Yield gap variation in rice cultivation in Indonesia

2.1 Data collection and compilation

Through Google Scholar and Web of Science, we conducted a comprehensive literature search of peer-reviewed studies published during 2012–2022. The keywords used in the search include paddy soils, rice yield, and the combination thereof, using the Indonesian language. Studies were included in the analysis if (i) they were relevant to paddy fields in Indonesia, (ii) the technology package and rice yield were reported, and (iii) details of experiment location, design, and condition were provided.

In addition to the published article, we collated technical reports containing the rice yield of previous field trials. The Assessment Institutes of Agricultural Technology conducted these trials under the supervision of the Indonesian Ministry of Agriculture to assess RCTPs’s performance. Most authors of these reports were principal investigators or research team members in the respective trials.

In the original published articles and reports, there was no information on the coordinate location, only the village name and type of paddy field were available. Therefore, we reconstructed the coordinate site according to the village name, masked with a village boundary map and paddy field map. The web application assisted the coordinate assignment, namely, the coordinate locater (www.coordinatelocator.com).

A total of 295 individual observations from 130 studies were included in our analysis; of the 295 studies, 66% were conducted in irrigated paddy fields, 23% were conducted in tidal paddy fields, and 10% were done in rainfed paddy fields. The trial was performed in almost all geo-regions (Figure 1), in the humid tropic (Sumatra, Java, Kalimantan, Sulawesi, Maluku-Papua), and Semi-arid (Nusa Tenggara). In humid tropical, the annual rainfall is 2,000 mm, and in semi-arid, it is less than 1,000 mm. Kalimantan and Sumatra get more rain (∼2,900 mm) than Java as a whole (∼2,600 mm), while the eastern part of the Archipelago receives much less: Sulawesi (∼2,300 mm) and the Little Sunda Islands (<1,500 mm) [26]. The differences in precipitation determine the water availability in the fields. The trial location extended from south to north latitude and from 1 to 612 m above sea level (asl) in 23 provinces. The sites were dominated by low land with an altitude of 200 m asl or less and in areas with latitudes >1° South.

Figure 1

The distribution of field trial sites in 23 provinces in Indonesia, 2012–2022.

In each trial, an RCTP was compared to the farmer’s practice. A total of eight RCTPs were evaluated from 2012 to 2022. Of 295 observations, 54% used PTT, and 38% used Jarwo Super. One technology package can be the modification of the previous one to adapt to local conditions. The applied RCTP is briefly described in Supplementary Table S1.

2.2 Statistical analysis

The dataset contained location up to village level, type of paddy fields (irrigated, rainfed, tidal), RCTP, rice varieties, PY, and FY. The PY was the rice yield that can be achieved by applying specific RCTP with no constraints, and the FY was the rice yield reached using farmers’ technology. If the FY was not recorded, FY was estimated from the statistical average farmers’ yield in a given district during the trials. As yield units differed, the yield unit was standardized to a metric ton of unmilled rice per hectare (Mg ha^–1). Next the YG was calculated by reducing PY to FY. Instead of using an absolute figure, the YG was expressed as a percentage of FY because when discussing food security, observed world grain production and likely increases were directly linked to FY, not PY [14]. Supplementary Table S2 provides a brief definition of these yields in comparison with other terminologies.

We did exploratory data analysis and variance analysis. The mean comparison was made using the Fisher-LSD test at alpha 0.05. Statistical analysis was assisted by the R programming language [27].

3.1 Effect of paddy field type, altitude, and latitude on rice yield and YG

Table 1 presents the effect of the type of paddy field on the variation in rice yield (FY and PY) and YG. FY was significantly different statistically at the level of 0.05. The irrigated paddy field shows the highest rice yield (5.24 Mg ha⁻¹), followed by the rainfed (4.04 Mg ha⁻¹) and tidal (3.42 Mg ha⁻¹). The PY also showed a similar pattern with FY, where PY was significantly different statistically at level 0.05. The irrigated paddy field produced the highest yield (6.82 Mg ha⁻¹), followed by the rainfed (6.18 Mg ha⁻¹) and tidal (4.96 Mg ha⁻¹). Surprisingly, the rainfed gave significantly the highest YG (62%), followed by the tidal (52%) and the irrigated paddy field (32%).

Table 1 also shows the minimum and maximum yield. At the farmer level, the irrigated shares the highest minimum and maximum yield by 3.16 and 7.84 Mg ha⁻¹, respectively, followed by rainfed by 2.00 and 5.85 Mg ha⁻¹, respectively, and tidal by 1.00 and 5.72 Mg ha⁻¹, respectively. The PY also gathers a similar pattern: the irrigated shares the highest yield of minimum and maximum yield by 3.73 and 9.90 Mg ha⁻¹, respectively, followed by 4.04 and 8.40 Mg ha⁻¹ for rainfed and 2.20 and 7.60 Mg ha⁻¹ for tidal paddy fields. Statistically, the irrigated and the tidal paddy fields are significantly different from the rainfed at a level of 0.05.

The altitude position did not statistically affect the FY and YG but significantly affected the PY. The highest PY was found on paddy fields with an altitude of 400 m asl or more, which reaches 7.23 Mg ha⁻¹ (Table 1). The lowest rice yield in the lowland paddy fields was 2.2 Mg ha⁻¹, 1.0 Mg ha⁻¹, and 1% for PY, FY, and YG, respectively. Likewise, the highest yield was 9.7 Mg ha⁻¹, 7.84 Mg ha⁻¹, and 310% for PY, FY, and YG, respectively.

The latitude statistically affects the PY, FY, and YG (Table 1). The PY and FY were the highest in the Southern Latitudes, reaching 6.54 and 5.01 Mg ha⁻¹, respectively. The lowest YG was also in the Southern Latitude region at 35%.

3.2 Geo-region

Table 2 shows the rice yield (FY and PY) and the YG in the six geo-regions of Indonesia, namely, Java, Nusa Tenggara, Sumatra, Sulawesi, Kalimantan, and Maluku-Papua. The paddy fields of Java contributed to the highest FY of 5.66 Mg ha⁻¹, which was significantly different from the other geo-regions, followed by Sulawesi, Sumatra, and Nusa Tenggara geo regions with 5.05, 4.82, and 4.76 Mg ha⁻¹, respectively. These three geo-regions differed significantly from Maluku-Papua and Kalimantan, which shared the lowest FY of 3.54 and 3.38 Mg ha⁻¹, respectively. Meanwhile, the average PY generated a different pattern from the FY. Java geo-region generated the highest PY, insignificantly different from Kalimantan by 7.27 and 5.47 Mg ha⁻¹, respectively. Furthermore, the geo-region of Kalimantan was insignificantly different statistically from the other geo-regions.

Surprisingly, the geo-regions of Kalimantan and Maluku-Papua resulted in the highest YG, 74 and 49%, respectively, significantly different statistically from the other geo-regions. Meanwhile, the lowest YG was gathered by Sulawesi and Java geo-regions (27 and 31%, respectively), which was insignificantly different statistically from the other geo-regions.

Table 3 presents the YG variation based on the provinces in Indonesia. It can be observed that the North Kalimantan province had the highest YG compared to all provinces in Indonesia, and it was significantly different statistically at level 0.05. The high YG in North Kalimantan (141%) was followed by East Kalimantan (105%) and West Kalimantan (66%). Otherwise, the lowest YG was found in West Sumatra (19%) and South Sulawesi (20%), followed by East Java, South Sumatra, and Central Kalimantan.

In Java, DKI Jakarta province had the highest YG (51%), while East Java had the lowest YG (20%). In this region, significantly few technical issues limit high yield. The limitations are mainly in the social and economic aspects, such as the low price of rice, which results in less profit than would be made from producing other commodities.

The highest YG in Sumatra was found in Riau province (51%), and the lowest occurred in West Sumatra province (19%). In Kalimantan, the highest YG occurred in North Kalimantan province (141%), and the lowest was found in Central Kalimantan province (20%). The highest YG in Sulawesi was in North Sulawesi province (47%), and the lowest was in South Sulawesi province (19%). The YG variation in several provinces is influenced by, among others, the planted varieties, fertilization, climate, topography, and soil resource properties [28]. From a technical cultivation aspect, early planting and dosage of N fertilizer significantly reduce YG [29].

Meanwhile, in the semi-arid Nusa Tenggara region, East Nusa Tenggara province (52%) had a greater YG than West Nusa Tenggara province (33%). The variation in the YG in Papua was higher than in North Maluku province for the Maluku-Papua region, by 53 and 33%, respectively. In rainfed paddy fields, changes in water regime and climate significantly affect rice yields because sufficient water supports rice plants’ growth and development [12].

Due to abundant water and nutrient supply from fertilizer applications, Indonesian farmers usually rely on the rainy season to grow rice. However, water becomes scarce during the dry season, primarily during drought. Among abiotic stress, drought is one of the essential factors affecting one-third of the total rice area in Asia and causing significant economic losses to poor rice producers [30]. Farmers have no liquid assets to buy necessary equipment in this scenario immediately, thus lowering rice productivity [31]. To avoid this, farmers must think ahead about the best timing to plant and harvest their crops and efficient use of water when the drought arrives. To reduce the risk of increasing YGs in some regions, selecting appropriate climate change mitigation measures and policies is necessary to close YG [32].

3.3 Introduced RCTPs

As many as eight RCTPs were tested from 2012 to 2022. From them, we only selected The Site-Specific Nutrient Management (PHSL), Jarwo Super, PTT, BPRL, and Intensive tidal rice cultivation (RAISA) for further analysis because they were assessed in three locations or more (Table 4). Other packages, i.e., Sonor, SRI-based organic paddy, and KATAM were evaluated only in 1–2 sites. The PY was significantly different among technological packages (P < 0.001).

The PY was influenced by technology packages (P < 0.01), type of paddy fields (P < 0.001), and interaction between type and technology packages (P < 0.01) (Table 4). PHSL, BPRL, or Intensive tidal rice cultivation (RAISA) was the technology used in specific paddy fields. Still, Jarwo Super and Integrated Rice Crop Management (PTT) were used in all paddy fields. The PHSL and Jarwo Super showed higher results in irrigated paddy fields, followed by the PTT and BPRL.

In the rainfed paddy fields, the Jarwo Super performs better, resulting in higher yield and significantly different from PTT. The application of fertilization technology supports the Jarwo Super technology, such as (1) nutrient management by applying balanced fertilization adjusted to the development and needs of plants, (2) utilizing agricultural waste, especially rice straw, as a source of organic matter, and (3) utilization of various superior microbes to decompose organic waste, seed treatment, and plant growth promoters. Therefore, the growth of rice plants is optimized, which impacts increasing rice productivity [33].

In Jarwo Super, the rice straw being returned to paddy fields can increase organic matter in the soil. Saikia et al. [34] concluded that using farmyard manure and crop waste (rice straw) in paddy fields long-term could help the stability of organic carbon in the soil. Thus, applying organic matter can increase microbial and enzyme activities that can accelerate the transformation of mineralized nitrogen in the soil. Organic-dominated nutrient management can restore soil health and offset 20% of the recommended NPK fertilization. Conversely, in the tidal paddy fields, the Jarwo Super shows a lower yield and significant difference from RAISA and PTT. RAISA is specific for tidal paddy fields.

3.4 Contribution of rice variety

Our study indicated that the rice variety significantly contributed to the variation in FY (P < 0.001) and PY (<0.001). This study identified 80 rice varieties that were tested, and Table 5 shows only varieties showing high yield, more than 5.0 Mg ha⁻¹, and were tested only at 5 locations or more. These varieties cover Inpari 46, Ciherang, Inpari 32, Inpari 19, Inpari 33, and Inpari 20. The Ciherang variety has been introduced for a long time but can still provide the same results as new superior varieties such as Inpari 34 and Inpari 42. These varieties also reflect the preferred variety by farmers. The Indonesian government has released more than 160 superior rice varieties for lowlands, highlands, and tidal paddy fields. These varieties are high in yield, resistant to pests and diseases, and tolerant to abiotic stress [35,36].

When this variety used a technology package, the rice yield increased by about 1.4–2.0 Mg ha⁻¹. Changing the rice variety increased rice yield, and if the same variety used the introduced technology package, the yield of this variety increased, too. Using superior rice varieties supported by site-specific technology packages produced productivity close to PY and reduced YG.

Each variety showed a different yield. Inpari 46 had the highest yield compared to other varieties, namely, 6.78 Mg ha⁻¹ at the farmer level and 8.68 Mg ha⁻¹ when planted using an improved technology package. Inpari 46 has an upright shape with a plant height of approximately 101.5 cm. Inpari 32 and 33 rice varieties can produce more than 6 Mg ha⁻¹ in tidal paddy fields [37]. The texture of the cooked form of this rice is fluffier. In addition, this type of Inpari rice is somewhat resistant to the brown planthopper biotype. Each variety has a different response for each given technology package.

Selecting suitable varieties to overcome environmental stress is very important, but soil improvement efforts such as applying fertilizers and soil conditioners must also be carried out. Besides increasing soil fertility, fertilization can increase plant resistance to pests and diseases. Rice varieties resistant to certain pests and diseases are as follows: Ciherang is resistant to blight and Inpari 33 is resistant to brown planthoppers. Integrated pest control, arrangement of planting schedules, and spacing of plants also need to be done to reduce pest and disease attacks. However, using one continuously planted variety is not recommended because some pests and diseases can have adaptability and form biotypes/strains. The superior varieties are proven to be able to increase production, but their development is still limited [38].

Some of the obstacles in the development of superior varieties are the availability of seeds in the right amount and at the right time, the level of grain loss, price competition by traders, the need to increase the level of farmer adoption, the taste of rice, and the shape of the grain. Growing and strengthening breeder institutions at the regional level are necessary to meet the demand for seeds of superior varieties. In addition, socialization and assistance regarding the advantages of certain varieties need to be carried out to increase the adoption rate of farmers.

4.1 YG variation

Our study indicated that the YG varied widely from 1 to 310% (Table 1). This variation was due to differences in the environment (latitude, geo-region, province), the type of paddy field (irrigated, rainfed, tidal), planted rice variety, and RCTP. The yield discrepancy also varied with the type of paddy field: rainfed was 64%, tidal was 54%, and irrigated was 32% of FY (Table 1). The YG of the provinces considered rice production centers, namely Java and Sulawesi, was smaller than other provinces (Table 3).

Our study also found exciting results in Nusa Tenggara. The region has a semi-arid climate, which is drier and hotter than other regions. But, the YG was similar to Sumatra statistically, at about 40% (Table 2). Farmer’s yield of Nusa Tenggara is also similar to those of Sumatra statistically, although Sumatra has a humid climate and abundant water. Interestingly, the PY was the second highest after Java. High temperatures and drier climates provide an abundance of solar radiation for photosynthesis. When water is available, the rate of photosynthesis increases, producing more yields.

Using crop modeling technique for estimating PY, Yuan et al. [12] reported that in Indonesia, the YG of irrigated paddy fields was, on average, 37% of PY (59% of FY), and rainfed was 49% of PY (104% of FY). In irrigated paddy fields of Yogyakarta, using crop model simulations, the YG was 57% of PY (75% of FY), primarily due to input efficiency and technology application (sowing date) [39]. In the rainfed paddy fields of Central Java, using crop modeling, the YG was 0–28% of PY (0–39% of FY) caused by water limitation, and due to N limitation was from 35 to 63% of PY (54–170% of FY) due to N limitation [40]. Our study confirmed that the YG in the rainfed was higher than in the irrigated.

The YG occurred because the available production technology was not applied to farmers’ fields [41]. It can happen, among others, due to: (a) characteristics of farmers such as lack of knowledge and skills, risk aversion; (b) agricultural characteristics, e.g., poor soil, rugged terrain, inaccessibility; and (c) technological incompatibility with farmers’ circumstances (e.g., labor-intensive, high investment costs, poor access to inputs). The existence of limiting factors in soil nutrient status and insufficient water availability in rainfed paddy fields has the potential to widen the YG. These constraints must be analyzed at the field level to design management strategies to improve and maintain yield stability and reduce YG [40].

Minimizing the YG and increasing profit and product quality are becoming increasingly difficult to achieve using a single-technology-centric approach. Although the rainfed shows a larger YG, it is also necessary to pay attention to the end of the YG in irrigated land. This is because it can cause a more significant impact on annual yield due to higher cropping intensity [42]. Combining and simultaneously applying the best component technologies is crucial for maximizing overall benefits to farmers [43].

Crop management options such as setting planting dates, using short-lived varieties, and water management can help plants minimize the adverse effects of drought and provide adequate N fertilizer to increase yields of rainfed lowland rice in Indonesia. The addition of N fertilizer should be given as needed. A systems approach using the concept of production ecology can be applied in yield constraint analysis to identify management strategies to increase yields and yield stability in lowland or other tidal and rainfed paddy fields [40].

The strategies to close YG include implementing agricultural technology innovation in the form of RCTP. These RCTPs are proven to close YG even for the next 20 years, especially in rainfed paddy fields [44]. Moreover, the YG in the regions that have implemented or used various old programs for a long time is lower compared to areas that have just been cleared and used, including sub-optimal areas. In irrigated paddy fields, the growing season can also affect the YG. In Java, the YG in the dry season is around 6–17% higher than in the wet season.

Our findings also support Yuan et al. [12], who explained that interventions to improve plant management practices, especially nutrient and water management. Ran et al. [28] indicated that more precise fertilization recommendations should be formulated based on comprehensive factors (soil, climate, terrain, and variety). We confirmed in Table 4 that applying the technology (in Supplementary Table S1) increased yields from 1.3 to 2 Mg ha⁻¹. The research results of Sulaeman et al. [45] found that the use of varieties such as Inpara 8, Suppadi 89, Inpara 3, and Inpara 2 gave higher yields than 5 Mg ha⁻¹, namely, 6.31, 5.54, 5.46, and 5.36 Mg ha⁻¹. The rice yield from tidal paddy fields depends on varieties, cultivation technology packages, and soil and water characteristics. A similar thing was also expressed by Erythrina et al. [46] that the average yield with farmer practices gave low yields, namely, 3.4 Mg ha⁻¹, compared to the recommendation package, which was 5.3 Mg ha⁻¹. By adopting better practices, YGs can be reduced. The average YG using the recommended package, namely, crop management and nutrition, is 37%. The recommendation package is the use of superior varieties such as Inpari, which are more suitable and tolerant to pests and diseases, compared to the use of IR64 and Ciherang varieties, which are more susceptible to pests and diseases, as well as managing plant nutrition [47]. In reality, farmers in nutrient management practices are not based on recommendations.

4.2 Re-orientation of subsidies and other interventions to boost rice production

The spatial variation of YG may be used as the base for allocating agro-input subsidies and other agricultural interventions to boost rice production. Paddy field with large YG (>40% of FY) indicates significant opportunities for research and the need for improvement in crop management, infrastructure, enabling institutions, and market [5]. Increasing rice yield in a high YG is affordable because options for intervention are available. Considering this large YG, subsidies and other agricultural interventions should go directly to rainfed and tidal paddy fields. Farming subsidies and technical assistance should be directed to East Kalimantan, North Kalimantan, South Kalimantan, West Kalimantan, East Nusa Tenggara, and North Sumatra provinces (Table 3).

These high YG were mainly due to the lack of capital and disability to apply recommended technologies. Regarding pest and disease control, farmers were unaware of practicing Integrated Pest Management, the right time, and insecticides to be used for certain kinds of pests and diseases. In this high YG area, closing YG can be done by building and improving irrigation systems, including groundwater exploitation, drainage systems, and amelioration soil, prioritizing rice yield increases in areas with low average yields, as well as the application of site-specific technologies such as precision farming, the use of superior adaptive varieties [48]. By using specific superior varieties, managing plant nutrition, and adding irrigation sources, rice yield increased to 1.9 Mg ha⁻¹ [46]. Agricultural intensification must occur in locations with large YGs and low resource use efficiency.

In the paddy field with low YG (<30% of FY), closing YG for increasing FY stagnates. The yield has closed to the maximum for a given technology package in this area. Nevertheless, yield growth can be maintained by crop breeding and more favorable inputs and output practices, motivating farmers to invest in technologies required to bridge the theoretical YG [5]. Meanwhile, in locations with high profitability, small YGs, and high resource use efficiency, increasing sustainability would be more appropriate based on continuous intensification (or even extensification) [49].

Specifically, the YG in tidal paddy fields can be closed through water management, soil improvement, fertilization, and adaptive varieties. The management water systems include regulation and control of water in the quarter and tertiary canals, which can facilitate the washing of toxic materials and prevent salt water from entering the plot of land [50]. Tidal paddy soils have low soil pH, high Fe, Al, and Mn solubility, low availability of nutrients, especially P and K, and low base saturation [51]. Hence, efforts are being made to improve soil quality and fertility through soil amendments, such as agricultural lime and rock phosphate [52,53].

4.3 Optimizing carbon (C) sequestration through technology package implementation

C sequestration captures atmospheric C through biomass production and storage in the soil [54]. Depending on different crop and land management practices, soil can be either a source of atmospheric CO₂ or a C sink. Technical interventions, such as applying organic and inorganic fertilizers or adding biomass and using certain varieties to narrow the YG, positively impact retaining carbon in the soil. It will be a challenge to optimally utilize plants as one of the inputs to increase yields.

Implementing a technology package improves rice growth performance and hence increases carbon sequestration. In addition, genjah (short-lived) rice variety can decrease the potential methane emission from paddy soil. Therefore, both may contribute to climate change adaptation and mitigation strategies. Nevertheless, more detailed studies need to be conducted to quantify this contribution. Furthermore, soil management is a significant driving factor in the dynamic form of soil organic carbon and greenhouse gas (GHG) emissions [55,56].

Most of the technology package in this study uses organic fertilizers either by returning the rice straw or adding manures to the field. Replacing rice straw and organic fertilizers in the soil-plant system can potentially increase soil organic carbon storage and carbon sequestration [57,58,59]. The continuous addition of organic fertilizers improves soil carbon sequestration and carbon pools, which are essential for tropical soils to sustain soil quality and agricultural productivity with intensive farming systems [58].

Nutrient management practices such as PHSL, Jarwo Super, PTT, and BPRL may strongly influence various carbon fractions, influencing carbon sequestration. The precision application of N fertilizer enhances crop yield, N use efficiency, and soil organic carbon storage and mitigates GHG emissions [60]. Package technologies such as PHSL, Jarwo Super, PTT, and BPRL apply optimum and balanced nutrients to maximize crop yields ( Supplementary Table S1), leading to relatively more C inputs from above- and below-ground plant biomass to the soil. In addition, the long-term practice of PHSL and the added organic matter contributed higher microbial biomass, labile, water-soluble, and particulate fractions of soil carbon reported elsewhere [61,62,63]. However, these carbon sequestrations need more detailed studies in Indonesia’s long-term practices of various technology packages.

On a global level, several initiatives and agreements have been established regarding good soil management, especially with food security and climate change issues. The Paris Agreement, Koronivia join work on agriculture (KJWA) and nearly 4p1000 initiatives encourage world governments to prioritize soil management as an adaptation strategy and co-benefit of climate change mitigation. Good plant growth will absorb a lot of carbon from the atmosphere and reduce the concentration of GHGs.

4.4 Future direction

Our study suggested that implementing a proper technology package at the right site increases FY to 1.6 Mg ha⁻¹. RCTPs are among sustainable intensification practices; therefore, closing YG to increase FY could be achieved by implementing RCTP. The performance of each technology varies with the location and type of paddy field. Some of the most promising RCTP to increase farmers’ yield to 1.6 Mg ha⁻¹ are Jarwo Super (Irrigated paddy fields), Jarwo Super and PTT (Rainfed paddy fields), and RAISA (Tidal paddy fields).

Hidayat et al. [64] concluded that PTT was the most effective alternative to increase rice productivity compared to other technologies in North Maluku Province. The paddy field has unique water, soil, and agroclimatic characteristics, which, alone or in combination, controls the effective implementation of all technology packages to increase rice yield. Such land resource characteristics determine the carrying capacity and the maximum limit of paddy fields to produce rice. Thus, in the future, we can estimate the maximum yield generated from the best technologies, either land management technology, cultivation technology, or genomic technology for variety assembly.

Each paddy soil has a critical limit and maximum rice production capacity. The maximum yield is greatly influenced by the genetic potency of rice varieties and the number and type of agro-input (fertilizer, soil ameliorant, and pesticide), soil tillage, and crop protection. Understanding each paddy plot’s critical point and capacity is necessary to avoid overestimating production. For example, Talio Hulu soils can only produce a maximum of 6 Mg ha⁻¹ with Inpara 2 variety and the most optimum technology. It means that the expectation of rice production of more than 6 Mg ha⁻¹ in this swampland is challenging to achieve unless new swamp rice varieties are introduced with higher productivity. Released varieties must have a yield of 8 Mg ha⁻¹ or more. If less than that, the rice variety cannot be released to the public by law.

The productivity achieved in various types of paddy fields with the application of site-specific cultivation technology packages (PTT, RAISA, BPRL, and Jarwo Super) still shows high YG compared to the possible results that can be achieved. Developing an approach to introducing a new technology innovation model is necessary, focusing on implementing critical technologies as leverage points in each technology package. Technology that can reduce seed use is adapted to the typology of paddy fields in Indonesia, such as using an Amator drum seeder in tidal lowlands [65,66].

To overcome the efficiency and technology gap, several technology packages, such as PHSL, Jarwo Super, PTT, BPRL, and RAISA, have been applied to increase rice yields and reduce YGs, and the results are pretty successful. Using a technology package for continuous intensification, which optimizes the interaction between genotype, environment, and management, holds promise. Once popularized, they would potentially increase resource use efficiency, producing more rice at a lower environmental cost. Our findings help develop climate-resistant cultivars and agronomic management practices optimal for site-specific environments to increase PYs, close YGs, and accelerate sustainable agriculture intensification. In addition, this study demonstrates advanced approaches’ capabilities to define the scope for yield improvement, identify limiting factors, and optimize interactions between genotype, environment, and management. These ideas and strategies can be applied to other cropping systems and areas.

4.5 Limitation of this study

In this study, we exercised the historical yield data for one planting season from different years (10 years, 2012–2022) and locations (130 field trial sites, countrywide). In addition, we mixed between yield from the dry season and the wet season planting. The previous research highlighted that rice yield in the dry season was higher than in the wet season [39], indicating that the growing season may contribute to yield variation. Also, the 10-year difference may record climate and rainfall change spatially. Hence, the multi-year analyses based on different growing seasons will be challenging to asses yield stability and spatiotemporal variation.

The precise coordinate location of field trial sites is crucial for relating yield and any aspect of the experiment to other auxiliary geospatial information. Yet, such information is rarely available. Thus, in this study, we reconstructed the coordinate position of each field trial from the village and regency name, which were available in the publication. We masked the coordinate location with the paddy field map and the type of paddy field. Therefore, the coordinate position of trial sites was rough and potentially misleading for studies that need a more precise location. But, for regional and global research or visualization, as in this study, the current coordinate location was helpful. Thus, the coordinate site should be used wisely for future studies.

This study also did not consider climatic temporal and spatial variation. The rainfall intensity, temperature, solar radiation, and elevation potentially contribute to spatial and temporal rice yield variations. Relating temporal and spatial climatic variation to rice yield while considering the contribution of technology packaged to increase yield may be future exciting research topics to understand climate change behavior in tropical regions such as Indonesia.