The black soldier fly (BSF), Hermetia illucens L. (Diptera: Stratiomyidae), is an insect decomposer used to valorize food and agricultural waste. Until now, large investments have been directed towards developing industrial-scale insect farming systems. Although there is potential for BSF farming to be adopted also by small and medium farmers, agricultural cooperatives, food processing facilities, and university campuses, an affordable and easy-to-adopt BSF farming system has not been developed yet. Here, we designed and tested a modular steady-state BSF rearing method for the rapid conversion of cafeteria waste into residual decomposed matter known as frass, a product made of insect excrements, residual undigested organic matter, insect exuviae, mineral nutrients, and a rich microbiota. We used the prepupae output (PP kg/m²-day) to track the efficiency of the system and compared it between a minimal and a multimetric monitoring approach. This comparison demonstrated how monitoring moisture, temperature, and pH in the system can produce a continuous, reliable yield of prepupae and frass, which increased by 47 % and 42 %, respectively, after the start of monitoring these metrics. The enhanced recovery of both outputs constitutes a potential source of revenue to offset operating costs. By enabling adult reproduction and egg laying to occur in the same space as larval fattening, and removing the need for rearing BSF in batches, our system can save on space and labor costs relative to established rearing options. This will increase BSF farming accessibility for smaller-scale users with limited resources and diverse waste management needs.

We evaluated prepupal yield over a one-year period using two different approaches: a minimal monitoring protocol that involved addition of inputs based only on visual assessments at each loading cycle (months 1-6), and a multimetric monitoring protocol in which input quantities were varied depending on sensor measurements of the moisture, pH, and temperature of BR1 contents, as well as a substrate consistency rating scale (months 7-12). Each method involved a loading cycle conducted approximately every four days, during which inputs were added and prepupal yield was assessed by weighing. Prepupae collected during each loading cycle were transferred to a plastic bin below the BR1 to allow adult flies to emerge (Fig. 2 in associated manuscript).

Regardless of the monitoring method, each loading cycle involved the same basic steps. First, food waste was added to the BR1 in the form of the contents of the oldest open bucket containing young larvae. Second, bulking agents were added to provide substrate aeration. Third, water was added to keep the substrate moist and to replace water that had evaporated between loading cycles. Fourth, the contents of the BR2s were added back to the BR1. Once all these components were placed in the BR1, the contents were mixed using a shovel and spading fork. Following mixing, excess BR1 substrate containing the combined materials and larvae was transferred back to BR2s leaving the depth of waste in the BR1 at approximately 14 cm. Additional maintenance tasks for the system (e.g., cleaning) can be found in Appendix 1.

The loading cycle for the minimal monitoring approach involved visual assessments of BR1 substrate and bulking agent needs without sensor-based monitoring or a substrate rating scale. This phase lasted 161 days, from 8 September 2023 through 16 February 2024. Water and bulking agents were added based on a visual evaluation of the consistency of the substrate housed in the BR1 unit. The amount of water varied between 9.5 and 37.8 liters per loading cycle.

The loading cycles for the multimetric monitoring approach involved decision making based on measurements of substrate temperature, gravimetric moisture, pH, and a friability index inspired by the soil texture-by-feel analysis method to allow operators to quickly assess moisture levels of the substrate with a gloved hand to decide how much water to add to the BR1 without having to measure moisture gravimetrically (Appendix 2). This phase lasted 160 days, from 8 March 2024 through 15 August 2024.

We collected data on substrate characteristics at each loading cycle and used these data to estimate the quantity of bulking agent and water to add to the system. We measured pH using a portable pH probe by slurring substrate from the BR1 combined with five parts of water. We used a compost thermometer to record ambient and waste temperatures (7 cm into the substrate, 3 BR1 points averaged). Water additions and temperature measurements are shown in Figures S1 and S2, respectively. At periodic intervals (15 out of the 42 loading cycles included in the multimetric monitoring phase), we quantified gravimetric moisture content of BR1 contents by collecting 50 g of material across the substrate depth before and after the loading cycle and drying in 8 oz metal tins at 105 °C for 24 hr. The average gravimetric moisture of the substrate at different depths, pre- and post-loading, are shown in Figure S3.

We collected and removed frass from the system by dedicating 1 to 2 BR2 as storage units instead of remixing them into the BR1. At each loading cycle, we added 1 to 2 kg of BR1 material to one BR2 at a time until full, before filling the other one. Once full, these units were left to dry to about 10% moisture content for about four weeks before the contents were weighed and stored in a 121-L vented plastic trash can. We measured pH of BSF-F as described above for BR1 contents and performed elemental analysis at the UCR Environmental Science Research Laboratory.