Rice Bund Management by Filipino Farmers and Willingness to Adopt Ecological Engineering for Pest Suppression

Abstract

Ecological engineering is defined as the design of ecosystems for the benefit of human society and the environment. In Asia, the ecological engineering of rice fields by establishing vegetation on bunds/levees for natural enemies has recently gained traction; however, its success depends on farmers’ willingness to implement changes. We surveyed 291 rice farmers in four regions of the Philippines to assess their bund management practices and willingness to establish bund vegetation that restores rice regulatory ecosystem services for pest management. During pre- and post-open field day (OFD) interviews, we assessed farmers’ perceptions of ecological engineering practices and sought their advice concerning bund vegetation. Over 60% of the farmers grew crops or allowed weeds on their bunds. Vegetables were grown as a source of extra food or income, and flowers were grown for pest management. Among the remaining farmers, their willingness to try ecological engineering increased from 36 to 74% after the OFDs. Furthermore, after the OFDs, willing farmers increasingly (from 2.6 to 15%) cited pest management as a reason to grow vegetables on bunds, and farmers almost exclusively focused on growing vegetables rather than flowers to adapt the system. While 46.5% of farmers who grew vegetables on their bunds applied insecticides, only ca 20% indicated that they would do so after the OFDs, if needed. Farmers had differing opinions on how vegetables would be incorporated into their rice farms. This range of options could be encouraged wherever farmers recognize the potential harm from pesticides for biodiversity and the restoration value of a diversified farm habitat.

1. Introduction

Rice is the main staple of the Asian diet and provides income and livelihoods for millions of people, including resource-poor farmers and farm laborers [1,2]. Since the beginning of the millennium, Asian farmers have dramatically increased their pesticide inputs, which has negative impacts on human health and on the biodiversity and related services of rice ecosystems [3,4,5,6]. Furthermore, with a ready availability of affordable pesticides, many farmers now apply pesticides prophylactically (i.e., based on calendar dates or crop stages), without monitoring their fields for potential risks [7,8,9,10,11]. For many farmers, this means that increasing pesticide use has not improved rice productivity or farm profits [6,12]. Insecticide use in particular has been associated with a loss of rice regulatory ecosystem services, leading to largescale outbreaks of herbivores such as planthoppers and leaf-folders and consequent rice yield losses [13,14]. This paradoxical behavior, where farmers apply pesticides without need and despite negative consequences, is attributed to a strong aversion to risks [4,15] and to efficient communication and marketing by the proponents of chemical-intensive agriculture [8,16]. Indeed, several non-chemical alternatives to insecticides in rice are already available to farmers, but, even where these have been widely adopted, farmers are still slow to reduce chemical inputs [7,17,18]. These trends highlight the need to better communicate with farmers on the drawbacks of using chemical pesticides and on the benefits of alternatives. However, it is also clear that, for improved upscaling and to overcome any reservations which farmers might have, alternatives should be designed in consultation with the farmers.

Farmer training and participatory research are essential components of sustainable crop production and integrated pest management (IPM) [19,20]. Much of the training that farmers receive on IPM is provided by government extension officers, advocates of the chemical industry, or pesticide suppliers [8,21]. Largescale farmer-participatory training programs offered through national and international campaigns have also been prominent at different times and in different regions. In particular, the Farmer Field School (FFS) approach to IPM training, as developed by the Food and Agriculture Organization of the United Nations (FAO), had prominence among Asian rice farmers in the 1990s and has since expanded to include other crops, regions, and topics [19]. International research organizations and universities have also played a significant role in promoting sustainable practices such as agroecology, ecologically based pest management, and ecological engineering [22,23,24]. However, such programs are expensive and usually restricted to a project lifespan of 3–5 years, after which pesticide use sometime returns to or surpasses pre-program levels [25,26]. To overcome problems related to sustainability impacts, FFSs and other programs emphasize the need for farmers to “experience” alternatives and, based on their accumulated and shared knowledge, participate in the adaptation of alternatives to better suit their needs and preferences [19,27,28]. Open field days (OFDs) are an essential component of co-learning (i.e., where researchers, extensionists, and trainers learn together with the farmers); they also provide a convenient platform for farmers to participate in the design of farming systems [24,26,29,30] as alternatives to chemically intensive pest management.

In response to widescale pesticide-induced outbreaks of rice pests across Asia in the early 2000s, several institutes began researching ecological engineering practices for rice pest management [7,17,18,31,32,33,34,35,36,37], some of which included strong advocacy components [38,39]. Ecological engineering is defined as the design of sustainable ecosystems, using engineering principals, for the benefit of human societies and the environment [40]. In terms of pest management, ecologically engineered production landscapes generally include functional habitats, such as flower strips or vegetable plots, which enhance the regulatory activities of natural enemies. These habitats can include refuge and feeding plants for natural enemies or trap and repellent plants for pests [41,42]. In 2013, the Philippines Government (Department of Agriculture-Bureau of Agricultural Development (DA-BAR)) initiated a program to incorporate ecological engineering as part of its Palayamanan model for rice-based farming, which promotes farm diversification by including fruit and vegetable production, livestock rearing, and aquaculture [43]. The program included the establishment of ecological engineering farm-sites (described below) as research and demonstration plots in four regions (Laguna, Rizal, Iloilo, and Bukidnon) [18,44]. The project also trained DA extension and research staff to identify beneficial arthropods in rice fields and aimed to assess farmer willingness to adopt the proposed ecological engineering practices.

In the present study, we describe farmers’ bund management practices and report on their perceptions and suggestions for adjusting bund management using ecological engineering practices to manage pests in rice and vegetables. We applied farmer surveys before and after OFDs at ecological engineering demonstration plots to gain unbiased information on current practices (before the OFDs) and farmers’ informed advice and perceptions of the proposed technology (after the OFDs). The OFDs offered an opportunity to gain ideas from the farmers that would be combined with traditional field studies to assess the impacts of ecological engineering on pest damage and rice yields, natural enemies and other wildlife, as well as farm economies. The results of the field trials will be published in future papers. During the structured interviews, we initially assessed farmers’ knowledge, practices, and perceptions related to ecological engineering and the planting of flowers or vegetables on rice bunds under their own initiative or as part of the Palayamanan program. The farmers were also assessed for changes in their perceptions of pest management as a result of the OFDs and their willingness to consider changes to their current pest management practices. The information and advice received from the farmers and documented here can be used to co-design rice landscapes and develop amenable strategies for the promotion of agroecology.

2. Materials and Methods

2.1. Demonstration Plots

Research was conducted at four sites on three islands of the Philippines, each of which had areas of rice fields with established patches of vegetation as part of an ecological engineering design for rice pest management. Three (Laguna, Rizal, and Iloilo) of the four sites were at research stations, with the remaining site (Bukidnon) on a commercial rice farm. High-diversity vegetation patches (HDVPs) that consisted of several small areas (1–2 m²) of vegetables at intermittent distances (ca 10 m) along rice bunds were established at each of the sites. The HDVP design departed from linear vegetation strips, which are continuous strips of flowering plants grown directly on the bunds, that many farmers would have already been familiar with through the Palayamanan and other programs [43,45,46]. The new patch configuration was designed to reduce the dispersion of rodents through the rice landscape and provide supplementary food or incomes to the farming households, thereby representing a form of positive feedback to enhance the sustainability of the system [17,35,47].

The HDVPs were standardized at all the sites: each patch had ladyfinger (Abelmoschus esculentus (L.) Moench), mungbean (Vigna radiata (L.) R. Wilczek), bitter gourd (Momordica charantia L.), and stringbean (Vigna unguiculata subsp. sesquipedalis (L) Verdc.) planted on raised earthen platforms. The bitter gourd and stringbean were supported on bamboo stakes. The size of patches ranged from 0.93 to 2.40 m², with the smallest patches at the Laguna site and the largest at the Bukidnon site. The areas with HDVPs ranged from 0.42 ha (Laguna) to 1.31 ha (Bukidnon), with the proportion of each area under patches ranging from 1.0 (Iloilo) to 1.4% (Bukidnon). All the rice fields were planted with the rice variety Rc298 and had adjacent areas of about the same size without HDVPs, which were also planted with Rc298 (i.e., treatment controls). No pesticides were used at any of the sites during the rice cropping period. The demonstration plots had been established during the previous rice cropping season (dry season) at each site and were now in their second season (wet season). Therefore, data on dry-season rice and fruit yields at the sites were available for the OFDs. At the time of the OFDs, the rice crops were at the tillering stages, and the HDVPs were producing fruits. Further information about the sites has been presented by Horgan et al. (2016) [44].

2.2. Open Field Days

The OFDs were designed to showcase the demonstration plots while also informing participants about the functions of ecological engineering and the HDVPs in the rice landscape and gaining farmer feedback on the proposed technology. The participants were informed about the reasons why the HDVPs were established and the benefits of ecological engineering in terms of augmenting natural-enemy diversity and abundance and providing extra food and income for farmers. The OFDs were standardized by including the same information in poster, presentation, and video form, available to all the participants, and allotting the same time for field visits and open discussions. The OFDs therefore included welcome remarks by local organizers and researchers from participating institutes that outlined the course of the day’s events and organized farmers for interviews. After the pre-event interviews, the participants were transported to the demonstration plots, where the plot designs and establishment process were explained in local languages by researchers. The participants were able to walk among the plots and could ask questions or make comments in plenary or in small groups with the researchers. The participants also received information leaflets on ecological engineering (Table S1) and the control of apple snails (Pomacea canaliculata (Lamarck)) and received letter-sized printouts of all the posters. Visits to the demonstration plots lasted ca 90 min. The participants then returned to a shaded and centralized area (i.e., auditoriums, barns, or tarps) to view a series of information posters and an insect zoo (ca 30 min), after which there was a further information session with an expert’s presentation and video, followed by a discussion forum with researchers and participants. The posters were all in English; however, research and extension staff at the sites were present to help explain the information in local languages if needed. The presentations and videos were in local languages, and any open discussions were in both English and local languages. A list of the information imparted to the participants is presented in Table S1.

2.3. Farmer Surveys

Pre-event interviews were conducted after the welcome remarks (see above) at centralized locations (e.g., schools, village halls, or DA facilities) with farmers from adjacent towns and villages invited by DA staff through village leaders. The farmers were contacted by the village leaders through a series of farmer groups, often linked to specific communities, and were actively engaged in rice farming. To increase the representativeness of our samples and reduce biases, farmers were invited from around 14 to 21 villages that were distant from the demonstration plots (i.e., the farmers would not have previously seen the plots). In Laguna Province, farmers from 22 villages were initially interviewed at Victoria, Pila, and Nagcarlan and later transported to the demonstration plots at the International Rice Research Institute (IRRI). In Rizal Province, farmers from 14 villages were initially interviewed at Pililia and Tanay and transported to a demonstration plot at the Rizal Agricultural Station (RAS). In Iloilo Province, farmers from 16 villages were initially interviewed at Zarraga, Dumangas, and Dingle and later transported to a demonstration plot at the Western Visayas Agricultural Research Center (WESTVIARC). In Bukidnon Province, farmers from 21 villages were initially interviewed at a school hall in Malaybalay and then transported to the Alugboc Rice Farm, which had two separate demonstration plots. The organizers of the field events and surveys in each province were instructed not to disclose the topic of the interviews to the participants or village leaders before the pre-event interviews took place. After viewing the demonstration plots, the farmers were interviewed again (post event) near the sites. Further details of the farmers and farm villages, based on the pre-event interviews only, are presented in a related paper by Horgan et al. (2023) [18].

The farmers were interviewed using structured and standardized questionnaires. In total, 270 farmers completed the pre- and post-event interviews, and a further 21 completed only the post-event interviews (i.e., 291 in total; 55 in Laguna, 56 in Rizal, 95 in Iloilo, and 85 in Bukidnon). To facilitate the interviews, the questionnaires were translated from English into three local languages: Tagalog for Rizal and Laguna, Ilonggo for Iloilo, and Cebuano for Bukidnon. These translations were carried out by native speakers with a knowledge of rice farming and pest management. The interviewers consisted of DA staff at each centralized location and staff from IRRI based in Los Baños, Laguna. The interviewers were trained in interview methods, and each interviewer was familiar with rice production systems and crop management. A total of 61 interviewers conducted the one-to-one interviews. This large number of interviewers was required because of the different languages used at the sites and to improve cultural sensitivity during the capturing of data; many were also involved in explaining concepts from the posters and the demonstration plots to the participants.

The questionnaires were developed according to the knowledge, attitudes, and practices (KAP) survey technique [48]. This robust technique is widely used during project planning for agricultural development and, particularly, for pest management interventions [49,50,51,52,53]. It is also relatively resilient to varying levels of interviewer experience. Furthermore, the technique allows the assessment of group perceptions and practices at a level (i.e., village or community) that is suitable for landscape approaches to pest management [49,50,53]. The pre-event questionnaire was developed based on three focus group discussions (FGDs) conducted between February and June of 2014, one with farmers in Laguna and two with farmers, seed suppliers, and DA staff in Iloilo. Pre-testing of the survey was conducted in Laguna prior to the Laguna OFD. The surveys were further refined after the Laguna OFD (which was the first of the four OFDs). Responses during the FGDs were used to code the questionnaires and thereby facilitate both the interview process and the translations prior to data entry. Because KAP surveys are subject to problems of association between categories of information [54], we incorporated a number of triangulation (within and between the pre- and post-event questionnaires) and verification questions, particularly when assessing farmers’ current management practices and perceptions of ecological engineering. The interviewers were also encouraged to record qualitative information to aid in cross-checking farmer responses. To avoid possible cultural biases [54] in interpreting the responses, the local interviewers who conducted the surveys also assisted in coding responses prior to global project data entry. During the OFD welcome sessions and, again, prior to each interview, the farmers were informed about the general objectives of the interviews and how the data would be stored and used and were advised that they were not obliged to answer any questions. The farmers’ names were only recorded to match the completed pre- and post-event questionnaires and have since been deleted from all records.

For the purpose of the present study, the farmers’ profiles (i.e., age, gender, education farming experience, and incomes other than rice), (current) management of their rice bunds, and willingness to continue or establish vegetation on rice bunds were taken from the responses to the pre-event questionnaire. Information related to farmers’ perceptions of the HDVPs (what they liked or did not like) and their opinions for improving the technology were taken from their responses to the post-event questionnaire. Indicators of farmer willingness to grow vegetation on bunds for pest management and reduce pesticide use, as influenced by the OFDs, were also taken from the post-event questionnaire and were based on expressed reservations by farmers about ecological engineering and/or HDVPs. The number of valid responses to each question is indicated in the result tables. Information retrieved using the pre-event questionnaire that is related to farmers’ rice pest and weed management, farm diversity, and factors that influence pesticide use on rice and field vegetables has been presented in a related paper by Horgan et al. (2023) [18]. The present study focuses only on bund management and farmers’ willingness to adopt ecological engineering practices, for which information was largely retrieved using the post-event questionnaire and has not been reported previously.

2.4. Data Analysis

Farmers’ responses were analyzed with region as the main factor. Numerical variables were analyzed using univariate general linear models (GLMs). We used Tukey’s post hoc tests to identify homogenous region groups. Categorical and binary variables (e.g., educational attainment, growing vegetables on bunds, growing flowers on bunds) were analyzed using 2 × 2 contingency tables. Tests of mutual and partial independence were conducted for any significant associations using χ² analyses. In cases where farmers indicated more than one category for the categorical variables, we used multivariate GLMs on the data ranked by the subject (farmer) and Tukey’s tests to indicate homogenous groups. To assess perceptions of insecticide-associated risks to health and the environment, we asked the farmers to indicate which category best reflected insecticides—1, always harmful; 2, sometimes harmful; 3, have no effect; 4, sometimes beneficial; and 5, always beneficial to their own health, to the health of their family members, their neighbors’ health, water quality, and non-target animals, thereby creating a 5 × 5 matrix of responses. The average score per farmer was reported as a “ranked risk perception score” (1 = highest perception of possible risks, 5 = lowest perception of possible risks). The ranked risk perceptions were analyzed using GLMs. We used Friedman’s 2-way ANOVAs by ranks to compare the frequency distributions of vegetables and flowers grown by farmers in different regions and compare farmers’ preferred vegetables before and after the OFDs. We used binary logistic regression with backward elimination to determine the best predictors of farmers’ current practices and willingness to plant vegetables or flowers on their bunds or establish HDVPs in the future. The residuals were examined following all parametric analyses to verify their normality and homogeneity.

3. Results

3.1. Farmer Profiles

The farmers we interviewed were mainly men, >50 years, and with >20 years of rice farming experience. The farmers in Iloilo tended to be older, and those in Bukidnon were generally younger (F_3,291 = 4.518, p < 0.001; Table S2). Proportionally more women were interviewed in Iloilo (65.8% compared to <25% at the other sites; χ² = 42.570, p < 0.001). All the participants were rice farmers, and a large percentage (>87%) depended on rice production as their main source of income. Nevertheless, a large number of farmers (>74%) also produced other crops or livestock or had other jobs (Table S2). The farmers grew these other crops for markets and home use: more farmers in Bukidnon produced other crops for sale at markets compared to the other sites (ca 40% compared to <20% for the other sites: χ² = 27.360, p < 0.01), while fewer farmers in Laguna produced other crops for markets (10%) or in general compared to the other sites (44% compared to >82% at the other sites; χ² = 44.604, p < 0.001) (Table S2).

The farmers in Rizal and Iloilo applied more insecticides to their rice crops (means ± SEM: 2.11 ± 0.11 and 1.97 ± 0.13 per rice crop, respectively) compared to those in Laguna (1.37 ± 0.14) and Bukidnon (1.16 ± 0.12) (F_3,270 = 9.995, p < 0.001; Table S2). Consequently, fewer farmers in Rizal and Iloilo were producing insecticide-free rice crops at the time of the survey (13.7% and 9.3%, respectively; χ² = 63.449, p < 0.001) or during the previous six years (3.2% and 7.1%, respectively; χ² = 29.905, p < 0.001) compared to the farmers in the other sites (>24% and >10% for current and previous years, respectively). In general, the farmers in Iloilo had fewer reservations regarding the environmental and health impacts of pesticides (ranked risk perception score = 2.26 ± 0.14), and those in Bukidnon had the greatest reservations (1.07 ± 0.03; F_3,291 = 25.095, p < 0.001; Table S2). Further details concerning the farmers’ profiles and their rice production practices are presented in Table S2, and further related information has been presented by Horgan et al. (2023) [18].

3.2. Farmers’ Current Management of Bunds

3.2.1. Bund Dimensions and Herbicide Use

Farms at all sites were divided into plots with an average area of ca 0.2 ha (the plot size in Laguna was not reported). The surrounding bunds were ca 30 cm wide at all sites and averaged between 27.9 and 34.3 cm in height, with the lowest bunds reported in Iloilo and the highest in Bukidnon (

Table 1. Responses to questions concerning farmers’ current management of rice bunds.

Questions	Sub-Categories	Sites 1				Test Statistics 2	DF	Valid Cases
		Laguna	Rizal	Iloilo	Bukidnon
Average plot size (ha)		na	0.17 ± 0.04	0.24 ± 0.03	0.23 ± 0.03	1.232 ns	2	216
Average width of bunds (cm)		29.96 ± 1.49	32.97 ± 1.60	33.36 ± 1.55	31.57 ± 1.35	0.935 ns	3	270
Average height of bunds (cm)		28.62 ± 1.37 ab	27.85 ± 1.87 a	33.81 ± 1.32 ab	34.32 ± 2.20 b	3.884 **		270
How many herbicide applications are made to bunds each season?	Zero	87.04	77.78	86.32	83.58	3.581 ns	6	270
	1×	7.41	16.67	10.53	13.43
	2×	5.56	5.56	3.16	2.99
Have you heard of planting vegetables on rice bunds? (% yes)		70.37 a	88.89 b	93.68 b	93.85 b	23.688 ***	3	268
Have you planted vegetables on your rice bunds? (% yes)		33.33 a	74.07 c	76.84 c	53.85 b	30.720 ***	3	268
If you plant vegetables on bunds, do you apply insecticides? (% yes)		55.56 ab	70.00 b	40.00 ab	28.57 a	18.019 ***	3	166
If you currently do not plant vegetables on your bunds, would you consider planting in future? (%)	Yes	14.81 a	54.54 c	35.71 b	60.00 c	20.453 ***	6	82
	No	62.96	45.45	64.29	36.67
	Maybe	22.22	0.00	0.00	0.00
Why would you plant vegetables on your bunds (%)	Extra food	25.00	40.00	100.00	63.16	2.036 ns	3	34 3
	Extra income	50.00 ab	25.00 ab	60.00 b	15.79 a	3.148 *	3
	Pest management	0.00	12.00	0.00	21.05	0.685 ns	3
	Other benefits 4	25.00	13.00	0.00	10.53	1.772 ns	3
Have you heard of planting flowers on rice bunds? (% yes)		40.74 b	24.07 a	27.37 a	81.54 c	56.995 ***	3	268
Have you ever planted flowers on your rice bunds? (% yes)		7.40 a	14.81 a	20.00 b	27.69 b	8.749 *	3	268
If you currently do not plant flowers on your bunds, would you consider planting in future? (% yes)		32.00 a	36.96 a	46.67 b	65.96 c	134.092 ***	3	219
Why would you plant flowers on your bunds (%)	Extra income	27.27 ab	0.00 a	11.43 a	41.94 b	5.970 ***	3	96
	Pest management	45.45	76.47	77.14	51.61	1.938 ns
	Other benefits 5	27.27	23.53	8.57	9.68	2.631 ns
Do you allow weeds/wild flowers to grow on your rice bunds? (% yes)		9.26 a	20.37 b	34.74 c	15.38 a	15.743 ***	3	268
If you currently do not allow weeds/wild flowers on your bunds, would you consider allowing them in future? (% yes)		22.22 a	37.21 a	26.23 a	50.91 b	18.835 ***	3	213

¹: Numbers are means ± SEM or percentages, as indicated; na = data not recorded; lowercase letters indicate homogenous site groups based on Tukey’s LSD tests for continuous variables and χ² partial analyses for contingency tables. ²: Test statistics are F-values for continuous variables and χ² for frequencies (converted to percentages for the table); ns = p > 0.05, * = p < 0.05, ** = p < 0.01, and *** = p < 0.005. ³: Results of multivariate GLM because farmers indicated multiple reasons (Wilks’ λ = 2.075, p = 0.031). ⁴: Other benefits of planting vegetables on bunds, as mentioned by farmers = to optimize space and as a barrier to livestock. ⁵: Other benefits of planting flowers on bunds, as mentioned by farmers = to attract pollinators, to suppress weeds, and for esthetics.

). In Iloilo, 72.6% of the farmers mowed their bunds (we did not ask farmers at the other sites whether they mowed bunds). About 16% of the farmers applied herbicides to their bunds (Table 1).

3.2.2. Growing Vegetables on Rice Bunds

Among the farmers we interviewed, 88% had heard of planting vegetables on rice bunds; the farmers in Laguna were less familiar with the concept (Table 1). About 46% of the farmers produced vegetables on their bunds. These farmers mainly grew stringbean, ladyfinger, mungbean, eggplant (Solanum melongena L.), and bitter gourd. The frequency distributions of preferred vegetables were similar across the sites (p = 0.208) (Figure 1). Vegetables were mainly produced on bunds to supplement household food (67%) and incomes (24%); only 2.6% of the farmers indicated that they grew vegetables for rice pest management (Table 1). A total of 46.5% of the farmers applied insecticides to their bund-grown vegetables. A higher percentage of farmers in Rizal (70%) applied insecticides compared to Bukidnon (28.6%) (Table 1).

Of the farmers who did not grow vegetables on their bunds, only 21% suggested that they might consider doing so in the future. Compared to the other locations, more farmers in Bukidnon (60%) and Iloilo (54.5%) suggested that they would consider growing vegetables (Table 1). The farmers saw the vegetables as a source of extra food (62%) or income (29.4%), alongside pest management (14.7%) and other benefits (11.8%). The main reasons given by the farmers for not planting vegetables were that their bunds were too narrow (33%), that the vegetable plants might attract pests (16.7%), and that their bunds were used as pathways (14%) (Table 1).

3.2.3. Growing Flowers on Rice Bunds

Among the farmers we interviewed, 42.5% were familiar with the concept of growing flowers on rice bunds, and 18.3% were already growing flowers on bunds (Table 1). The farmers in Bukidnon (81.5%) were more familiar with the concept than the farmers at the other sites, and a higher percentage of Bukidnon (27.7%) and Iloilo (20%) farmers currently planted flowers on their bunds compared to the farmers in Laguna (7.4%) and Rizal (14.8%). All the farmers (100%) who planted flowers did so for pest management (Table 1).

Marigolds (Tagetes spp., 44%) and cosmos (Cosmos spp., 34%) were the most commonly planted bund flowers. There was no statistically significant difference between the frequencies of preferred flowers across the sites (p = 0.454) (Figure 2). Overall, 45% of the farmers who did not plant flowers indicated that they were willing to do so in the future. A greater percentage of Bukidnon (66%) and Iloilo (47%) farmers compared to those at the other sites suggested that they would consider growing flowers on their bunds; 65% of these farmers indicated pest management as their reason for planting flowers, but 21% (42% in Bukidnon) felt that flowers could represent a further source of household income (Table 1). The main reason for not planting flowers was that the farmers did not see what benefits they would bring (27.4%). The farmers also felt that their bunds were too narrow (14%) or had other uses (i.e., pathways, 15%) or that the flowers might damage the bunds (14%). About 2% of the farmers felt that the flowers would attract pests (Table 1).

3.2.4. Allowing Weeds and Wild Flowers to Grow on Rice Bunds

Only 22% of the farmers, mainly in Iloilo and Rizal, allowed weeds and wild flowers to grow on their bunds. This was mainly for pest management (66%), as forage for carabao or other livestock (13.5%), or as a refuge or food source for pollinators (10%). Of those farmers who did not allow weeds/wild flowers on their bunds, 33.8%, mainly in Bukidnon (51%), were willing to do so in the future for pest management (60.7%), to strengthen the bunds (9%), or to attract pollinators (7%) (Table 1).

3.3. Farmers’ Opinions about Ecological Engineering after the Open Field Days

The farmers were generally enthusiastic about ecological engineering by the end of the OFDs: 95% of the farmers indicated that they would consider growing vegetables or flowers on their bunds and saw the benefits in terms of extra food (68.7% of the farmers) and income (52%) and/or pest management (50%). The farmers preferred to grow vegetables on their bunds, with >92% of the farmers in Laguna, Iloilo, and Bukidnon preferring vegetables and 78% in Rizal preferring to plant vegetables (

Table 2. Responses to questions during the post-event interviews concerning vegetable production on rice bunds ¹.

Questions	Sub-Categories	Sites 2				Test Statistics 3	DF	Valid Cases
		Laguna	Rizal	Iloilo	Bukidnon
What type of vegetation would you prefer to grow on bunds (%)	Vegetables	92.45 b	78.00 a	96.70 b	98.25 b	26.232 ***	6	251
	Flowers	0.00	6.00	3.30	1.75
	Both	7.55	16.00	0.00	0.00
Why would you grow vegetables on your bunds? (%)	Extra income	58.49	63.46	53.76	40.74	2.376 ns	3	275 4
	Extra food	54.72 ab	51.92 a	74.19 b	62.96 ab	3.695 *	3
	Pest management	16.98 ab	11.54 ab	7.53 a	23.46 b	3.262 *	3
	Other benefits 5	1.89	0.00	1.08	3.70	0.984 ns	3
	Because advised 6	3.77	1.92	0.00	1.23	1.141 ns	3
How would you manage insect pests on your bund vegetables? (%)	Without insecticides	74.07	56.36	76.34	76.83	13.585 ns	12	284
	With insecticide	16.67	30.91	16.13	14.63
	Using concoctions	3.70	5.45	5.38	6.10
	Biocontrol/agroecology	3.70	1.82	1.08	1.22
	Cultural/physical	1.85	5.45	1.08	1.22
How would you manage weeds on your bund vegetables? (%)	Without herbicide	75.61 b	58.18 a	56.99 a	75.61 b	27.743 ***	9	271
	With herbicide	21.95 c	12.73 b	8.60 a	7.32 a
	Using concoctions	2.44	0.00	1.08	1.22
	Biocontrol/agroecology 7	na	na	na	na
	Cultural/physical	0.00 a	29.09 b	31.00 b	15.85 b
How would you manage plant diseases on your bund vegetables? (%)	Without fungicide	66.67 a	85.19 b	58.06 a	64.63 a	21.309 *	12	282
	With fungicide	12.96	16.67	24.73	18.29
	Using concoctions	3.70	0.00	7.53	10.98
	Biocontrol/agroecology	1.85	0.00	1.08	0.00
	Cultural/physical	11.11	0.00	8.60	6.10
Will you need extra help/labor to manage bund vegetables (% yes)		51.92 a	55.77 a	75.27 b	78.21 b	16.577 ***	3	276
Who is most likely to carry out the extra labor (%)	Hired worker(s)	37.04	34.48	47.14	31.15	10.922 ns	6	187
	Family member(s)	66.67	55.17	51.43	67.21
	Extension support	3.70	10.34	1.43	1.64
Do you anticipate any negative effect of bund cropping? (% yes)		18.52	14.81	5.75	11.39	5.938 ns	3	274
What negative effects do you anticipating?	Added drudgery	65.00	28.57	14.29	56.25	12.654 ns	9	50
	Lack of Knowledge	7.00	21.43	14.29	25
	Added Costs	7.00	7.14	0.00	12.5
	Limits to Implementation 8	21.00	42.86	71.43	6.25

¹: Few farmers were willing to plant flowers on their bunds; therefore, the information in this table principally concerns vegetables. ²: Numbers are percentages; lowercase letters indicate homogenous site groups, based on Tukey’s tests for ranked preferences and χ² partial analyses for contingency tables. ³: Test statistics are F-values for multivariate GLM and χ² for frequencies (converted to percentages for the table); ns = p > 0.05, * = p ≤ 0.05 and *** = p ≤ 0.005. ⁴: Results are for multivariate GLM because each farmer indicated multiple reasons (Wilks’ λ = 1.948, p = 0.017). ⁵: Other benefits of planting vegetables on bunds, as mentioned by farmers = to optimize space, to strengthen bunds, to suppress weeds, and to attract pollinators. ⁶: “Because advised” refers to advice received outside the OFDs. ⁷: na = no farmers mentioned biological control or agroecology as approaches to manage weeds. ⁸: Limitations to implementation as mentioned by farmers = limited space, disturbs rice management operations such as plowing, difficult to manage water on bunds, pesticide drift from rice will affect vegetables, issues with rodent management, and bunds used as pathways.

). The planting of vegetables was seen as a source of extra income (61%) and/or food (54%). Only ca 15% of the farmers mentioned growing vegetables for pest management, with higher percentages in Bukidnon (23.5%) than Iloilo (7.5%) of farmers mentioning pest management (Table 2).

The farmers’ preferences for vegetable species during the post-event interviews were similar across the sites (p = 0.629) and to the frequency distributions of species currently grown by farmers as indicated during the pre-event interviews (p = 0.356) (Figure 1). When asked how they might manage bund-grown vegetables, 71%, 66.6%, and 68.6% of the farmers indicated that they would avoid insecticides, herbicides, and fungicides on their bunds, respectively. However, 19.6%, 12.7%, and 18.2% of the farmers mentioned that they would apply insecticides, herbicides, and fungicides, respectively, on the bunds. A higher percentage (31%, not significant) of the farmers in Rizal indicated that they would use insecticides if needed, and, in Laguna, a higher percentage (22%) than at the other sites mentioned that they would use herbicides if needed (Table 2).

After viewing the demonstration plots, 63.5% of the farmers had no suggestions to improve the HDVP system (Table S4); however, 21.7% suggested that the configuration of patches should be changed either by changing the density of the patches, changing their size, growing vegetables adjacent to the rice—but not on bunds—or reverting to strips of vegetables on bunds. A total of 10.5% suggested that they would change the composition of the plants grown by choosing species more adequate to each region or which produce a marketable product or by planting only one species per patch. Finally, 4.3% of the farmers made suggestions related to the management of the patches, such as using organic fertilizers, using pest control (sometimes including pesticides), removing the bamboo stakes, or adding trellising.

Across regions, 65.3% of the farmers indicated that they would need extra help/labor to install and manage the patches (>70% in Iloilo and Bukidnon), and this would likely be carried out by family members (60% of the farmers), with nearly 80% indicating that help would come from the family unit (spouse or children); 37.5% suggested that they would need to hire extra labor (Table 2). A total of 12.6% of the farmers across the sites identified potential negative effects of HDVPs and growing vegetables on rice bunds in general. These potential issues included the extra work and difficulties required for the system (42.3% of the respondents), their lack of the knowledge required for the system (17.7%), the extra costs involved (7.4%), and limitations to implementing the technology (40.1%) due to narrow bunds, rats, potential theft, water, and other issues (Table 2).

3.4. Farmers’ Willingness to Grow Vegetation on Bunds before and after the OFDs

During the pre-event interviews, 39% of the farmers who did not currently grow bund vegetables indicated that they were willing to grow vegetables on their bunds; a higher percentage of Bukidnon farmers (61.5%) and a lower percentage of Laguna (11.4%) farmers were willing to plant vegetables on bunds (

Table 3. Farmers’ willingness to adopt ecological engineering practices on their rice farms based on pre- and post-event interviews.

Willingness to Adopt Ecological Engineering	Farmer Category Based on Pre-Event Interviews	Sites 1				χ2 Region 2	χ2 Category 2	Valid Cases
		Laguna	Rizal	Iloilo	Bukidnon
Pre event
Willing to grow vegetables on bunds (% yes)	Not growing flowers or vegetables 3	11.43 a	44.44 b	38.46 b	61.54 c	17.032 ***		83
Willing to grow flowers on bunds (% yes)	Not growing flowers or vegetables	25.71 a	18.18 a	23.81 a	53.85 b	7.914 *	9.534 ***	234
	Growing vegetables	41.18	45.71	54.55	50.00	5.302 ns
Post event
Willing to grow vegetables on bunds (% yes)	Not growing flowers or vegetables	65.71	61.54	73.91	82.98	4.422 ns		118
Willing to adopt HDVPs (% yes)	Growing vegetables	75.00	74.42	87.67	68.42	6.540		174
Insecticide on bunds (% yes)	No insecticide on bunds	10.00	20.00	9.09	13.79	1.358 ns	5.861 *	167
	Insecticide on bunds	20.00	35.71	26.09	25.00	1.940 ns
Herbicide on bunds (% yes)	No herbicide on bunds	18.75 b	21.88 b	3.13 a	6.25 a	10.484 *	0.429 ns	173
	Herbicide on bunds	0.00	18.18	22.22	0.00	2.156 ns

¹: Numbers are percentages; lowercase letters indicate homogenous site groups based on χ² partial analyses for contingency tables. ²: ns = p > 0.05, * = p ≤ 0.05, and *** = p ≤ 0.005. ³: Excluding farmers who answered “maybe”.

). The percentage of farmers who had not planted vegetables or flowers on bunds before the OFDs but were willing to plant on bunds after the events increased from 36.3% to 74.4% (χ² = 30.435, p < 0.001); this was based on the farmers who expressed no reservations about the technology (otherwise, ca 95% of the farmers in all categories mentioned that they were willing to adopt the technology). During the post-event interviews, 76.4% of the farmers who planted vegetables on bunds indicated that they would consider growing vegetables in HDVPs, with no statistically significant differences between the farmers from the four regions (Table 3).

Before the OFDs, a higher percentage of the farmers (pre event 47.9%) who grew vegetables on bunds were willing to grow flowers on the bunds compared to the farmers who did not grow vegetables on their bunds (30.4%). More farmers in Bukidnon (51%) were willing to grow flowers compared to farmers at the other sites (Table 3). After the OFDs, very few of the farmers indicated a willingness to grow flowers on their bunds; this was because of a strong preference for growing vegetables instead. Only three farmers mentioned growing flowers together with vegetables.

After the OFDs, a higher percentage of the farmers who used insecticides (26.7%, based on the pre-event interviews) compared to those who used no insecticide on their bund vegetables (13.2%) indicated that they would consider using insecticides on bund vegetables if needed. There were no statistically significant differences between the regions in the farmers’ willingness to use insecticides on the bunds (Table 3). After the OFDs, 11% of the farmers (who used and did not use herbicides according to the pre-event interviews) considered using herbicides if needed on their bund vegetables. More farmers in Laguna (18.8%) and Iloilo (21.9%) indicated that they would use herbicides on the bunds (Table 3), despite many indicating, in the pre-event interviews, that they did not currently use herbicides on their bunds.

The best predictors for the current planting of vegetables and/or flowers on bunds were the survey region (higher in Rizal and Iloilo), planting other crops (higher among those who planted other crops), farmer age (higher among older farmers), and using biocontrol in rice (higher among the farmers who used biocontrol) (

Table 4. Results of binary logistic regressions indicating the best models for the predictors of farmers’ practices and their willingness to adopt ecological engineering.

Models and Dependent Variables	Predictor Variables 1	Log-Likelihood	DF	Valid Cases	p-Values
Currently grow vegetables or flowers on bunds	Region	28.736	3	270	<0.001
	Plant other crops	13.585	1		<0.001
	Age of farmer	4.812	1		0.028
	Apply biocontrol in rice	3.835	1		0.050
Willing to grow vegetables (pre event)	Region	21.563	3	90	<0.001
	Other incomes	12.438	1		<0.001
	Education achieved	5.498	2		0.064
	Rice farming experience	2.879	1		0.090
Willing to grow vegetables (post event)	Region	10.892	3	224	0.012
Willing to adopt ecological engineering/HDVPs	Plant other crops	5.743	1	133	0.017

¹: Based on binary logistic regression with backward elimination; the initial independent variables were (numerical, continuous) age, farming experience, land area, visits to farm per week, ranked risk perception score, number of pesticide applications to rice per season, (categorical) region, educational attainment, pesticide trends in last 6 years, (binary) gender, rice as main occupation, have other incomes, grow other crops, have grown vegetables in the last 5 years, and use biocontrol agents.

). Region (highest for Bukidnon) and other incomes (higher among those who had other incomes) were the best predictors of willingness to grow vegetables and flowers on bunds, with education (higher among those who advanced beyond primary school) and rice farming experience (higher among those with more experience) also included in the model (Table 4). Finally, the farmers who did not grow other crops were more willing to adopt HDVPs than those who currently planted other crops (Table 4).

4. Discussion

In the Philippines, two campaigns, Palayamanan [43] and ecological engineering [35], have likely influenced rice farmers to establish vegetation on their bunds. However, the use of bunds for vegetable production was also independently adopted by farmers outside the context of these two programs [55,56]. Based on our survey, 88% of the farmers were familiar with the concepts of planting vegetables or flowers on bunds (Figure 3). This broad category included farmers who may have seen their neighbors grow plants on bunds, farmers that had been introduced to the idea through training and extension, or that already used their bunds to plant mainly vegetables for extra food or income. Indeed, nearly 60% of the farmers already grew vegetables or flowers on their bunds, and a further ca 2% allowed weeds and wild flowers on their bunds, mainly as a refuge and food source for natural enemies and pollinators (Figure 3). Our results also indicated nuanced responses by farmers to the technologies outlined in the OFDs, marked by differing regional priorities (Table 1 and Table 2). Research has shown that ecological engineering can reduce pesticide inputs, thereby reducing farm expenditure, while, at the same time, benefiting wildlife and increasing the nutritional diversity of farming households [7,12,17,31,32,33,44]. We selected stringbean, mungbean, bitter gourd, and ladyfinger as the species to include in HDVPs based on interviews with local DA officers. Before the farmers had seen the demonstration plots, those who grew vegetables on their bunds indicated that they also planted these four species, with similar preferences across the four regions (Figure 1). During our OFD discussions, many farmers were particularly interested to gain further advice on the choice of vegetation to grow on their bunds for the most effective management of rice pests.

4.1. Farmers’ Preferences for Growing Vegetatables on Rice Bunds

The issue of what to grow on rice bunds has been a key topic of research in ecological engineering for rice pest management [33,41,57,58]. A number of studies have shown how flowering fruits and vegetables are associated with an increased diversity of natural enemies and lower pest damage to rice [14,31,32,33,58,59,60,61]. Many farmers were aware of these potential benefits and, therefore, planted flowers, but, because they did not recognize the same benefits from flowering vegetables (Figure 3), they were apparently unaware of the underlying mechanisms. Bund-grown vegetation is thought to benefit natural enemies in several different ways. Flowers may provide nectar food sources for the free-living stages of parasitoid wasps and predatory bugs that kill the eggs, larvae, or pupae of rice pests [57,58]. This mechanism is supported by observations that parasitoids, such as Trichogramma wasps, survive longer and achieve higher rates of parasitism in the presence of sugar and nectar food sources [58,60,62,63,64]. A number of field studies have convincingly shown increases in the mortality of rice pests such as planthoppers and stemborers in fields with ornamental flowers and flowering vegetables on their bunds [31,32,60,61]. A diversity of vegetation in the rice ecosystem may also provide natural enemies with physical structures [65] or alternative food sources during fallow periods when rice pests occur at relatively low densities. For example, wild grasses have been shown to support populations of alternative prey for key natural enemies in African and Asian rice fields [66,67,68]. Farmers can also establish trap plants that attract ovipositing stemborers, but on which the larvae fail to complete development [42,69,70]. Trap plants have been shown to effectively reduce stemborer populations and damage to maize and other crops in “push–pull” systems in Africa [71,72] and have reduced stemborer damage to rice in China [39,42].

The main flowers grown by the farmers were marigolds and cosmos (Figure 2). Marigolds have been noted previously to repel pest insects [73,74] and reduce pest damage to rice when grown together with other flowers on bunds in Bangladesh [31]. A number of papers have emerged in mainly regional journals to support the inclusion of cosmos in seed mixtures for rice pest management [31,75,76,77,78]. Flowers are unlikely to be treated with insecticides because they are rarely a source of extra income [17]; furthermore, they are easy to establish on rice bunds and can be grown relatively easily from seeds. However, considerably more research is needed to better direct farmers toward the ornamental flowers with the greatest benefits for pest management in specific crops. Evidence from Vietnam also suggests that the sustainability of ecological engineering using flower strips is threatened by apparently low returns on labor and costs (the farmers largely depended on the government to supply flower seeds) [17]. Difficulties among farmers in comprehending the impacts of flower patches or strips on rice pests represent a major obstacle for adopting this kind of intervention (mentioned by 27% of the farmers who were unwilling to adopt this strategy).

According to the information collected during the pre- and post-event interviews, the OFDs influenced the farmers’ opinions regarding planting on bunds in three main ways (outlined in Figure 3). Firstly, the farmers who were largely uninterested in growing vegetables or flowers on their rice bunds were now more willing to do so (Table 2 and Table 3). Based on those farmers who had mentioned no major obstacles to implementation, willingness increased from 36 to 74% over the course of the OFDs (Figure 3). The second major change was noted among the farmers who already grew vegetables on bunds: before the OFDs, only 2.6% of these recognized the pest management potential of their planted bunds, but, after the OFDs, 50% of farmers recognized that planting vegetables was beneficial in terms of pest management, and 15% listed pest management as a reason underlying their willingness to grow vegetables on bunds in future (Table 2). The third major shift was from a focus on growing flowers to a new focus on growing vegetables (Figure 3). Indeed, many of the farmers who already grew flowers on their bunds now considered shifting to growing vegetables due to their multiple benefits, including as a source of extra income and food, as well as for pest management (Figure 3). The greater attention by these Filipino farmers to vegetables compared to flowers on bunds was similar to that of Cambodian and Vietnamese farmers [17,33], who also saw a greater range of benefits from growing vegetables.

4.2. Avoiding Pesticides on Bund-Grown Vegetation

Initially, very few farmers (practicing and willing to practice) recognized the possible benefits of bund-grown vegetable plants for rice pest management (pre event: Figure 3). Perhaps as a consequence of growing vegetables and because pests were sometimes obvious, 46.5% of the farmers who grew vegetables on their bunds applied insecticides to said bunds (Table 1 and Table 3). Furthermore, although the farmers appeared more averse to using insecticides on their bunds after the OFDs, over 20% of the farmers indicated that they would still continue to use insecticides if needed; this percentage was higher among those who already used insecticides on their bunds (Table 3). A similar trend has been noted among rice farmers in Vietnam. In the Mekong Delta Region, government support encouraged thousands of farmers to plant flower strips on rice bunds for pest management. However, many farmers gradually transitioned to growing vegetables, and, of these, 24% used insecticides on their bund-grown vegetables [7,17]. The use of pesticides on ecological engineering rice bunds runs counter to the objective of restoring rice regulatory ecosystem functions because it potentially destroys alternative prey and/or natural enemies.

To avoid resorting to pesticides on rice bunds, farmers might be further encouraged to grow flowers instead of vegetables or integrate flowers with the bund vegetables [18]. Only ca 18% of the farmers we interviewed grew flowers on their bunds, but 100% of these did so for pest management purposes (Table 1, Figure 3). It is possible that farmers who grow bund vegetables would further avoid pesticides if they saw clear benefits from the vegetables in terms of rice pest management. The OFD demonstration plots showed the farmers that flowering bund vegetables, like ornamental flowers, could enhance natural enemies and thereby protect rice crops. We presented preliminary results from the dry season at each site to indicate that stemborer damage had, on average, declined in the presence of HDVPs and that planthopper mortality had increased (the preliminary data were presented in posters, and some have since been published by Vu et al., 2018 [60]). Furthermore, we presented the rice and vegetable harvests at the demonstration plots and compared these with harvests from each adjacent control plot, which had similar rice yields but produced no supplementary vegetables (see also Horgan et al. (2015) [35]). Despite these messages, the farmers still saw the vegetables mainly as a source of extra food and income, with only moderate increases in the percentage of farmers indicating the pest management potential of flowering vegetables as a reason for diversifying their farms. This may also explain why the farmers who currently did not use herbicides on their bunds indicated that they were more likely to use herbicides when growing bund vegetables (Table 3). Like insecticides, herbicide use was presumably carried out to protect any investments in bund vegetable production. Despite these trends, the percentage of farmers willing to grow vegetables and avoid any pesticides was generally high (i.e., >66% of all the farmers interviewed after the OFDs, Table 2).

Based on the pre-event survey, we noted that the farmers who grew vegetables for home consumption (either in field plots or on bunds) were less likely to use pesticides than those who produced vegetables for the market [18]. Similarly, among the farmers in the Mekong Delta Region, there was a strong positive relation between growing bund vegetables for extra income (as opposed to growing them as food for home use) and the use of insecticides to manage potential vegetable pests [17]. Therefore, encouraging vegetable production for improved household nutrition and not for markets could reap multiple benefits in terms of farm diversification without added pesticide inputs and with possibilities of reducing the overall pesticide use on rice. However, we also noted that farmers from certain regions (i.e., Rizal, Laguna, and Iloilo in the present study) may be less averse to using insecticides (i.e., lower ranked risk perception scores) and less conscious of their associated hazards (Table S2; see also Horgan et al. (2023) [18]). These farmers were associated with higher pesticide use in this and a previous study (Table 1, [18]), and the farmers in Rizal made relatively frequent pesticide applications even to vegetables grown for home use [18]. These results suggest that the best approaches to ensuring that farmers continue to diversify their rice fields but do not use pesticides on their bunds is to strengthen their knowledge of the harms caused by pesticides to beneficial arthropods, the restoration services provided by vegetables (see also Horgan et al. (2023) [18]), and the incompatibility of the two. Policymakers should recognize that knowledge in all these areas will increase the sustainability of implementing ecological engineering and, thereby, could strive to improve the links between research in agronomy and agroecology for pest management. Furthermore, the common objectives of environmental and agriculture ministries could be emphasized under future national biodiversity action plans, for example, by promoting set-aside plots for both the natural enemies of rice pests and other wildlife (see below). This would require greater intergovernmental dialog during the search for alternatives to chemically intensive rice production.

4.3. Farmers’ Appreciation of HDVPs and Other Interventions

The OFDs clearly improved farmers’ knowledge of ecological engineering, which, together with the necessary skills, is an essential component of behavioral change [79]. During the post-event evaluations, the farmers responded best to the demonstration plots compared to the other aspects of the OFDs and called for further events with more details on related topics that had not been addressed, such as organic rice production and other sustainable pest management options (Table S4). This suggests that many of the farmers were not yet confident in their skills and, therefore, although willing, were not yet capable of implementing ecological engineering. Our demonstration plots were based on establishing HDVPs [35]. Most of the farmers (ca 64%) made no suggestions to improve the HDVP system after the OFDs. Among those who gave advice, the main concerns were related to the configuration of the patches (Table S4). Many of these farmers already had experience in growing vegetables on rice bunds, and all the suggestions involved reducing the work involved in establishing patches, either by reducing the density of patches, reducing their size, or moving patches to dry ground adjacent to the rice crop but not on the bunds. Furthermore, a number of farmers preferred growing vegetables as strips along their rice bunds. All these scenarios can probably be implemented without losing the most tangible (see below) benefits of diversifying the habitat, although issues around the scale effect of bund vegetables require considerably more research attention. In areas prone to rat damage, long strips of dense bund vegetation are not advisable [35,47,80].

Without more detailed knowledge of the mechanisms, the main impact of bund crops may simply be in their encouragement of farmers to spray less and thereby conserve natural enemies [7,17,32]. This implies that biologically relevant scale effects may be largely inconsequential in the face of severe pesticide-related hazards for biodiversity. Several studies have already shown that pesticide reduction is a major contributor, or sometimes the main contributor, to the benefits of flower strips in rice [14,31,33,61]. It is therefore important to recognize and act on farmers’ suggestions to considerably reduce the costs and efforts required to establish HDVPs or other systems (e.g., vegetation strips) that incorporate diverse habitats. One suitable alternative might be to use dry, set-aside areas near rice fields as diverse habitat patches that include flowers and trap plants or robust vegetable species, much like seed mixtures are used as a habitat for pollinators [81,82]. Linking pollinator conservation and the conservation of natural enemies could also benefit from the considerable research that has already been conducted for the conservation of insect pollinators [83,84,85], as well as the large advances that have been made in implementing citizen science-based pollinator conservation practices [86,87].

Willingness implies a positive attitude towards the proposed technologies. The high willingness of the farmers in our study, although measured in the short term around OFDs, suggests that they were motivated to adopt ecological engineering [88] despite the reservations of some farmers about some of the practices. Future work is required to compare the attitudes of adopters and non-adopters in the longer term. Furthermore, it would be counterproductive to ask farmers to adopt any intervention (such as establishing flower strips or vegetation patches) without proof of concept or due diligence in ensuring that the cost–benefit ratios favor the farmer. For example, the farmers we interviewed indicated that the work to establish and maintain bund crops would require extra help and that most of this would come from the immediate family, often the female spouse (Table 2). Vo et al. (2015) [12], in an assessment of the benefits of flower strips for rice farmers in Vietnam, found that these strips increase farm profitability by reducing pesticide inputs. Furthermore, Sattler et al. (2021) [33] found that ecological engineering was more profitable for Cambodian rice farmers than insecticide use. However, farmers might not perceive such benefits, as shown by what occurred among some ecological engineering farmers in Vietnam, who were dissatisfied with the intervention because of the labor required to establish flower strips [17]. While planting vegetables gives immediate returns in terms of extra food or income [14,32,33,39,44,59], the largescale planting of ornamental flowers must be more carefully considered because the financial benefits can be more abstract and difficult to perceive. Indeed, studies have shown that the financial benefits of ecological engineering compared to the chemical management of rice pests can be similar to control plots without bund vegetation or pesticides [14,33].

5. Conclusions

A high percentage of the farmers we interviewed already produced vegetation on their rice bunds for food, income, and pest management. The farmers who grew vegetables on their bunds often applied pesticides, including insecticides, to their bund crops, which is incompatible with ecological engineering. During the OFDs, the farmers gained an appreciation of the pest management potential of bund-grown vegetables, and many were willing to avoid pesticides on their bunds in the future. The farmers’ preferences for bund-grown plants were highly consistent across the four regions and before and after the OFDs. The farmers overwhelmingly preferred planting vegetables to planting flowers on their bunds. Further research into pest- and disease-resistant vegetables for planting on bunds is warranted if pesticide applications are to be avoided. The farmers indicated that considerable costs and labor would be required to establish and manage HDVPs, and, despite their benefits in terms of added food and income for the farming households, alternative configurations were recommended by the farmers to alleviate any related drudgery. The farmers’ willingness to reduce pesticide applications by adopting ecological engineering demonstrates that the planting of flowers or vegetables on bunds allows farmers to be pro-active in pest management and, therefore, responds to their risk aversions.