The research, led by Riska Tiasmalomo and an interdisciplinary team of Indonesian agricultural scientists, takes aim at a persistent blind spot in agricultural sustainability research. Most previous studies of cut flower farming have examined only one or two dimensions at a time, such as profitability or plant physiology, using conventional statistical tools like regression analysis and analysis of variance. The team instead embraced system dynamics, a modeling technique pioneered by MIT engineer Jay Forrester, which is designed to capture the feedback loops, delays, and nonlinear interactions that make real-world systems so difficult to predict. Their framework integrates four dimensions simultaneously: economy, social conditions, technology and innovation, and ecology, a combination the authors abbreviate as ESTIE and anchor in the United Nations Sustainable Development Goals.

At the core of the modeling effort are 54 variables drawn from official statistics compiled by Batu City’s Central Statistics Agency and its Agriculture and Food Security Agency. The researchers first built a causal loop diagram, a conceptual map in which arrows show whether one variable strengthens or weakens another. They then translated this map into a stock and flow diagram using the Vensim PLE software, converting qualitative relationships into quantitative equations that could be simulated over a 15-year horizon from 2020 to 2035. Each of the four submodels centers on a single crucial variable: profit for the economy, the total number of farmers for the social dimension, agricultural tools and machinery for technology, and total rose production for ecology.

The economic submodel treats profit as a stock that fills with income and drains with expenditure. Income depends on production volume multiplied by price, while expenditure combines fixed costs, including land rent, taxes, and equipment depreciation, with variable costs such as seeds, fertilizer, pesticides, and labor. The numbers are striking: pesticide costs alone run between roughly 87 and 94 million Indonesian rupiah per hectare per year, and seed costs between 43 and 53 million. Because these inputs consume such a large share of revenue, even modest changes in input prices or selling prices ripple powerfully through the system, a dynamic the simulations make vividly clear.

The social submodel captures a quieter but equally threatening trend: farmers leaving the profession. Batu City counts 9,707 farmers overall, but only 279 of them grow cut roses, organized into 22 farmer groups and a single association. The model tracks an occupational transition of roughly 2,000 to 2,200 people per year, driven by the harsh arithmetic of farming that no longer meets daily needs. When farmers sell their land and move into manufacturing or service jobs, agricultural land is converted to other uses, which in turn pushes remaining farmers to seek better-paying work elsewhere. The researchers describe this as a self-reinforcing spiral in which land conversion and farmer attrition feed each other, eroding the industry’s foundations year after year.

Ecological pressures compound the problem. Agricultural land in Batu City is being converted to non-agricultural uses at a rate of 5 to 10 percent annually, driven by population growth, housing demand, and a booming tourism industry that needs hotels, supermarkets, and infrastructure. Meanwhile, rose production is buffeted by climate anomalies, modeled through temperature, humidity, rainfall, and rainy days, and by pest and disease attacks that the model allows to range from 15 to 50 percent of output. The technology submodel reveals its own paradox: of nearly 48,000 agricultural tools and machines in the city, more than 2,100 sit idle, often because subsidies arrived without repair facilities, spare parts, or training, leaving equipment damaged and unused.

To ensure the model faithfully reflected reality, the team subjected it to behavioral validation, comparing simulated outputs against historical data using two statistical tests: a comparison of averages, which must deviate by no more than 5 percent, and a comparison of amplitude variation, which must stay within 30 percent. All four submodels passed comfortably. The economic model’s error was just 3 percent, the social model 1.74 percent, the technology model 2.35 percent, and the ecological model 0.81 percent, indicating that the simulations track the actual behavior of Batu’s rose system with remarkable precision.

With a validated model in hand, the researchers ran three parameter scenarios, pessimistic, moderate, and optimistic, by adjusting key variables and projecting outcomes to 2035. The optimistic scenario emerged as the clear winner across every dimension. Economically, it requires raising the selling price of cut roses by 795 rupiah per stalk, from 788, while cutting fertilizer costs by about 33.5 million rupiah per year and pesticide costs by roughly 93 million rupiah per year. Under these conditions, profit in 2026 would climb to 378.3 million rupiah per hectare, well above the actual figure of 357.8 million, and continue rising toward more than 1.1 billion rupiah by 2035.

The social and technological prescriptions are equally concrete. Cutting the number of farmers abandoning the profession to 1,734 people per year, expanding the ranks of rose farmers to 360, and doubling extension service visits from 12 to 24 times per year would lift the projected farming population to 55,735 people by 2026, exceeding the baseline of 53,218. On the technology front, increasing actively used machinery to 24,079 units, reducing idle equipment, and raising the share of farmer groups adopting agricultural technology from 48.96 percent to 51.50 percent would push the total tool count to 205,090 units by 2026. Ecologically, the optimistic scenario calls for expanding the rose harvest area by 545 hectares, planting an additional 54.5 million rose trees, and slashing pest and disease attacks from 40 percent to 8.6 percent, which would raise 2026 production to 124.2 million stalks.

The study’s authors are careful to frame these figures not as predictions but as strategic guidance, a rational benchmark against which farmers, extension workers, and policymakers can evaluate decisions. Their central message is that sustainability cannot be achieved by fixing one variable in isolation. Lowering pesticide costs, for instance, must be paired with better pest management so that production does not fall; retaining farmers requires both economic viability and stronger social support through training and counseling. By modeling the four dimensions together, the dynamic systems approach reveals trade-offs and synergies that single-factor analyses miss entirely. For a city whose identity and economy are intertwined with a single flower, the stakes could hardly be higher. The simulations suggest that with coordinated action on prices, inputs, technology adoption, land protection, and farmer welfare, Batu’s rose industry can not only recover from its 2020 shock but grow steadily for decades, offering a replicable template for sustainable horticulture across Indonesia and beyond.

Subject of Research: Dynamic systems modeling of sustainable cut rose farming in Batu City, Indonesia

Article Title: Model and scenario of sustainable cut rose farming development in Batu City: A dynamic system approach for Indonesia's cut rose center

Article References: Tiasmalomo, R., Salam, M., Iswoyo, H., Jamil, M. H., Tenriawaru, A. N., Dermawan, R., Kamarulzaman, N. H., Akhsan, Heliawaty, Fudjaja, L., Rahmadanih, Ridwan, M., Ali, H. N. B., & Syam, S. H. (2026). Model and scenario of sustainable cut rose farming development in Batu City: A dynamic system approach for Indonesia's cut rose center. Environmental and Sustainability Indicators, 32, Article 101533. https://doi.org/10.1016/j.indic.2026.101533

Image Credits: AI Generated

DOI: 10.1016/j.indic.2026.101533

Keywords: cut roses, system dynamics, sustainable agriculture, Batu City, Indonesia, horticulture, simulation modeling, farmer welfare, agricultural technology, land conversion, pest management, scenario analysis