4. Discussion

4.1 Pesticide Use

We interviewed 174 farmers to obtain an impression of pesticide use on the farmed crops. The majority (154) was using pesticides and the most frequently used pesticide groups were insecticides and fungicides. Most of the pesticides that we found in our survey were introduced on the market in the 20th century with the insecticides Chlorantraniliprole (2008) and Spinetoram (2007) as notable exceptions. Fourteen of the 31 pesticides that we identified were also reported by Sekiyama et al. (2007) who performed a study on the use of pesticides in the Citarum River Basin in 2006. The widest used pesticides in our survey were Profenofos (in 13 of 21 crop types) and Mancozeb (in 15 of 21 crop types) which is in line with the results of Sekiyama et al. (2007) who reported 13.5% and 24.3% of their respondents using these two pesticides, respectively. Of the 10 most frequently used pesticides reported by Sekiyama et al. (2007), we did not find Permethrin (insecticide), Spinosad (insecticide), Iprodione (fungicide), Dimethomorph (fungicide) and Bacillus thuringiensis (biological). This illustrates the dynamic nature of pesticide use which is governed by a variety of factors such as supply by industry, authorization by the government and farmer-specific considerations (Mariyono et al. 2018).
The average pesticide usage was influenced by the frequency of application on each crop type. The frequency of pesticide application on vegetables was highest (7-10 times/ month) while for rice the lowest (1-3 times/growing season). The annual average of pesticide usage in UCRB range from 2.10-4 kg/ha (Brodifacoum on rice) to 32.2 kg/ha (Chlorothalonil on tomato). On average, 24.6 kg/ha pesticide is applied annually on UCRB agricultural land, which is lower than Bahamas and Mauritius with 59.4 kg/ha and 25.5 kg/ha, respectively (Ly, 2013). But it is relatively higher compare to other Asian countries, such as 14 kg/ha in China (Yang et al., 2014), 7.2 kg/ha in Malaysia, 13.1 kg/ha in Japan, and 0.2 kg/ha in India (Ly, 2013). This high estimation is plausible because our study area represents a densely populated and intensively farmed landscape.
Maggi et al. (2019) estimated crop-specific pesticide use (kg/ha) globally. When comparing overlapping crop types and pesticides used in Maggi et al.(2019) and our study, we notice a mismatch: for rice and corn all applied pesticides differ; for cabbage we share one common pesticide (Chlorothalonil); Chlorpyrifos and Azoxystrobin are also present in Maggi et al.(2019) but for different crops. We conclude that pesticide use is very region specific and are not sure a global map of pesticide use distribution is representative for actual use.
Our results on prescribed versus actual use on rice show that farmers use pesticides for rice that are not recommended for rice farming. Most types of pesticides are used (per hectare or as diluted with water) more than the lowest recommended amounts; about a quarter are used more than the highest recommended amount. For rice farmers in Sulawesi, Indonesia, Batoa et al. (2019) found that the prescribed frequency (influencing use-per-hectare) and dose were followed by about 1/3 of the interviewed farmers, while 2/3 deviated from recommended frequency and dose in both higher and lower than recommended. Zhang et al. (2015) reported under- and overuse for Chinese farmers for various crops. Mariyono et al. (2018) reported overuse on Java Island, Indonesia, but they did not specify the pesticide type. A study by Fan et al. (2015) in China showed that most of the farmers surveyed lacked the ability to understand the instruction manuals and pesticide labels. Additionally, the farmers often failed to select an appropriate pesticide to resolve a specific pest problem (Akter et al., 2018). These kinds of problems are also common in other agricultural areas (Fan et al., 2015; Houbraken et al., 2016; Akter et al., 2018). It stresses the importance of having transparent national pesticide usage guidelines and training farmers thoroughly in pest management, i.e. the diagnosis as well as the application of pesticides and alternative pest control strategies.
The survey showed that some rice farmers still used Endosulfan, usually to control stem borers, and green and brown leafhoppers (Fabro & Varca, 2012; Derbalah et al., 2019). Endosulfan is an organochlorine compound that was internationally banned in 2011 via the Stockholm Convention (UNEP, 2011; Balmer et al., 2019). Another banned insecticide found in the survey was Chlorpyrifos. The use of Chlorpyrifos in Indonesia is banned in rice agriculture (Ministry of Agriculture Republic Indonesia, 2011; Ministry of Agriculture Republic Indonesia, 2015). Sousa et al. (2018) found that concentrations of Chlorpyrifos and Endosulfan in most developing Asian countries, e.g. India, exceeded the values of the European Environmental Quality Standards (EQS) suggesting potential harm for aquatic ecosystem. Therefore, it is very important to monitor and enforce the usage guidelines, especially for these two pesticides.

4.2 Gathering Usage Statistics

Public availability of pesticide use data is generally scarce, i.e. because of proprietary data issues, poor registration and lacking regulations. Sales statistics in combination with recommended use of national institutions offer some insight in the types and amounts of pesticides used, but such data are generally only available at higher spatial scales. More detailed pesticide use statistics are needed for local environmental risk assessments, consumer protection (guiding residue monitoring), operator protection (improving or optimizing use) and monitoring the potential movement of pesticides into water (Eurostat, 2008). For example, our results show that farmers do not always apply the pesticides to the prescribed crop types. Secondly, the amounts applied vary, sometimes exceeding the highest recommended dose. In some cases, brands containing the same pesticide are applied simultaneously. Finally, the frequency of application also varies per farmer.
Although pesticides are among the most toxic substances released into the environment, very little public information is available on their use patterns, especially at the level of brands, active ingredients and at refined spatial scales. Information on which pesticide is used where and when, and in what quantities, is essential for protection of human health and the environment, as well as for effective pest management. In our opinion, a data should be public because people have a right to know when, where, and how pesticides are being applied so that they can take the appropriate measures to protect themselves and the environment. Accurate information on pesticide use enables better risk assessments and supports the identification of problematic use practices so they may be targeted for developing alternatives (PAN Germany, 2003). Comparison of our results with a previous study on pesticide use in the UCRB (Sekiyama et al. 2007) shows considerable differences in pesticide use over time, indicating that results of single surveys are representative for a limited timeframe only. Gathering representative data over a longer timeframe requires the establishment of a pesticide use reporting system. California’s pesticide-use reporting system represents the largest undertaking of this kind, and can act as a model for future pesticide disclosure programs (CDPR, 2000).

4.3 Reducing Pesticide Use

Our results may be used to identify management options for reducing the pesticide use. For example, the results show that crops like tomatoes, chili and cabbage require more pesticides than rice, cauliflower and eggplant. Also, Mariyono et al. (2018) reported that pesticide use even differs between local varieties and cultivated varieties within a crop type, where local varieties need more pesticides. Managers may consider to stimulate the production of crops, or crop varieties, that demand less pesticides. Another option is to replace more toxic pesticides by less toxic alternatives. However, most of the pesticides used in the UCRB fall in WHO class 5 (“may be harmful if swallowed”), with only a few pesticides falling in categories 2 or 3 (“fatal/toxic if swallowed”; IPCS, 2010). A more refined identification of management interventions would be possible if we would understand why farmers choose various pesticides, why they use the dosages and application frequencies as they do and sometimes overrule the prescriptions. In Sulawesi, Indonesia, Batoa et al. (2019) found that 73% of rice farmers interviewed state to know the use rules, whereas about 27% knows little or nothing about prescribed use. So the majority seem to know the recommendations and knowingly deviate. However, in contrast, Zhang et al. (2015) reported both under- and overuse for Chinese farmers and say it may be related to lack of knowledge. Bagheri et al. (2019) studied the drivers of farmers’ intentions to use pesticides. Including an assessment of knowledge and motivations of use could improve understanding and estimations of pesticide use especially when extrapolating survey data. With insights in farmers’ motivations, the extrapolation of the data to other regions can be more precise or can be applied in intervention scenarios to estimate effects of social- or financial interventions.