high reproductive and adaptive potential and reinforced the need for regular inspection, monitoring and controls. Therefore, phytosociological analysis is not only important for quantifying weed pressure but also critical for identifying invasive threats to native ecosystems. From a conservation standpoint, this floristic profile within a fully protected unit underscores the need to limit propagule pressure from the corridor into adjacent native fragments, reinforcing systematic monitoring and targeted suppression of high-risk exotics.

Phytosociology is a set of ecological evaluation methods that aim to provide a comprehensive view of the distribution of plant species in a given environment, establishing qualitative (relationship between species in the study area) and quantitative (number of individuals and weed density) characteristics per m²^53,54. For phytosociological studies of weeds, it is recommended to follow some steps to obtain a more robust analysis: assessment of the general infestation; development of phytosociological tables or graphs; characterization of the diversity of the site; and characterization and grouping by similarity relative to species outside of the site⁵⁴.

The combination of pre- and post-emergence herbicides provided high levels of weed control without the need for frequent reapplications. With this type of application, a high percentage of control was obtained from the beginning of the evaluations (15 DAA) due to the action of the herbicides glyphosate and saflufenacil in controlling already established plants and because the herbicides with preemergence action (imazapyr and indaziflam) were effective in keeping the soil seed bank under control for a long period (180 DAA) (Table 5). However, when analyzing the isolated application of these herbicides, it was observed that in all treatments in which the herbicides with exclusive postemergence action (glyphosate and saflufenacil) were applied, lower efficacy was observed due to the high seed flow in the area and the presence of glyphosate herbicide-resistant and glyphosate herbicide-tolerant species such as G. max and Commelina spp., respectively (Table 3). Beyond operational safety, sustained suppression curtails seed rain and seed-bank replenishment along the right-of-way, which directly supports conservation outcomes by reducing the outward spread of exotics from the corridor into surrounding forest areas.

Table 5.

General weed control at 15, 30, 60, 90, 120, 150 and 180 days after application of the herbicides glyphosate (Gly), Imazapyr (Imaz), Indaziflam (Ind) and Saflufenacil (Saflu).

Treatment	Days after application (DAA)
	15	30	60	90	120	150	180
Glyphosate (720)	70 Ca	55 Eb	25 Cc	15 Dd	7 Fe	0 Gf	0 Gf
Glyphosate (1440)	60 Db	75 Ca	35 Bc	15 Dd	14 Ed	0 Ge	0 Ge
Gly (720) + Imaz (2500)	90 Bb	99 Aa	100 Aa	99 Aa	88 Bbc	86 CDc	88 Cbc
Imazapyr (1500)	40 Fe	65 Dd	99 Aa	91 Bb	88 Bb	75 Ec	74 Ec
Imazapyr (2500)	50 Ee	88 Bc	99 Aa	99 Aa	92 Bb	82 Dd	82 Dd
Ind (125) + Gly (1440)	95 Abc	99 Aab	100 Aa	100 Aa	100 Aa	92 Bc	92 Bc
Ind (200) + Gly (1440)	90 Bb	100 Aa	100 Aa	99 Aa	100 Aa	76 Ec	76 Ec
Imaz (1500) + Ind (125)	35 Gc	100 Aa	100 Aa	100 Aa	100 Aa	88 Cb	88 Cb
Imaz (2500) + Ind (200)	70 Cb	100 Aa	100 Aa	100 Aa	100 Aa	98 Aa	98 Aa
Saflufenacil (180)	40 Fc	5 Ge	5 De	88 Ba	70 Db	9 Fd	7 Fde
Saflufenacil (390)	50 Eb	15 Fc	5 Dd	81 Ca	80 Ca	7 Fd	7 Fd
Imaz (1500) + Saflu (180)	50 Eb	99 Aa	100 Aa	100 Aa	98 Aa	98 Aa	97 Aa
Imaz (1500) + Saflu (390)	70 Cb	100 Aa	100 Aa	100 Aa	100 Aa	99 Aa	100 Aa

Values in parentheses refer to the concentration of the active ingredient (g ha⁻¹). Control means followed by the same letter, uppercase by column and lowercase by row, did not differ statistically according to the t test (LSD) (α ≤ 0.05).

The species G. max, Commelina spp, Digitaria spp and Paspalum compressum followed the same trend of general control up to 180 DAA, with the highest levels of efficacy observed in treatments with the combined application of postemergence and preemergence herbicides (Tables 6 and 7). In all the herbicide mixtures analyzed, control was maintained at 100% up to 120 DAA for the four species studied, except for Digitaria spp., for which high control was observed up to 180 DAA. For the species G. max and P. compressum, 100% control was maintained until the end of the evaluations with the mixtures of imazapyr (1500) with saflufenacil (180 and 390), imazapyr (2500) with indaziflam (200) and indaziflam (125) with glyphosate (1440) (Table 6). For Commelina spp., 100% control was maintained until 180 DAA in mixtures of imazapyr (1500) with saflufenacil (180 and 390) and imazapyr (2500) + indaziflam (200) (Table 7). These results highlight the importance of herbicides used for weed management along railways, especially imazapyr as a preemergence option.

Table 6.

Control of Glycine max and Paspalum compressum after application of the herbicides glyphosate (Gly), Imazapyr (Imaz), Indaziflam (Ind) and Saflufenacil (Saflu) at 15, 30, 60, 90 and 120, 150 and 180 days after application (DAA).

Treatment	Days after application (DAA)
	15	30	60	90	120	150	180
	Glycine max
Glyphosate (720)	15 Fa	0 Eb	3 Cb	3 Cb	7 Cab	0 Hb	0 Gb
Glyphosate (1440)	0 EDb	0 Eb	0 Cb	3 Cb	15 Ca	0 Hb	0 Gb
Gly (720) + Imaz (2500)	80 Ac	100 Aa	100 Aa	99 Aa	97 Aab	90 BCDb	90 BCDb
Imazapyr (1500)	20 De	50 Cd	90 Bb	99 Aa	97 Aab	72 Fc	72 Fc
Imazapyr (2500)	50 Bc	85 Bb	100Aa	99 Aa	99 Aa	82 DEb	82 DEb
Ind (125) + Gly (1440)	80 Ab	97 Aa	100 Aa	99 Aa	100 Aa	92 ABCa	93 ABCa
Ind (200) + Gly (1440)	20 Dc	100ª	100 Aa	99 Aa	100 Aa	76 EFb	76 EFb
Imaz (1500) + Ind (125)	10 Ec	100ª	100 Aa	100 Aa	100 Aa	87 CDc	87 CDc
Imaz (2500) + Ind (200)	20 Db	100 Aa	100 Aa	100 Aa	100 Aa	97 ABa	97 ABa
Saflufenacil (180)	10 Eb	5 Ebc	1 Cc	86 Ba	88 Ba	7 HGbc	7 Gbc
Saflufenacil (390)	30 Eb	15 Dc	4 Cd	86 Ba	85 Ba	9 Gbc	7 Gbc
Imaz (1500) + Saflu (180)	80 Ab	99 Aa	100 Aa	100 Aa	100 Aa	99 Aa	97 ABa
Imaz (1500) + Saflu (390)	54 Bb	100 Aa	100 Aa	100 Aa	100 Aa	100 Aa	100 Aa
	Paspalum spp
Glyphosate (720)	70 Ca	67 Ca	5 Cb	7 Db	7 Db	0 Gb	0 Fb
Glyphosate (1440)	80 Ba	85 Ba	32 Bb	7 Dcd	12 Dc	0 Gd	0 Fd
Gly (720) + Imaz (2500)	95 ABa	100 Aa	100 Aa	99 Aa	100 Aa	85 BCDb	85 BCDb
Imazapyr (1500)	30 Ec	98 Aa	100 Aa	76 Cb	100 Aa	78 DEb	78 DEb
Imazapyr (2500)	40 Dc	100 Aa	100 Aa	99 Aa	100 Aa	82 CDEb	82 DECb
Ind (125) + Gly (1440)	75 BCb	100 Aa	100 Aa	100 Aa	100 Aa	92 ABa	92 ABa
Ind (200) + Gly (1440)	80 Bb	100 Aa	100 Aa	100 Aa	100 Aa	76 Eb	76 Eb
Imaz (1500) + Ind (125)	20 Fc	100 Aa	100 Aa	100 Aa	100 Aa	87 BCb	87 BCb
Imaz (2500) + Ind (200)	80 Bb	100 Aa	100 Aa	100 Aa	100 Aa	97 Aa	98 Aa
Saflufenacil (180)	20 Fc	5 Ede	2 Ce	86 Ba	65 Bb	11 Fd	7 Fde
Saflufenacil (390)	30 Ec	14 Dd	2 Ce	85 Ba	75 Cb	7 FGde	7 Fde
Imaz (1500) + Saflu (180)	40 Db	99 Aa	100 Aa	100 Aa	93 Aa	99 Aa	98 Aa
Imaz (1500) + Saflu (390)	70 Cb	100 Aa	100 Aa	100 Aa	100 Aa	100 Aa	100 Aa

Values in parentheses refer to the concentration of the active ingredient (g ha⁻¹). Control means followed by the same letter, uppercase by column and lowercase by row, did not differ statistically according to the t test (LSD) (α ≤ 0.05).

Table 7.

Control of Commelina spp. And Digitaria spp. Along railways after application of the herbicides glyphosate (Gly), Imazapyr (Imaz), Indaziflam (Ind) And Saflufenacil (Saflu) at 15, 30, 60, 90, 120, 150 And 180 days after application (DAA).

Treatment	Days after application (DAA)
	15	30	60	90	120	150	180
	Commelina spp
Glyphosate (720)	10 Hc	25 Db	32 Ca	2 Cd	7 Ec	0 Gd	0 Gd
Glyphosate (1440)	10 Hc	70 Ca	15 Db	2 Cd	14 Db	0 Gd	0 Gd
Gly (720) + Imaz (2500)	80 Ac	100 Aa	100 Aa	99 Aa	100 Aa	92 Bb	92 Bb
Imazapyr (1500)	30 Gd	89 B	100 Aa	99 Aa	100 Aa	74 Ec	74 Ec
Imazapyr (2500)	40 Fc	100 Aa	99 Aa	99 Aa	100 Aa	82 Db	82 Db
Ind (125) + Gly (1440)	65 Cc	100 Aa	100 Aa	99 Aa	100 Aa	92 Bb	92 Bb
Ind (200) + Gly (1440)	40 Fc	100 Aa	100 Aa	100 Aa	100 Aa	76 Eb	76 Eb
Imaz (1500) + Ind (125)	50 Ec	100 Aa	100 Aa	100 Aa	100 Aa	88 Cb	88 Cb
Imaz (2500) + Ind (200)	70 Bb	100 Aa	100 Aa	100 Aa	100 Aa	98 Aa	98 Aa
Saflufenacil (180)	30 Gc	5 Fd	1 De	84 Ba	65 Cb	7 Fd	7 Fd
Saflufenacil (390)	80 Aa	14 Ec	1 De	81 Ba	75 Bb	7 Fd	7 Fd
Imaz (1500) + Saflu (180)	50 Eb	100 Aa	100 Aa	100 Aa	100 Aa	99 Aa	98a A
Imaz (1500) + Saflu (390)	60 Db	100 Aa	100 Aa	100 Aa	100 Aa	100 Aa	100 Aa
	Digitaria spp
Glyphosate (720)	80 Ba	25 Db	15 Bc	2 Cd	2 Dd	2 Dd	0 Dd
Glyphosate (1440)	75 Cb	80 Ba	15 Bc	2 Cd	2 Dd	2 Dd	0 Dd
Gly (720) + Imaz (2500)	95 Ab	100 Aa	100 Aa	99 Aa	100 Aa	100 Aa	100 Aa
Imazapyr (1500)	30 Dc	76 Cb	100 Aa	99 Aa	100 Aa	100 Aa	100 Aa
Imazapyr (1500)	80 Bb	99 Aa	100 Aa	99 Aa	100 Aa	100 Aa	100 Aa
Ind (125) + Gly (1440)	80 Bb	100 Aa	100 Aa	100 Aa	100 Aa	100 Aa	100 Aa
Ind (200) + Gly (1440)	80 Bb	100 Aa	100 Aa	100 Aa	100 Aa	100 Aa	100 Aa
Imaz (1500) + Ind (125)	20 Eb	100 Aa	100 Aa	100 Aa	100 Aa	100 Aa	100 Aa
Imaz (2500) + Ind (200)	15 Fb	100 Aa	100 Aa	100 Aa	100 Aa	100 Aa	100 Aa
Saflufenacil (180)	20 Ee	5 Ef	4 Cf	84 Ba	75 bB	62 Cc	39 Cd
Saflufenacil (390)	30 Dd	14 Fe	2 Cf	85 Ba	71 Cb	69 Bb	54 Bc
Imaz (1500) + Saflu (180)	30 Db	100 Aa	100 Aa	100 Aa	100 Aa	100 Aa	100 Aa
Imaz (1500) + Saflu (390)	80 Bb	100 Aa	100 Aa	100 Aa	100 Aa	100 Aa	100 Aa

Values in parentheses refer to the concentration of the active ingredient (g ha⁻¹). Control means followed by the same letter, uppercase by column and lowercase by row, did not differ statistically according to the t test (LSD) (α ≤ 0.05).

The high efficacy of integrated herbicide applications also underscores the importance of using products with different modes of action to delay the development of herbicide resistance. The presence of glyphosate resistant and tolerant species such as Glycine max and Commelina spp. in the study area reflects a growing challenge in weed management along disturbed corridors. By combining preemergence and postemergence herbicides, the weed seed bank is depleted over time while emerged individuals are effectively controlled, resulting in prolonged weed suppression and reduced selection pressure for resistant biotypes.

Analysis of herbicides and some metabolites in water and soil

The analysis of herbicide concentrations in water reveals that the levels detected are much lower than the concentrations to which the test organisms can be exposed without adverse effects (NOEC) (Table 8). To provide a conservative, discussion-level frame of reference appropriate to a Protected Area, water concentrations at each timepoint were interpreted against PPDB benchmarks by comparing the maximum measured concentration of each herbicide, summed with detected metabolites when present, to the lowest chronic aquatic NOEC reported for the most sensitive standard taxon (Table 1). During the tests, 152 water samples were collected, and imazapyr was the active ingredient detected in the largest number of samples (77%); however, the maximum sampled value (17.3 ng mL⁻¹) occurred 15 days after application, which represents only 0.04% of the NOEC limit for the most sensitive test organism (Tables 1 and 8). The high number of samples in which herbicides were detected is related to the large number of treatments containing the active ingredient and consequently the larger area treated with the product. After 15 DAA, collections were performed close to the application date, and the concentrations of this herbicide were less than 0.01% of the NOEC limit (Fig. 3).

Table 8.

Number and percentage of samples with detection, maximum concentration of the herbicides glyphosate, imazapyr, Indaziflam and Saflufenacil and some of their metabolites in water and the relationship between the maximum concentration and NOEC limits.

Herbicide	Number of samples with detection	Percentage of samples with detection	Maximum concentration in water	NOEC limita	Percentage of the NOEC limit
		%	ng mL−1	ng mL−1	%
Glyphosate	35	23,0	5,610	1000	0,56
Imazapyr	117	77,0	17,300	43,100	0,04
Indaziflam	36	23,7	1,160	340	0,34
Saflufenacil	9	5,9	0,772	997	0,08
AMPA b	0	0,0	-	-	-
FDAT c	9	5,9	0,210	-	-

^aNOEC limit for the most sensitive aquatic organism; ^baminomethylphosphonic acid – AMPA (glyphosate metabolite); ^cfluoroethyldiaminotriazine – FDAT (indaziflam metabolite).

Fig. 3.

Percentage of the highest concentration in the water at each collection time of the herbicides imazapyr, indaziflam (and FDAT metabolite), saflufenacil and glyphosate relative to the NOEC of the most sensitive aquatic organisms for each herbicide (Table 1).

The highest concentration of glyphosate detected represents 0.56% of the NOEC limit. Glyphosate is the compound with the highest percentage observed, followed by indaziflam (0.34%) and saflufenacil (0.08%). At 180 DAA, the maximum concentrations of glyphosate and indaziflam corresponded to 0.09 and 0.06%, respectively, of the NOEC limit for the most sensitive test organism. When combined with its metabolites, saflufenacil was detected up to 150 DAA at a concentration of 0.001% (Fig. 3).

Regarding the metabolites of the respective herbicides, AMPA (glyphosate metabolite) was not detected, and FDAT (indaziflam metabolite) was detected in approximately 6% of the samples. (Table 8).

The highest concentrations of the analyzed compounds in water were generally detected in standing water (Fig. 4), pointing to microhabitats with longer residence time and, potentially, higher exposure for early life stages. This pattern highlights practical mitigation in a Protected Area: maintaining drainage, avoiding depressions that retain runoff, and scheduling applications outside periods of intense rainfall to limit pulse exposures.

Fig. 4.

Concentration in standing and running water of the herbicides glyphosate, imazapyr, saflufenacil, indaziflam and its metabolite fluoroethyldiaminotriazine (FDAT).

For glyphosate, the maximum concentration was observed at 90 DAA, in both running and standing water (Fig. 4). In a study analyzing the environmental fate of glyphosate used in Swedish railways, the authors concluded that glyphosate can reach shallow groundwater directly below or in the vicinity of the track where it was applied. Monitoring stations bordering the railway indicate that horizontal mobility in the groundwater zone appears to be limited, and this in combination with the no-spray zones that surround the groundwater bodies indicates that groundwater resources are unlikely to be affected by leaching⁵⁵. In general, glyphosate is expected to be leached easily from the upper ballast of the railway, which consists of gravel, but to be quite immobile in the materials that compose the lower and subballast layers⁵⁵.

For the concentration of the herbicide indaziflam in water, the highest levels were observed at 15 DAA in standing water, with a reduction throughout the evaluations (Fig. 4). The herbicide indaziflam, as well as its metabolite FDAT, had very low levels in running water, with higher concentrations of indaziflam at 15 DAA and of FDAT at 60 DAA, both in standing water (Fig. 4). The herbicide saflufenacil was detected in water up to 60 DAA, predominantly in still water. In a study evaluating the use of aquatic macrophytes in the decontamination of water contaminated by saflufenacil, it was found that in the absence of these organisms, the half-life of the compound in water was 72.2 days, and in the presence of these organisms, there was a reduction in DT50 of up to 94.8%⁵⁶.

The low concentrations of herbicides detected in water throughout the study demonstrate that environmental risk was minimal, even in a sensitive area like the Atlantic Forest. The maximum values recorded for glyphosate, imazapyr, indaziflam, and saflufenacil remained well below the NOEC thresholds for aquatic organisms, reinforcing the environmental safety of these products. However, the more frequent detection of residues in standing water, compared to flowing water, suggests that low-flow environments may be more vulnerable to localized accumulation, with potential ecological implications for less sensitive aquatic species, this pattern reinforces the need for special attention to drainage structures and water retention zones along railways. In practice, maintaining low connectivity between retained waters and adjacent streams, together with routine clearing of culverts and side ditches, helps keep exposures well below conservative benchmarks in a State Park setting.

There was no significant movement of the herbicides in the soil outside the application area, as the compounds were detected only in the areas treated with each herbicide and on the sides of the respective plots. Indeed, even in the application areas, the values were well below the thresholds (Figs. 5, 6 and 7), which demonstrates that after application, the herbicides were confined within the application area. The concentrations to the sides of the application areas can be explained by a possible deposition of the spray solution by the drift process.

Fig. 5.

Concentration of imazapyr within the applied area, regardless of treatment, at 15, 30, 60, 90, 120, 150 and 180 days after application (DAA) and around the application areas at 15, 150 and 180 DAA.

Fig. 6.

Concentration of indaziflam and its metabolite (fluoroethyldiaminotriazine, FDAT) within the applied area, regardless of treatment, at 15, 30, 60, 90, 120, 150 and 180 days after application (DAA) and on the sides of the applied areas at 15, 150 and 180 DAA.

Fig. 7.

Concentration of saflufenacil within the application area, regardless of treatment, at 15, 30, 60, 90, 120, 150 and 180 days after application (DAA) and to the sides of the application areas at 15, 150 and 180 DA.

The herbicide glyphosate and its metabolite (AMPA) were not detected in soil. Thus, glyphosate may have been degraded and/or strongly sorbed under the conditions under which the study was conducted. Cederlund (2022)⁵⁵ did not detect glyphosate in the matrix of a rail bed and attributed these results to possible differences in the Koc and/or the sorption properties of the compounds in addition to the degradation of the product. In a study evaluating the influence of organic matter on glyphosate sorption and desorption in soils with different mineralogical attributes, Prata (2002)⁵⁷ highlighted that the herbicide was sorbed to a very large extent into the soils, regardless of the effect of organic matter, and that there was no desorption, with glyphosate remaining in the soils as a bound residue. Silva (2019)¹³, when evaluating glyphosate leaching in soils of permanent preservation areas crossed by a railway line, did not observe herbicide or AMPA, and there were no symptoms of exposure in bioindicator plants.

The highest concentration peak for imazapyr (275.6 ng g⁻¹ soil) was observed 15 DAA within the area treated with the isolated dose of 2.5 L a.i. ha⁻¹; at 180 DAA for the same treatment, the concentration was 33.9 ng g⁻¹ of soil, which represents 12.3% of the maximum detected amount. Outside the treated area, the largest imazapyr detection amount (26.6 ng g⁻¹ soil) also occurred in the isolated application with a dose of 2.5 L a.i. ha⁻¹ at 150 DAA (Fig. 5).

For indaziflam, the maximum concentration in soil (0.59 ng g⁻¹ soil) occurred within the area applied at the dose of 200 mL a.i. ha⁻¹ mixed with glyphosate (1440 g aa ha⁻¹) at 15 DAA. Over time, the concentration dropped to 0.31 ng g⁻¹ soil at 180 DAA, which represents 52.5% of the concentration observed shortly after application (Fig. 6). The smaller variation between the periods with higher detection and the concentrations reported in the latest evaluations (150 and 180 DAA) are related to the physicochemical characteristics of indaziflam, which has a long residual period, low solubility and low mobility in soil, which makes it poorly extractable with water and consequently not easily desorbable and leachable in the soil^21,31.

While the maximum concentration of indaziflam within the applied area was 0.59 ng g⁻¹ soil, for its metabolite (fluoroethyldiaminotriazine - FDAT), the highest peak was 1.19 ng g⁻¹ soil (Fig. 6). To the sides of the application areas, the little product detected was observed in greater amounts at 150 DAA and followed the same behavior as the product inside the application area, with a higher concentration of FDAT than indaziflam (Fig. 6). The degradation process of indaziflam and its transformation into its metabolites favored the high concentration of fluoroethyldiaminotriazine in soil (Fig. 6), which is also related to the fact that the metabolite has low sorption and is considered to be mobile to highly mobile in soil, with higher potential of availability in the soil solution, as it is readily desorbable throughout its persistence time^30–32,58.

The sorption of indaziflam and its metabolites is affected by soil properties and depends on the physicochemical properties of each compound. The metabolites are more polar than the parent compound and have lower sorption in the order of indaziflam > triazinesindanone > fluoroethyldiaminotriazine = indaziflam carboxylic acid³². Although FDAT has a higher risk of leaching because it remains less sorbed, its ecotoxicity indices, which represent the potential for contamination of the environment, are lower than those of indaziflam^30,31,33.

For saflufenacil, the highest herbicide concentration (72.86 ng g⁻¹ soil) within the treated area was detected at 15 DAA at a dose of 390 g a.i. ha⁻¹. In the second collection period, at 30 DAA, there was a reduction of 85% of the initial concentration, reaching a reduction of 99.3% by 180 DAA. To the sides of the application areas, the maximum detected was 0.226 ng g⁻¹ soil, which represents 0.3% of the maximum of the detected product (Fig. 7).

In a study evaluating the sorption-desorption and dissipation rates of saflufenacil in surface and subsurface soils, Papiernik et al. (2012)⁵⁹ found low sorption and relatively rapid dissipation, suggesting that this herbicide is readily available for degradation or uptake by plants in the root zone. Matallo et al. (2014)⁶⁰ corroborates this information in a study with two soil types with contrasting physicochemical attributes, where it was concluded that the low sorption constants indicate poor sorption of saflufenacil to soil colloids, predisposing the molecule to release into soil solution. These authors, as well as Monquero et al. (2012)⁶¹ highlight that the occurrence of rainfall shortly after saflufenacil application in soils may contribute to leaching.

Within Brazil’s National System of Protected Areas, only indirect use of natural resources is permitted; activities in pre-existing transport corridors, such as the licensed railway right-of-way, are governed by the Park’s management plan and environmental licensing⁶². Non-agricultural herbicide used for infrastructure vegetation control can be authorized when compliant with zoning and permit conditions⁶³. In this context, the low environmental exposures observed here, combined with drainage upkeep, drift control, buffer observance, and timing relative to rainfall, support practical, legally compliant vegetation management that aligns operational safety with conservation mandates.

Taken together, the very low fractions of the most sensitive NOECs (typically ≪1%) in surface waters, the spatial confinement and decline observed in soils, and the ecological benefits of reducing exotic propagule pressure indicate low environmental risk under the tested conditions in a fully protected unit. Continued attention to right-of-way hydrology and edge habitats, coupled with rotation of complementary modes of action, provides a pathway to sustain control while safeguarding Atlantic Forest ecosystems within the PESM.

These findings support the ecological compatibility of herbicide use under the tested conditions but also underscore the need for risk assessments that incorporate metabolite behavior, environmental heterogeneity, and potential sublethal effects on non-target species. Incorporating such dimensions can guide more nuanced and site-specific herbicide management strategies in conservation areas.

Conclusions

The weed composition present in the study area indicates the entry of exotic species that can compromise the establishment of native biodiversity along the banks of the railway line within the Mata Atlantica reserve. The analysis of the control information indicated that it is essential to apply herbicides or treatments that combine postemergence and preemergence actions, preventing rapid reinfestation in the area. The use of the herbicides imazapyr, glyphosate, indaziflam and saflufenacil under the evaluated conditions was safe, with the products being dissipated in the soil without promoting contamination of the surroundings of the application areas. These results support the viability of targeted chemical control as a safe and efficient strategy for vegetation management in ecologically sensitive railway environments, contributing to both operational safety and environmental conservation.

As to risk interpretation and study scope, two clarifications help frame these findings. Aquatic results were interpreted using literature NOECs as contextual benchmarks, offering a conservative reference suitable for a Protected Area. Monitoring focused on soil and surface water, with biota and sediment/periphyton outside the present scope. Spray drift was minimized operationally; incorporating quantitative drift measurements and, where relevant, microbial functional assays and additional sublethal endpoints in multi-season, multi-site settings would further refine exposure estimates and support continued optimization of herbicide portfolios and application windows in Protected Areas.

Author contributions

Conceptualization: Carbonari, Simões and Velini; data curation: Simões and Costa; identification of weeds: Simões and Costa; formal analysis: Carbonari, Simões and Costa; investigation: Carbonari, Simões, Pitelli and Velini; methodology development: Carbonari, Simões, Pitelli and Velini; supervision: Carbonari and Velini; writing—original draft: Simões and Costa; writing—review and editing: all authors.

Funding

No funding was received to assist with the preparation of this manuscript.

Data availability

The datasets used and analyzed during the current study are available from the corresponding author upon request.

Declarations

Competing interests

The authors declare no competing interests.