Agricultural weeds, particularly those with high seed production like *Amaranthus palmeri*, rapidly adapt to herbicides, evolving resistance mechanisms. These mechanisms are multifaceted and can involve target-site (TS) mutations (e.g., in the 5-enolpyruvylshikimate 3-phosphate synthase (EPSPS) gene for glyphosate resistance) and non-target-site resistance (NTSR), including gene duplication/overexpression and altered herbicide uptake/translocation. High herbicide doses typically select for TS mutations, while low doses may favor NTSR. The diversity of weeds and resistance mechanisms, coupled with complex human interventions, makes long-term weed control challenging. Population models offer valuable tools to predict the evolutionary dynamics of weed populations under various management strategies, but their accuracy hinges on detailed knowledge of weed biology and resistance mechanisms, and the ability to model spatio-temporal variations. Existing models frequently focus on theoretical principles rather than providing specific, field-applicable recommendations due to incomplete understanding of the factors involved. This study aims to address these limitations by integrating various research aspects into a comprehensive approach to manage *Amaranthus palmeri* resistance to glyphosate in Argentinian soybean fields.

Previous research on herbicide resistance has often focused on specific aspects: field investigation of resistance prevalence, glasshouse/laboratory confirmation of resistance, genetic/molecular analysis of resistance mechanisms, exploration of alternative control methods, and long-term sustainability evaluation via population models. This fragmented approach lacks the holistic perspective needed for effective, timely resistance management. Studies on glyphosate resistance in *A. palmeri* have revealed diverse mechanisms, including reduced uptake, EPSPS gene amplification, and P106S target-site mutations. The complexity of these mechanisms, combined with the influence of varied weed control practices, makes accurate prediction of long-term control sustainability difficult. Therefore, the current research adopts a novel integrated approach that encompasses all these aspects to provide a more comprehensive understanding and develop robust, sustainable control strategies.

The study utilized *A. palmeri* samples from a glyphosate-treated soybean field in Villa Valeria, Cordoba, Argentina, along with standard sensitive (ApS1 and ApS2) and resistant (ApR) populations. Resistance confirmation employed agar-based RISQ (Resistance In-Season Quick) and whole plant pot tests with varying glyphosate concentrations. The mechanism of resistance was investigated using biokinetic experiments with ¹⁴C-glyphosate to assess uptake and translocation. Quantitative real-time PCR (qPCR) determined EPSPS gene copy number and expression levels, while Sanger sequencing identified EPSPS gene mutations. An individual-based population model, modified to represent the *A. palmeri* population's characteristics (emergence, seed production, mortality rates), simulated the long-term effects of various herbicide programs, including glyphosate alone and combinations with alternative herbicides such as fomesafen, lactofen, S-metolachlor, and metribuzin, considering the initial frequency of both target-site and quantitative resistance mechanisms. The model incorporated parameters like herbicide application rates, timing, and residual activity duration to evaluate weed density and resistance evolution over a 20-year period.

The *A. palmeri* population from Villa Valeria (VV) exhibited resistance to glyphosate, confirmed by both RISQ and whole-plant pot tests. Biokinetic experiments showed no significant differences in glyphosate uptake between VV and ApS1 populations, indicating that resistance was not due to reduced uptake. qPCR analysis revealed a modest (~1.5-fold) increase in EPSPS gene copy number and expression in the VV population compared to ApS2, significantly less than the increase observed in the highly resistant ApR population (~9.0-fold). Sanger sequencing identified a P106S target-site mutation in the EPSPS gene in the VV population, with genotype frequencies of 6% (PP106), 22% (PS106), and 72% (SS106), indicating the prevalent role of this mutation. The population model simulations predicted that with glyphosate as a solo treatment, weed density exceeded the control threshold in approximately 9.7 years when both TS mutation and quantitative resistance were present. TS mutation-based resistance dominated, reaching >20% in ~7.4 years, while quantitative resistance remained very low (<0.04%). The model demonstrated that a single herbicide treatment primarily selects for a single resistance mechanism, depending on the initial frequency of available mechanisms. Alternative herbicides, such as fomesafen and lactofen, effectively controlled the glyphosate-resistant VV population. The model also demonstrated that programs incorporating chemical diversity and residual herbicides, such as those combining fomesafen or lactofen with S-metolachlor and/or metribuzin, were crucial for sustainable PPO-herbicide use. Programs that used only a single mode of action (even with multiple applications) led to rapid resistance evolution and control failure.

This holistic study successfully identified the mechanisms driving glyphosate resistance in an Argentinian *A. palmeri* population, contrasting with populations from the USA where gene duplication/overexpression typically dominates. The dominance of the P106S target-site mutation in this case highlights the influence of differing weed management practices. The model simulations confirmed that the use of a single herbicide favors the selection of a single resistance mechanism, underscoring the importance of herbicide mixtures and residual herbicides in delaying resistance development. The results emphasize the need for region-specific weed management strategies, accounting for local resistance mechanisms and weed control practices. The effectiveness of alternative PPO-inhibiting herbicides highlights their potential role in sustainable weed management, but the study also cautioned that inappropriate use can lead to rapid resistance evolution. The findings demonstrate that even when weed density seems controlled, resistance frequencies in the seedbank might be high, putting future weed management at risk if practices are not optimized.

This research demonstrates the value of a holistic approach to herbicide resistance management, integrating resistance detection, mechanistic studies, and population modeling. The findings highlight the importance of considering regional differences in resistance mechanisms and weed control practices. The effectiveness of alternative herbicides, coupled with the need for chemical diversity and residual activity to maintain sustainable control, are crucial for long-term weed management. Future work should expand the holistic approach by incorporating economic, environmental, and social factors to develop more effective and sustainable weed management strategies.

The study focused on a single *A. palmeri* population from a specific region. The generalizability of the findings to other regions or populations may be limited. While the model incorporated various parameters, it might not account for all factors influencing weed evolution and herbicide resistance. The specific parameters used in the model were drawn from other studies on *A. palmeri* and may not perfectly reflect the specific traits of the Villa Valeria population. The study's scope is limited to glyphosate and PPO inhibitors. Other herbicide modes of action and their associated resistance mechanisms weren't explored.