Crop diversity reduces pesticide use more efficiently with refined diversification strategies

Zhang, Yaoyun; Bohan, David A.; Zhang, Chaochun; Cong, Wen-Feng; Munier-Jolain, Nicolas; Bedoussac, Laurent

doi:10.1038/s43247-025-02418-7

Download PDF

Article
Open access
Published: 13 June 2025

Crop diversity reduces pesticide use more efficiently with refined diversification strategies

Communications Earth & Environment volume 6, Article number: 460 (2025) Cite this article

3699 Accesses
2 Citations
1 Altmetric
Metrics details

Subjects

Abstract

Pesticides assured food security for decades, but have left humanity with degraded soils, polluted water, and biodiversity losses¹. Enhanced crop diversity contributes to the regulation of insect pests, weeds, and diseases^2,3,4, and is therefore assumed to allow pesticide reduction. At the cropping system scale, pesticide use is affected more by crop species than by the number of crops⁵, because crops have contrasting sensitivities to pests and contrasting pesticide requirements. Here, we disentangled the effects of diversity from the effects of crop species, for 1285 cropping systems in French commercial arable farms, using 28 indicators of functional diversity. A composite diversity metric, combining indicators with the greatest explanatory power, accounted for 8% of the pesticide use variance, much less than the variance due to crop species. The results suggest that reducing agricultural reliance on pesticides through diversification is feasible when different components of diversity—namely, crop species and diversity features—are combined.

Fostering temporal crop diversification to reduce pesticide use

Article Open access 16 November 2023

Global evidence that plant diversity suppresses pests and promotes plant performance and crop production

Article 20 January 2026

Acting pre-emptively reduces the long-term costs of managing herbicide resistance

Article Open access 14 March 2024

Introduction

Pesticides are an effective solution to most of the pest outbreaks—except in cases of the development of pest resistance to pesticides, an increasing concern in many regions of the globe. However, the use of pesticides has unacceptable effects on the environment (environmental externalities)¹ and human health⁶. Reducing the reliance of agriculture on pesticides requires a redesign of cropping and farming systems⁷, where crop diversification is regarded as one of the main mechanisms to make this change^2,8,9,10,11. The impact of diversifying arable crops on the regulation of pest population densities and consequently on pesticide use has rarely been assessed to date. Studies have demonstrated decreased pest pressure as an effect of crop diversity^2,3,4, but with little or no evidence that the observed decreases in pest pressure were sufficient to allow farmers to reduce applications of pest control chemicals¹². A few studies conducted on experimental sites or at the landscape level have also related crop diversity to pesticide reduction^5,13,14,15. However, no generic quantification of the potential to decrease pesticide use through crop diversification has been demonstrated at the cropping system level, where cropping system in general refers to a succession of crops grown over a set of fields from a given farm that are managed with the same strategy^5,10,13.

Diversifying a cropping system changes both the diversity and crop species identity simultaneously. Crop species identity affects pesticide use because each crop is subject to a unique suite of pests and levels of infestation that can lead to different amounts of pesticide application to avoid yield loss^16,17,18,19. Some biodiversity studies have tried to disentangle the effects of diversity from the effects of species identity on ecosystem functioning^18,20, always with an experimental approach manipulating both the number and the nature of a few species. Understanding the effects of crop diversity requires that they are first clearly distinguished from the effects of crop species identity. Zhang et al.⁵ demonstrated that the simple number of crops grown, independent of the crop species, is associated with reductions in pesticide use at the cropping system scale. However, crop diversity cannot be limited to only the concept of crop richness. Here, we expanded on the analytical methodology of Zhang et al.⁵ to consider indicators of all components of crop diversity. A total of 28 indicators were used, covering: (i) richness (i.e. number of crops); (ii) evenness²¹; (iii) proportional abundance (i.e. Simpson inverse diversity index^22,23); (iv) dominance²⁴; and, (v) 24 indicators of functional diversity features related to regulation functions of crops regarding pests, weeds, and crop pathogens. The description of these indicators can be found in the ‘Methods’ and Supplementary Table 1. We use data from 1285 cropping systems on 795 non-organic commercial arable farms in the database of the DEPHY network, which is a French national network involving around 2000 volunteer farmers willing to reduce pesticide use at their farms. For each system, there is detailed information on the crops grown and pesticide use over 3 years (see ‘Methods’ for details). Pesticide use at the cropping system level was evaluated using the Treatment Frequency Index (TFI). This indicator quantifies the yearly number of treatments, averaged over the 3-year cropping system period. Each treatment is weighted by the actual dose applied relative to the official reference full dose and by the proportion of the field area treated. TFI is therefore an indicator of the frequency and intensity of resorting to chemical pesticides to solve pest problems. The total TFI value of cropping systems can be decomposed into the four groups of pesticides, namely herbicide TFI, fungicide TFI, insecticide TFI, and other pesticides TFI, which respectively accounted for 53%, 27%, 13% and 7% of the total TFI values, calculated based on the average total TFI value across 1285 cropping systems in the database. To assess the effect of crop diversity on pesticide use at the cropping system level, independent of the crop species grown, we used a specific method developed in Zhang et al.⁵ to remove the effect of crop species. The effect of crop species was represented by the average TFI of each crop species over the whole dataset. This effect was removed from the total cropping system TFI after being weighted by the proportions of crop species within the system (i.e. TFI_diff; see ‘Methods’ for details).

Each of the 28 indicators was initially assessed alone for its correlation with pesticide use at cropping system level, once the effect of crop species was removed (TFI_diff). A decrease in the value of TFI_diff indicates a reduction in TFI after removing the effect of crop species in the cropping system. We then used the Lasso regression method²⁵ to select the combination of indicators for each pesticide group with the greatest explanatory power. The combination of Lasso-selected indicators and their coefficients produced by Lasso led to a group of diversity indicators that linked with pesticide use for each pesticide group. We used this group of indicators as the composite diversity metric (CDM) for each pesticide group and examined their relationships with pesticide use. Lastly, the cropping system TFI variances were partitioned into different sources of variation, namely crop diversity (i.e. CDM), crop species and other factors, to quantify their relative importance.

Results

Individual diversity indicators

Increasing richness (i.e. the number of crops grown in a given cropping system), irrespective of the crop species identity, or the proportional abundance (Simpson inverse diversity index), was correlated with a decrease of the total TFI_diff (Slope = −0.07 and −0.11 respectively, P < 0.001) with only the herbicide group being unaffected (Fig. 1). A similar result was found for increases in evenness (Slope = −0.36, P = 0.033), but with only fungicides unaffected. Conversely, dominance was positively correlated with total TFI_diff (Slope = +0.61, P = 0.001), with only herbicides unaffected.

**Fig. 1: Individual assessments of the relationship between crop species diversity indicators and TFI_diff for each pesticide group.**

Cropping systems with temporary grassland or intercrops (i.e. mixture of crop species in the same field) had significantly lower TFI_diff in all pesticide groups, indicating that other crops were less treated when grown in rotations with temporary grasslands (Fig. 2, box plots). No effect was found for cover crops (i.e. non-harvested crops grown between primary cash crops). Increasing the proportion of temporary grassland, intercrops or cover crops had no effect on total TFI_diff (P = 0.261, 0.438 and 0.690, respectively; Fig. 2 Scatter plots). The increasing proportion of temporary grassland did have an effect on the TFI_diff of the insecticide and other pesticide groups (Slope = +0.10 and +0.07, P = 0.002 and 0.001 respectively).

**Fig. 2: Individual assessments of the relationship between crop functional diversity features and TFI_diff for each pesticide group.**

Increasing the number of crop botanical families had no effect on total TFI_diff (P = 0.526). An increasing proportion of Poaceae and Brassicaceae was associated with an increase in the total TFI_diff (Slope = +0.36 and +3.26; P = 0.007 and P < 0.001, respectively), while increasing the Fabaceae was associated with a decrease in the total TFI_diff (Slope = −1.71, P < 0.001). Other families had no effect (P = 0.071).

Diversifying crops by including a crop with a different sowing season, as compared to the other crops of the cropping systems (e.g. introducing a summer crop into a system dominated by winter-sown crops) was associated with a 0.12 decrease in total TFI_diff (P = 0.001). Including summer- or winter-sown crops was significantly associated with a decrease in the total TFI_diff (P < 0.001), while spring-sown crops were associated with an increase in the total TFI_diff (P = 0.002) (Fig. 2, box plots). Increasing the proportion of spring-, autumn- or winter-sown crops was correlated with an increase in the total TFI_diff (Slope = +1.32, +0.83 and +0.43, P = 0.004, P < 0.001 and P = 0.018, respectively), and the proportion of summer-sown crops had no effect on total TFI_diff (P = 0.392).

The composite diversity metric and its constituent indicators

Once the effect of crop identity was removed, the Lasso method selected 21, 20, 28, 16 and 24 crop diversity indicators for constructing the CDM for the total, herbicide, fungicide, insecticide and other pesticide groups, respectively (Fig. 3). Richness, the proportion of Fabaceae and the proportion of intercrops were the three indicators found to increase the most CDM_total (+2.97, +1.08 and +1.00 unit of the CDM_total, respectively). Conversely, the number of taxonomic families, the proportion of Brassicaceae, and dominance were the three that contributed the most to a decrease in the CDM_total (−2.98, −2.26 and −0.82 units of the CDM_total, respectively). These six indicators affected all components of pesticide use, except that the proportion of Fabaceae, the proportion of intercrops and the number of taxonomic families were not selected for CDM_insecticide, and dominance was not selected for CDM_herbicide.

**Fig. 3: Coefficients and confidence intervals of Lasso-selected indicators that constitute the composite diversity metric (CDM) for each pesticide group.**

The slope and intercept of the regression of total TFI_diff, as a function of CDM_total (Fig. 4), were close to −1.0 and 0.0, respectively. The intercepts for all pesticide groups were not significantly different from zero, with the slopes ranging between −1.1 (CDM_herbicide and CDM_insecticide) and −1.0 (CDM_fungicide and CDM_other), respectively.

**Fig. 4: Regressions of the composite diversity metric (CDM) for each pesticide group.**

Diversity explains 7.9% of the variance

CDM_total explained 7.9% of the variance in total TFI, while 37.1% and 31.9% were respectively explained by crop species and the other factors (i.e. all remaining factors affecting pesticide use except crop diversity and crop species, including production situations and management practices) (Fig. 5). Crop species explained 46.4% of the variance for fungicide TFI, 31.0% for herbicide, 19.2% for insecticide and 8.0% for other pesticides. The CDM for herbicide, fungicide, insecticide and other pesticide groups explained 4.2%, 4.9%, 7.2% and 11.9% of the pesticide variance, respectively. Other factors contributed to 70.5% of the variance of other TFI, 59.6% for insecticide TFI, 50.6% for herbicide TFI and 26.6% for fungicide TFI. The covariance between crop diversity and other factors (cov_diversity_other) remained lower than 1% for each category of pesticides. The covariance between crop species and other factors (cov_species_other) remained very low, with the highest value for fungicide TFI (3.9%). The covariances between crop species and crop diversity were the highest for the total TFI (10.1%).

**Fig. 5: Variances of the cropping system TFI explained by crop species (var_species), crop diversity (var_diversity), the other factors (var_other), and their covariances for each pesticide group.**

Discussion

Our work suggests that increasing crop diversity could allow to reduce the use of pesticide, potentially due to the mitigation of the local pest pressure at the field scale. This result is in line with previous studies on herbivore²⁶ and pathogen load²⁷. However, as we did not directly monitor pest population dynamics in this study, we cannot conclusively demonstrate the underlying biological mechanisms. Increasing the number of different crops grown could disrupt the life cycle and reduce the frequency of favourable conditions for invertebrate pests and fungal diseases. Indeed, most insect pests and fungal pathogens are highly crop-specific^16,17,18. Although weed communities are also associated with crop types¹⁹, a given weed species can germinate and compete in many different crops, and produce seeds that replenish the seed bank every year, even with different succeeding crops, particularly if these crops are sown in the same season^28,29. This could explain why we did not detect any significant correlation between crop richness and herbicide TFI_diff. We would rather expect sowing seasonality to be the main factor driving the composition of the weed community of a given crop^29,30, and impacting weed demography and herbicide needed (but see the discussion below). The specific mechanisms by which evenness fosters pest regulation remain unclear^31,32. As the proportional abundance had a similar effect on pesticide use as richness, and this measure combines both richness and evenness²³, we expect that the measured effects of pesticide use are driven by the increasing richness, rather than evenness.

Incorporating temporary grassland into cropping systems could contribute to reduce the need for pesticides^33,34 in subsequent or adjacent crops, potentially due to its mitigation effect on weeds^30,35, fungal pathogens and insect pests³⁶. Previous studies have also shown that increasing the proportion of grassland leads to increases in the rate of predation of weed seeds and aphids³⁶, but high proportions of temporary grassland can also provide a favourable habitat for insect pests, serving as a reservoir for infestations³⁷. In our study, we confirmed that cropping systems that included temporary grassland had lower pesticide inputs as compared to those without, after controlling for the intrinsically low pesticide reliance of temporary grassland management (box plots of Fig. 2). Increasing the proportion of grassland led to a significant increase in the insecticide and other pesticide groups (e.g. molluscicides) (Fig. 2), suggesting that grasslands increasingly act as sources of invertebrate pests than as sources of natural enemies.

Intercropping has been demonstrated to suppress weeds^38,39, control diseases^40,41 and limit pest dispersal⁴², thereby contributing to decreased use of pesticides⁴³. Our results also suggest increases in the proportion of intercrops further decreasing herbicide use, a finding that is confirmed in studies demonstrating strong weed regulation on the whole cropping system³⁸.

The effect of increasing the number of botanical families only significantly decreased fungicide use, but had no effect on all other groups of pesticides (Fig. 2). This result was predominantly an effect of diversifying Poaceae-based cropping systems (N = 1283 out of 1285), with Fabaceae, Brassicaceae, and other families. Increasing proportions of Brassicaceae have been found to increase the pressure of herbivores⁴⁴, which could explain increased pesticide use (Figs. 2 and 3). There is no clear evidence that the Fabaceae (i.e. legume crops) have a strong suppressive effect on weeds or pathogens, but growing them is expected to deliver higher soil organic matter and better structure⁴⁵ and disrupt the life cycle of Poaceae-specific pathogens⁴⁶. These effects, in combination, could lead to reduced needs for all groups of pesticides at the cropping system level, except insecticides.

Seasonal turnover is observed in insects⁴⁷, pathogens⁴⁸ and soil microbial communities²⁸. Diversifying the sowing seasons in a cropping system reduced fungicide and insecticide use (Fig. 2). Surprisingly, but in accordance with results from Guinet et al.⁴⁹, we found no benefits of diversifying sowing seasons on the herbicide use. Indeed, diversifying sowing seasons could stimulate complex outcomes such as increasing soil disturbance that could allow a wider range of weed species to coexist⁵⁰. Crops that have a high reliance on pesticides, such as potato and beetroot that rely on herbicides, tend to be grown in crop sequences with diversified sowing seasons (Supplementary Fig. 13). This indirect link might mask the potential benefits of diversifying cropping seasons on the regulation of weed populations.

Increasing the proportion of crops sown in one season increased pest pressure, and consequently pesticide use, except for summer-sown crops (Fig. 2). Some spring crops like beetroot (accounting for 29.1% of the spring-sown crops, see Supplementary Fig. 12 and 84.5% of the Amaranthaceae, see Supplementary Figs. 9 and 10 for details), are susceptible to fungal pathogens, and are therefore highly reliant on fungicides. Increasing the proportion of spring crops could therefore favour the development of plant pathogens and trigger increases in fungicide use. Conversely, summer crops such as maize, sunflower and soybean (69.2%, 10.8% and 4.2% of summer crops, see Supplementary Fig. 12) are associated with a limited number of pests and relatively low pesticide reliance⁵. Increasing proportions of these summer crops did not affect pesticide use. Overall, our results on functional diversity suggest that: (i) crops associated with limited pest pressure tend to regulate pests for other crops at the cropping system level by breaking pest life cycles (e.g. temporary grasslands, summer crops); and, (ii) crops with high sensitivity to pests and which are grown in intensive patterns (higher proportions) tend to select pests that are able to exploit them (e.g. Brassicaceaes, spring crops), and this leads to a higher chemical use to achieve the desired control.

The CDM_total, based on 21 crop diversity indicators, explained 7.9% of the variance of total pesticide use, whereas richness alone had previously been found to explain only 1.3% of the variance of pesticide use⁵. In this study, we demonstrate that crop richness is only one component of crop diversity that drives pesticide application to avoid crop yield loss. Refining the definition of crop diversification by selecting other functional aspects of diversity produces a more predictive metric of pesticide need and use. However, the contribution of the CDM remained quite low in the variance partitioning, and much lower than the contribution of crop species identity, and the combination of other variates such as climate, soil, and technical crop management options. This diversification effect was not expected to be so small, as crop diversification is a priori regarded as one of the main approaches for reducing pesticide use at the farm level^10,11, although the few attempts to quantify the links between crop diversity and pesticide use have similarly produced significant, but limited effects⁴⁹.

Higher crop diversity seems to be associated with a reduced pesticide use, regardless of other variates such as soil, climate and management practices, as demonstrated by the very low covariance between crop diversity and the other factors (Fig. 5). Indeed, this identified effect of crop diversity could still coincide with other changes, such as management practices, as the process of diversification could be associated with different farming strategies (e.g. reducing tillage when introducing temporary grassland in cropping systems). However, the low covariances between crop diversity and other factors indicate a limited confounding effect on crop diversity. Disentangling the impact of management strategies from diversity changes requires further studies. Disease susceptibility of crops could also be determined by regional factors, such as crop frequency in the region³⁶, which might explain the relatively high covariance between crop species and the other factors for fungicides (Fig. 5c). Sowing seasons of crops might affect the pesticide reliance of crop species, and this could partly explain the covariances between crop species and crop diversity.

Indeed, besides crops and their diversity levels, a wide range of other factors could also affect the level of pesticide use, as reflected in the 31.9% of TFI variance explained by other factors. The impact of farmers’ decision-making stategies¹², or to a broader extent, the socio-economic context, such as the economic value of crops⁵¹, or policy restrictions on pesticides⁵² could all contribute to varied pesticide use levels. This implies that farmers may not only respond to pest levels in their pesticide decisions, but also take into account other factors. However, due to the specific nature of the DEPHY network database, where farmers joined voluntarily to reduce pesticide use, the influence of other factors should be subtle and mitigated. Isolating farmers’ response to pest levels from other factors is an important direction for future research.

According to the CDM, reducing pesticide use by increasing diversity is a question of balancing different forces at work in the cropping system. For example, incorporating soybeans every 2 years in a wheat monoculture changes the proportion of crops (i.e. change from 100% of wheat to 50% of wheat and 50% of soybeans), and therefore changes the various components of crop diversity. This change would affect the CDM_total sub-components, for richness (+0.13 unit of CDM_total), for the proportion of Fabaceae (+0.54), for the presence and proportion of summer crops (+0.24 and +0.2), for evenness (−0.33), for the proportion of Poaceae (+0.18), for dominance (+0.41), and for the number of botanical families (−0.33, a negative value since the number of botanical families was positively correlated with TFI_diff, and was therefore considered as a negative component of CDM_total). This leads to a value of −1.04 unit for TFI_diff as the effect of crop diversity, and −0.81 as the effect of crop species (i.e. 100% of wheat with average TFI of 3.28 replaced by 50% of soybean with averaged TFI of 1.66 and 50% of wheat, which makes an average TFI of 2.47; 3.28−2.47 = 0.81, data from Zhang et al.⁵). The effect of crop diversity is higher than the effect of crop species in this example, whereas, at the whole dataset scale, the variance of pesticide use related to diversity was lower than the variance related to crop species. This discrepancy could illustrate the fact that the process of crop diversification from a standard crop sequence (i.e. wheat monoculture in this example) most often explores a small range of crop species, with potentially a limited range of specific reliance on pesticides as compared to the large variation in pesticide reliance across crop species over the whole dataset. It is worth mentioning that diversifying with Brassicaceae crops requires a careful balance of forces between introducing them (−0.50 unit of CDM_total) and increasing their proportion (+2.25), because the benefit of their introduction disappears when their proportion is more than 22%.

The slope of the relationship between CDM and total pesticide use is very close to −1.0, indicating a proportional decrease in pesticide reliance with increasing composite crop diversity, independently of the pesticide requirements of the crop species grown. Even though the variability is large and may limit accurate prediction of pesticide use based solely on crop diversity, the observed trend provides a useful indication of overall pesticide use levels. The information needed to calculate the CDM only requires coarse crop proportions with little data collection effort, which could be potentially interesting for guiding pesticide and diversity-related policies. In addition, a better understanding of diversity effects on the regulation of pest populations contributes to the decision-making process of farmers in designing their cropping systems to fulfil their production targets, while minimising pesticide use. This metric, CDM, computed from our data of French farms, is likely to have a generic value for all cropping systems in temperate climates, but it would still require validation in other countries and agricultural contexts.

Conclusion

Disentangling the effects of biodiversity in the cropping system translates the complex functioning of biodiversity into knowledge that can guide agricultural production towards enhanced sustainability. Taking advantage of a unique large database at the cropping system level, we found that diversification towards pesticide reduction is a rather delicate process involving many competing effects that need to be harmonised (e.g. introducing species of Brassicaceae contributes to increased crop richness, but too much Brassicaceous crops increases pest pressure; introduction of temporary grasslands contributes to decrease the need for treatments in the other succeeding crops, but too much grasslands could favour the development of insect pests…). Our results could provide complementary knowledge in helping farmers design cropping systems that not only attain their operational objectives but also provide extra benefits in pest regulation and pesticide reduction. This goal could be achieved with a careful consideration of all the important facets of diversity simultaneously. Reducing reliance on pesticides will require an increase in the diversity of succeeding crops, but the efficiency of this strategy will be strengthened if robust crops intrinsically requiring limited pest control are preferentially selected for diversification. Our results provide the first comprehensive, quantitative attempt to fill the existing knowledge gap on the effect of crop diversity. The potential confounding of the diversity effects requires further exploration. A better understanding of the role of crop diversity is an urgently needed step forward in the transition to sustainable agriculture.

Methods

Data source

We used data from the French DEPHY network database, which contains detailed pesticide application information on each crop from the participating farms. The DEPHY network database was constructed under the French national Ecophyto Plan, which set the target of halving pesticide use by 2030⁵³. Data were collected from farmers when they joined the network, and yearly thereafter. There are around 3000 active farms in the network, with the first farmers joining the network in 2010, and the most recent data used in this study correspond to 2021. A cropping system is here defined as the succession of crops grown over a set of fields from a given farm that are managed with the same strategy over a period of 3 years (i.e. from the earlier 2010–2012 to the latest 2019–2021). The 3-year period is chosen to (i) smooth possible inter-annual variations in pesticide use due to climate and pest pressure variability, but not related to crop diversity, and (ii) to maximise the chances of counting all crops of the crop sequence over fields belonging to the same cropping system. A given farm might include one (most frequent case) or several cropping systems (e.g. when the farmers manage fields with different soil types). In this study, we considered 1285 cropping systems from 795 non-organic commercial arable farms. A total of 67 different field crops were identified across all cropping systems. Most crops were monospecific (n = 64), while crops grown in the pattern of mixture were categorised into three types, namely (i) temporary grassland (merged into one group of ‘temporary grassland’), (ii) intercropping (merged into one group of ‘intercrops’), and (iii) a group of uncommon species for field crops of ‘Other’.

Quantifying pesticide use

TFI was used to measure the reliance on pesticides, both at the crop and the cropping system level. TFI measures the frequency and intensity of pesticide application at the farmers’ decision level, i.e. the reliance on chemicals to control invertebrate pests, weeds, and fungal diseases to avoid yield loss. TFI does not measure the ecotoxicity of pest control strategies.

TFI of a given crop i compiles the number of treatments of pesticides, weighted by the ratio to the reference dose, and the treated crop surface to the total crop surface as follows⁴⁹:

$${{TFI}}_{i}={\sum }_{k=1}^{K}\frac{{D}_{k}}{{{DR}}_{k}}\times \frac{{S}_{k}}{{S}_{T}}$$

(1)

where D_k and DR_k are the applied dose and reference dose for the pesticide treatment k for a given crop i with total number of K treatments. S_k and S_T are the sprayed surface for treatment k and the total surface, respectively. The ratio of applied surface and total surface was reported on a given crop by farmers yearly. As each pesticide product can be applied to several crops and their associated pests, the reference dose used in TFI computation is defined by the French Ministry of Agriculture as the lowest registered dose of a given pesticide product for a given crop and a given pest combination. The reference dose can be found on the E-phy platform⁵⁴. The total pesticide TFI is the sum of herbicides TFI, fungicides TFI, insecticides TFI, and others TFI (including growth regulators, molluscicides, rodenticides, nematicides, etc., excluding seed coating). TFI was computed considering all applications of synthetic chemicals, excluding the ‘low-impact’ pesticides, i.e. products from the so-called ‘Biocontrol list’ maintained by the French Ministry of Agriculture.

The TFI of a given cropping system j is calculated as the weighted average of TFI values computed at the crop level, with the weight of crop i being the proportion of crop i in cropping system j, as follows:

$${{TFI}}_{j}={\sum }_{i=1}^{N}{{TFI}}_{i}\times {p}_{i}$$

(2)

where proportion of crop i (p_i) is calculated as the ratio of the number of fields occupied by crop i to the total number of fields of cropping system j over the 3 years, and N is the total number of crops in cropping system j.

Disentangling crop species effect in TFI

To assess the impact of crop diversity on pesticide use, regardless of crop species, we used the method proposed by Zhang et al.⁵ to disentangle the effect of crop species on pesticide use of cropping systems. Mean TFI for each crop species i ($\overline{{{{{\rm{TFI}}}}}_{{{{\rm{i}}}}}}$) was calculated by averaging TFI_i over the whole database to represent the crop-specific pesticide reliance. For each cropping system j, a predicted cropping system $\widehat{{{{{\rm{TFI}}}}}_{{{{\rm{j}}}}}}$ was calculated as the sum of weighted average of $\overline{{{{{\rm{TFI}}}}}_{{{{\rm{i}}}}}}$, weights being the proportions p_i of all crops composing the cropping system j as follows:

$$\widehat{{{TFI}}_{j}}={\sum }_{i=1}^{N}\overline{{{TFI}}_{i}}\times {p}_{i}$$

(3)

The difference (TFI_diff) between the observed cropping system TFI_j and the predicted $\widehat{{{{{\rm{TFI}}}}}_{{{{\rm{j}}}}}}$ reflected the influences on pesticide use of all other factors except crop species. TFI_diff was computed as follows:

$${{TFI}}_{{diff}}={{TFI}}_{j}-\widehat{{{TFI}}_{j}}$$

(4)

TFI_diff was the metric used to test the impacts of crop diversity on pesticide use, with the removed effect of crop species composing the cropping system. The effect of crop diversity was tested for total pesticide (TFI_diff total), herbicide (TFI_diff herbicide), fungicide (TFI_diff fungicide), insecticide (TFI_diff insecticide), and other pesticides (TFI_diff other).

Selecting crop diversity indicators

We selected indicators measuring both species diversity and functional diversity at cropping system levels to consider all facets of crop diversity.

We considered the most important features related to species diversity, namely (i) richness^55,56, (ii) evenness^21,56, (iii) proportional abundance, combining simultaneously richness and evenness^22,23 and (iv) dominance^24,57. We found 9, 4 and 2 applicable indicators in the literature describing evenness, proportional abundance and dominance, respectively (Supplementary Tables 1 and 2). Indicators that could not be computed from our data were filtered out and listed in Supplementary Table 2. Except for richness, the potential indicators we found for each species diversity feature were highly correlated (Pearson’s correlation coefficient ≥ 0.7 or ≤ −0.7, see Supplementary Table 2 and Supplementary Figs. 2–4 for details). To avoid redundancy, only one indicator for each species diversity feature was included in the analysis. Therefore, for each species diversity feature, we chose the indicator based on its regression with pesticide use (TFI_diff) by the overall lowest residual standard error and the smallest P value (Supplementary Figs. 5–7). This led to the final selection of the evenness indicator from Chao and Ricotta²¹, the Simpson inverse diversity index²² and the Berger–Parker index²⁴ representing dominance. Richness was counted as the number of different crops (N) in the cropping system. Evenness, proportional abundance and dominance were calculated according to their equations based on the proportion of different crops (p_i) (see Supplementary Table 1 for details).

For functional diversity and functional features of agronomical relevance, incorporating temporary grassland^33,34, intercrops^43,58, cover crops^58,59, diversifying botanical families^46,49 and diversifying by including an additional crop sowing seasons (e.g. introducing summer crops into systems dominated by winter-sown crops)^28,29 could affect pest pressure. We included each functional diversity feature in the analysis both as presence/absence (e.g. presence/absence of temporary grassland) and as proportion (e.g. proportion of temporary grassland), to consider the potential non-linear effects of functional diversity features on pesticide use. For the binary comparisons of presence/absence of each diversity feature, box plots were used to represent the distribution of TFI_diff for cropping systems with vs. without the corresponding feature (e.g. with vs without temporary grasslands; see Supplementary Table 3 for details). Box plots represent 25th, 50th and 75th percentiles. Outliers (points) correspond to values that lie more than 1.5 times the interquartile range beyond the upper or lower quartile. Besides, the number of botanical families and the number of sowing seasons were also used as indicators of diversity in this study. The proportion of each functional diversity feature in a cropping system was calculated as the ratio of the number of fields occupied by the functional feature to the total number of fields of the cropping system over the 3 years (e.g. if 3 out of 10 fields were grown with temporary grassland, the proportion of temporary grassland was 30% in this cropping system). The proportion of the first three major botanical families (i.e. Poaceae, Fabaceae, Brassicaceae) and the proportion of the other botanical families (i.e. the proportion of other families excluding the first three) were included in the analysis (see Supplementary Figs. 8–10 for details). Each crop was classified as a spring-, summer-, fall-, or winter-sown crop based on its sowing date to calculate the proportion of sowing seasons (Supplementary Figs. 11 and 12). The proportion of cover crops was computed as the proportion of fields where a cover crop was sown and grown before the sowing of a cash crop. In sum, 28 indicators were included (see Supplementary Table 1 for details), and a correlation matrix was produced to check the correlation of these indicators (Supplementary Fig. 1).

Statistical analyses

All statistical analyses were carried out with R software version 4.3.1⁶⁰.

Assessing individual indicators

Linear regressions of each diversity indicator were conducted individually regarding TFI_diff for each pesticide group (TFI_diff herbicides, TFI_diff fungicides, TFI_diff insecticides, and TFI_diff other). Linear regressions of each functional diversity indicator were based on cropping systems with the presence of the corresponding functional feature (e.g. for the linear regression involving the proportion of temporary grassland, we only included the cropping systems with temporary grassland). The t-test was conducted on two groups of cropping systems (i.e. the cropping systems with the targeted functional features vs. the cropping systems without the targeted functional features) for analysing each functional diversity indicator.

Selecting indicators to compute the composite diversity metric

Due to the potential complex interaction between diversity indicators, we chose Lasso regression²⁵ to select the highly explanatory variables for each pesticide group. The combination of the selected indicators and the corresponding coefficients computed by Lasso formed a CDM. A specific CDM was computed for each pesticide group. A higher CDM value reflects an overall higher diversity level when considering all selected diversity indicators. Lasso can deal with high-dimensional data and shrink the non-impactful indicators towards zero. The regularisation parameter λ was selected to minimise leave-one-out cross-validation error (LOO-XVE). The number of crops, the number of botanical families, the Simpson inverse diversity index, and the number of sowing seasons were rescaled to the same scale as crop proportions (i.e. 0–1). Confidence intervals of the Lasso-selected indicators were computed by using the bootstrap method³³.

Linear regressions of CDMs were conducted with the fitted values of the Lasso-selected indicators applied in each cropping system (namely, TFI_fitted_CDM, i.e. the sum of each coefficient of the Lasso-selected indicators multiplied by the value of diversity indicators of each cropping system) for TFI_diff of each pesticide group.

The variance partitions of diversity

The variance of observed cropping system TFI over the whole dataset was analysed by distinguishing the variances of the different components⁵, namely (i) the variance of the predicted cropping system TFI related to crop species (var_species), (ii) the variance explained by crop diversity, represented by the fitted value of Lasso model i.e. TFI_fitted_CDM, (var_diversity), and (iii) the variance explained by the other factors (var_other). The TFI values that were affected by other factors were represented by the difference between the TFI_diff (i.e. TFI after controlling the effect of crop species) and TFI_fitted_CDM (i.e. TFI fluctuations that were brought by crop diversity). These variances were considered as ratios of the total variance of the cropping system TFI (i.e. Var_TFI_j). Covariances between the three components (namely, cov_species_diversity, covariance between crop species, and crop diversity; cov_species_other, covariance between crop species and other factors; and cov_diversity_other, covariance between crop diversity and other factors) were computed to finalise the analysis of variance partitioning.

Data availability

The dataset used in this study was deposited in the Data INRAE repository and is available at: https://doi.org/10.57745/SBTUXP. Any further questions about this paper can be addressed to the corresponding author.

Code availability

The R script used to analyse the data and generate figures was deposited in the Data INRAE repository, and is available at: https://doi.org/10.57745/SBTUXP.

References

Tang, F. H., Lenzen, M., McBratney, A. & Maggi, F. Risk of pesticide pollution at the global scale. Nat. Geosci. 14, 206–210 (2021).
Article CAS Google Scholar
Ratnadass, A., Fernandes, P., Avelino, J. & Habib, R. Plant species diversity for sustainable management of crop pests and diseases in agroecosystems: a review. Agron. Sustain. Dev. 32, 273–303 (2012).
Article Google Scholar
Letourneau, D. K. et al. Does plant diversity benefit agroecosystems? A synthetic review. Ecol. Appl. 21, 9–21 (2011).
Article Google Scholar
Kremen, C. & Miles, A. Ecosystem services in biologically diversified versus conventional farming systems: benefits, externalities, and trade-offs. Ecol. Soc. 17, 342 (2012).
Article Google Scholar
Zhang, Y. et al. Pesticide use is affected more by crop species than by crop diversity at the cropping system level. Eur. J. Agron. 159, 127263 (2024).
Article Google Scholar
Nicolopoulou-Stamati, P., Maipas, S., Kotampasi, C., Stamatis, P., & Hens, L. Chemical pesticides and human health: the urgent need for a new concept in agriculture. Public Health Front. 4, 148 (2016).
Hill, S. B. & MacRae, R. J. Conceptual framework for the transition from conventional to sustainable agriculture. J. Sustain. Agric. 7, 81–87 (1996).
Article Google Scholar
Altieri, M. A. The ecological role of biodiversity in agroecosystems. Agric. Ecosyst. Environ. 74, 19–31 (1999).
Article Google Scholar
Tamburini, G. et al. Agricultural diversification promotes multiple ecosystem services without compromising yield. Sci. Adv. 6, eaba1715 (2020).
Article Google Scholar
Vialatte, A. et al. Protéger les cultures en augmentant la diversité végétale des espaces agricoles. Rapport scientifique de l’Expertise scientifique collective https://hal.inrae.fr/hal-04127709 (Institut National de Recherche Pour l’agriculture, l’alimentation et l’environnement, 2023).
Bommarco, R. Ecological redesign of crop ecosystems for reliable crop protection. A review. Agron. Sustain. Dev. 44, https://doi.org/10.1007/s13593-024-00987-z (2024).
Czapar, G. F., Curry, M. P. & Wax, L. M. Grower acceptance of economic thresholds for weed management in Illinois. Weed Technol 11, 828–831 (1997).
Article Google Scholar
Alletto, L., Vandewalle, A. & Debaeke, P. Crop diversification improves cropping system sustainability: an 8-year on-farm experiment in South-Western France. Agric. Syst. 200, 103433 (2022).
Article Google Scholar
Larsen, A. E. & Noack, F. Impact of local and landscape complexity on the stability of field-level pest control. Nat. Sustain. 4, 120–128 (2021).
Article Google Scholar
Nicholson, C. C. & Williams, N. M. Cropland heterogeneity drives frequency and intensity of pesticide use. Environ. Res. Lett. 16, 074008 (2021).
Article CAS Google Scholar
Fitt, B. D. L., Huang, Y.-J., van den Bosch, F. & West, J. S. Coexistence of related pathogen species on arable crops in space and time. Annu. Rev. Phytopathol. 44, 163–182 (2006).
Article CAS Google Scholar
Smith, V. et al. Weed and invertebrate community compositions in arable farmland. Arthropod Plant Interact. 2, 21–30 (2008).
Article Google Scholar
Viketoft, M. et al. Long-term effects of plant diversity and composition on soil nematode communities in model grasslands. Ecology 90, 90–99 (2009).
Article Google Scholar
Perronne, R., Le Corre, V., Bretagnolle, V. & Gaba, S. Stochastic processes and crop types shape weed community assembly in arable fields. J. Veg. Sci. 26, 348–359 (2014).
Article Google Scholar
Downing, A. L. & Leibold, M. A. Ecosystem consequences of species richness and composition in pond food webs. Nature 416, 837–841 (2002).
Article CAS Google Scholar
Chao, A. & Ricotta, C. Quantifying evenness and linking it to diversity, beta diversity, and similarity. Ecology 100, e02852 (2019).
Article Google Scholar
Simpson, E. H. Measurement of diversity. Nature 163, 688–688 (1949).
Article Google Scholar
Stirling, G. & Wilsey, B. Empirical relationships between species richness, evenness, and proportional diversity. Am. Nat. 158, 286–299 (2001).
Article CAS Google Scholar
Berger, W. H. & Parker, F. L. Diversity of planktonic foraminifera in Deep-Sea sediments. Science 168, 1345–1347 (1970).
Article CAS Google Scholar
Tibshirani, R. Regression shrinkage and selection via the lasso. J. R. Stat. Soc. Ser. B Methodol. 58, 267–288 (1996).
Article Google Scholar
Otway, S. J., Hector, A. & Lawton, J. H. Resource dilution effects on specialist insect herbivores in a grassland biodiversity experiment. J. Anim. Ecol. 74, 234–240 (2005).
Article Google Scholar
Mitchell, C. E., Tilman, D. & Groth, J. V. Effects of grassland plant species diversity, abundance, and composition on foliar fungal disease. Ecology 83, 1713–1726 (2002).
Article Google Scholar
Hilton, S., Bennett, A. J., Chandler, D., Mills, P. & Bending, G. D. Preceding crop and seasonal effects influence fungal, bacterial and nematode diversity in wheat and oilseed rape rhizosphere and soil. Appl. Soil Ecol. 126, 34–46 (2018).
Article Google Scholar
Weisberger, D., Nichols, V. & Liebman, M. Does diversifying crop rotations suppress weeds? A meta-analysis. PLoS ONE 14, e0219847 (2019).
Article CAS Google Scholar
Meiss, H., Médienne, S., Waldhardt, R., Caneill, J. & Munier-Jolain, N. M. Contrasting weed species composition in perennial lucernes and six annual crops: implications for Integrated Weed Management. Agron. Sustain. Dev. 30, 657–666 (2010).
Article Google Scholar
Crowder, D. W., Northfield, T. D., Strand, M. R. & Snyder, W. E. Organic agriculture promotes evenness and natural pest control. Nature 466, 109–112 (2010).
Article CAS Google Scholar
Wittebolle, L. et al. Initial community evenness favours functionality under selective stress. Nature 458, 623–626 (2009).
Article CAS Google Scholar
Lechenet, M., Dessaint, F., Py, G., Makowski, D. & Munier-Jolain, N. Reducing pesticide use while preserving crop productivity and profitability on arable farms. Nat. Plants 3, 17008 (2017).
Article Google Scholar
Nandillon, R., Guinet, M. & Munier-Jolain, N. Crop management strategy redesign enables a reduction in reliance on pesticides: a diachronic approach based on a diversity of French commercial farms. Agric. Ecosyst. Environ. 366, 108949 (2024).
Article CAS Google Scholar
Dominschek, R. et al. Crop rotations with temporary grassland shifts weed patterns and allows herbicide-free management without crop yield loss. J. Clean. Prod. 306, 127140 (2021).
Article Google Scholar
Perrot, T., Rusch, A., Coux, C., Gaba, S., & Bretagnolle, V. Proportion of grassland at landscape scale drives natural pest control services in agricultural landscapes. Front. Ecol. Evol. 9, 607023 (2021).
Poggi, S. et al. Dynamic role of grasslands as sources of soil-dwelling insect pests: New insights from in silico experiments for pest management strategies. Ecol. Model. 440, 109378 (2021).
Article Google Scholar
Gu, C., Bastiaans, L., Anten, N. P. R., Makowski, D. & van der Werf, W. Annual intercropping suppresses weeds: a meta-analysis. Agric. Ecosyst. Environ. 322, 107658 (2021).
Article CAS Google Scholar
Bedoussac, L. et al. Ecological principles underlying the increase of productivity achieved by cereal-grain legume intercrops in organic farming. A review. Agron. Sustain. Dev. 35, 911–935 (2015).
Article Google Scholar
Boudreau, M. A. Diseases in intercropping systems. Annu. Rev. Phytopathol. 51, 499–519 (2013).
Article CAS Google Scholar
Luo, C., Lv, J., Guo, Z. & Dong, Y. Intercropping of faba bean with wheat under different nitrogen levels reduces faba bean rust and consequent yield loss. Plant Dis. 106, 2370–2379 (2022).
Article CAS Google Scholar
Puliga, G. A., Arlotti, D. & Dauber, J. The effects of wheat-pea mixed intercropping on biocontrol potential of generalist predators in a long-term experimental trial. Ann. Appl. Biol. 182, 37–47 (2022).
Article Google Scholar
Yan, E., Munier-Jolain, N., Martin, P. & Carozzi, M. Intercropping on French farms: Reducing pesticide and N fertilizer use while maintaining gross margins. Eur. J. Agron. 152, 127036 (2024).
Article CAS Google Scholar
Root, R. B. Organization of a plant-arthropod association in simple and diverse habitats: the fauna of collards (brassica oleracea). Ecol. Monogr. 43, 95–124 (1973).
Article Google Scholar
Preissel, S., Reckling, M., Schläfke, N. & Zander, P. Magnitude and farm-economic value of grain legume pre-crop benefits in Europe: a review. Field Crops Res. 175, 64–79 (2015).
Article Google Scholar
Kirkegaard, J., Christen, O., Krupinsky, J. & Layzell, D. Break crop benefits in temperate wheat production. Field Crops Res. 107, 185–195 (2008).
Article Google Scholar
Wolda, H. Insect seasonality: Why?. Annu. Rev. Ecol. Syst. 19, 1–18 (1988).
Article Google Scholar
Hu, W., Strom, N. B., Haarith, D., Chen, S. & Bushley, K. E. Seasonal variation and crop sequences shape the structure of bacterial communities in cysts of soybean cyst nematode. Front. Microbiol. 10, 2671 (2019).
Article Google Scholar
Guinet, M. et al. Fostering temporal crop diversification to reduce pesticide use. Nat. Commun. 14, 7416 (2023).
Article Google Scholar
Hernández Plaza, E., Navarrete, L. & González-Andújar, J. L. Intensity of soil disturbance shapes response trait diversity of weed communities: the long-term effects of different tillage systems. Agric. Ecosyst. Environ. 207, 101–108 (2015).
Article Google Scholar
Rosenheim, J. A., Cass, B. N., Kahl, H. & Steinmann, K. P. Variation in pesticide use across crops in California agriculture: economic and ecological drivers. Sci. Total Environ. 733, 138683 (2020).
Article CAS Google Scholar
Möhring, N. et al. Pathways for advancing pesticide policies. Nat. Food 1, 535–540 (2020).
Article Google Scholar
Lamichhane, J. R., Messéan, A. & Ricci, P. Research and innovation priorities as defined by the Ecophyto plan to address current crop protection transformation challenges in France. Adv. Agron. 154, 81–152 (2019).
Article Google Scholar
ANSES. E-Phy: Le catalogue des produits phytopharmaceutiques et de leurs usages, des matières fertilisantes et des supports de culture autorisés en France. https://ephy.anses.fr/ (2024).
Tilman, D. et al. The influence of functional diversity and composition on ecosystem processes. Science 277, 1300–1302 (1997).
Article CAS Google Scholar
Purvis, A. & Hector, A. Getting the measure of biodiversity. Nature 405, 212–219 (2000).
Article CAS Google Scholar
Hillebrand, H., Bennett, D. M. & Cadotte, M. W. Consequences of dominance: a review of evenness effects on local and regional ecosystem processes. Ecology 89, 1510–1520 (2008).
Article Google Scholar
Beillouin, D., Ben-Ari, T., Malézieux, E., Seufert, V. & Makowski, D. Positive but variable effects of crop diversification on biodiversity and ecosystem services. Glob. Change Biol. 27, 4697–4710 (2021).
Article CAS Google Scholar
Debaeke, P. et al. Non-chemical weed management: Which crop functions and traits to improve through breeding?. Crop Prot. 179, 106631 (2024).
Article Google Scholar
R Core Team. R: a language and environment for statistical computing. (R Foundation for Statistical Computing, 2023).

Download references

Acknowledgements

The authors thank the support from the China Scholarship Council (CSC), the AgroEcoSystèmes, Department of French National Research Institute for Agriculture, Food & Environment (INRAE) and Embassy of France in China for the funding of PhD fellowship and the platform for scientific collaboration of the International Associated Laboratory (LIA) dedicated to Agroecology and Agricultural Green Development (A-AGD). The work was supported by the MoBiDiv (ANR-20-PCPA-0006) project funded by the ‘Growing and Protecting Crops Differently’ French Priority Research Program (PPR-CPA) and DiroDur (OFB-24-1716) project funded by the French Biodiversity Office. We also thank farmers and farm advisors from the DEPHY network who shared and collected the data, as well as the AGROSYST team who extracted the data. The authors also thank Michael McCallum for his English revision.

Author information

These authors contributed equally: Nicolas Munier-Jolain, Laurent Bedoussac.

Authors and Affiliations

UMR Agroécologie, Institut Agro Dijon, INRAE, Université Bourgogne Europe, Dijon, France
Yaoyun Zhang, David A. Bohan & Nicolas Munier-Jolain
College of Resources and Environmental Sciences, State Key Laboratory of Nutrient Use and Management, China Agricultural University, Beijing, China
Chaochun Zhang & Wen-Feng Cong
UMR AGIR, ENSFEA, INRAE, INPT, El PURPAN, Université de Toulouse, Castanet-Tolosan, France
Laurent Bedoussac

Authors

Yaoyun Zhang
View author publications
Search author on:PubMed Google Scholar
David A. Bohan
View author publications
Search author on:PubMed Google Scholar
Chaochun Zhang
View author publications
Search author on:PubMed Google Scholar
Wen-Feng Cong
View author publications
Search author on:PubMed Google Scholar
Nicolas Munier-Jolain
View author publications
Search author on:PubMed Google Scholar
Laurent Bedoussac
View author publications
Search author on:PubMed Google Scholar

Contributions

Y.Z., N.M. and L.B. designed the study; Y.Z. conducted the research, and wrote the first draft; N.M. and L.B. revised the paper; D.B., C.Z. and W.C contributed to the discussion of the paper.

Corresponding author

Correspondence to Nicolas Munier-Jolain.

Ethics declarations

Competing interests

The authors declare no competing interests.

Peer review

Peer review information

Communications Earth & Environment thanks the anonymous reviewers for their contribution to the peer review of this work. Primary Handling Editors: Wenfeng Liu and Mengjie Wang. A peer review file is available.

Additional information

Publisher’s note Springer Nature remains neutral with regard to jurisdictional claims in published maps and institutional affiliations.

Supplementary information

Transparent Peer Review file

Supplementary material

Rights and permissions

Open Access This article is licensed under a Creative Commons Attribution-NonCommercial-NoDerivatives 4.0 International License, which permits any non-commercial use, sharing, distribution and reproduction in any medium or format, as long as you give appropriate credit to the original author(s) and the source, provide a link to the Creative Commons licence, and indicate if you modified the licensed material. You do not have permission under this licence to share adapted material derived from this article or parts of it. The images or other third party material in this article are included in the article’s Creative Commons licence, unless indicated otherwise in a credit line to the material. If material is not included in the article’s Creative Commons licence and your intended use is not permitted by statutory regulation or exceeds the permitted use, you will need to obtain permission directly from the copyright holder. To view a copy of this licence, visit http://creativecommons.org/licenses/by-nc-nd/4.0/.

Reprints and permissions

About this article

Cite this article

Zhang, Y., Bohan, D.A., Zhang, C. et al. Crop diversity reduces pesticide use more efficiently with refined diversification strategies. Commun Earth Environ 6, 460 (2025). https://doi.org/10.1038/s43247-025-02418-7

Download citation

Received: 21 November 2024
Accepted: 26 May 2025
Published: 13 June 2025
Version of record: 13 June 2025
DOI: https://doi.org/10.1038/s43247-025-02418-7