Assessment of post-harvest losses and carbon footprint in intensive lowland rice production in Myanmar

Globally about 500 million tons of milled rice are produced annually, 90% in Asia. For smallholder farmers, reducing post-harvest losses and improving grain quality are critical. Traditional post-harvest operations in Myanmar are characterized by low mechanization, manual harvesting, in-field stacking for weeks before threshing, sun drying, and storage in granary bags—practices that can cause substantial quantitative and qualitative losses (e.g., shattering, pests, mould, discoloration, broken grains). While mechanization can increase costs and potentially GHGE through machine production and fuel use, it may reduce losses relative to traditional practices. Prior studies quantified post-harvest efficiencies or LCA metrics in various countries, but no synthesis compared alternative post-harvest management practices across the full chain to identify options that minimize losses, costs, and environmental footprints. This study aims to assess, in Myanmar, (i) energy and costs of grain production in traditional systems, (ii) GHGE incurred by mechanized post-harvest options, and (iii) comparative GHGE, energy, and costs between traditional and improved scenarios.

The paper situates its work within evidence that post-harvest losses can reach 20–30% based largely on qualitative assessments (FAO, IRRI). Mechanization levels vary widely across Southeast Asia, with Myanmar lagging behind Thailand and Vietnam. Sun drying and poor storage are documented to cause quality and quantity losses; hermetic storage and mechanical dryers mitigate these. Previous research quantified economic and technical efficiency of rice post-harvest systems in several countries and applied LCA to rice production in multiple contexts (e.g., Philippines, Japan, Iran, Italy, Canada). Studies have examined specific processes such as drying, straw collection, and farm mechanization. However, a comparative assessment across alternative post-harvest chains integrating energy, cost-benefit, and GHGE, particularly in Myanmar smallholder systems, was lacking—defining the research gap addressed here.

Study site and seasons: Tar Pat Village, Maubin, Ayeyarwaddy delta, Myanmar (16.617°N, 95.680°E), in WS2014 and DS2015–2016 using variety Sin Thukha (135 days). Functional unit (FU): 1 ha of rice production. System boundary: cultivation through milling. LCA: Attributional LCA following ISO 14044; conversion factors from Ecoinvent v3.3; Cumulative Energy Demand (1.09) and IPCC GWP100 used for energy and GHGE factors. Experimental design: Completely randomized design with five replications per scenario. WS2014: three scenarios—FP1w (manual cutting, stack 1 week, farmer thresher, sun drying, granary bags), FP4w (stack 4 weeks), and IPR (manual cutting, immediate threshing ≤12 h with improved thresher TC-800, flatbed dryer, hermetic Super bags). DS2015 & DS2016: FP (manual cutting, immediate threshing with farmer thresher, sun drying, granary bags) and IPRc (combine harvester Kubota DC-70G, flatbed dryer, hermetic bags). Plot sizes: WS2014—15 plots of 270 m²; DS—10 plots of 390 m². Equipment: Farmer thresher (axial, 15 HP two-wheel tractor power), imported thresher TC-800 (7.5 HP), combine DC-70G (70 HP), flatbed dryer (4 t batch, locally made), hermetic bags (50 kg). Measurements: Harvesting losses quantified via five 1 m² quadrats per scenario capturing shattering and losses during cutting/handling, in-field stacking (when applicable), threshing (separation, cleaning, under-machine), and combine harvesting. Losses during drying and storage handling and losses to birds/rodents were not directly quantified (rodent losses referenced). Post-harvest quality: milling recovery (MR) and head rice recovery (HRR) measured using standardized lab milling of subsamples; HRR and MR calculated as percent of paddy sample. Discoloration quantified by separating discolored kernels from three random 25 g milled samples (>0.5% colored surface considered discolored). Energy accounting: Net energy value (NEV) = EVoutputs − EVinputs; Net energy ratio (NER) = EVoutputs/EVinputs. Outputs include energy of whole rice (15.2 MJ kg⁻¹), broken/discolored/bran (9.6 MJ kg⁻¹), husk (8.7 MJ kg⁻¹), and straw (6.5 or 3.5–6.5 MJ kg⁻¹; straw assumed ~50% of grain yield collected for mushroom production). Inputs include energy for cultivation (WS 12 GJ ha⁻¹; DS 16 GJ ha⁻¹), machine production (depreciated over 5 years), fuel/power, labor (converted via MET for 55 kg body weight), and transportation (tractor trailer, 15 km). GHGE accounting: Total GHGE per ha = (GHG_cultivation + GHG_harvest + GHG_postharvest)/product recovery ratio. Cultivation GHGE: WS ~2000 and DS ~1200 kg CO₂-eq ha⁻¹ (site-specific prior study). Off-field emissions translated per ha via yield. Product recovery ratio = (1 − Harvesting loss) × HRR × (1 − Discoloration). Electricity emission factor from Ecoinvent (ROW grid). Cost-benefit: Net income value (NIV) = income from whole, broken, discolored rice, bran, husk, straw minus cultivation plus harvest/post-harvest costs; Net income ratio (NIR) = NIV/total input cost. Prices: whole rice US$400 t⁻¹, discolored rice and bran US$140 t⁻¹, husk US$10 t⁻¹ equivalent (US$0.01 kg⁻¹). Cultivation cost US$650 ha⁻¹. Component costs (equipment price, lifespan, electricity US$0.05 kWh⁻¹, labor US$0.46 h⁻¹) per 2018 assessment. Sensitivity: linear effects of harvesting loss on NIV and NIR analyzed for improved practices. Statistics: ANOVA (single-factor, two-factor with replication) and F-Test for variances (Excel). LCA modeling via SIMAPRO v8.5 with Ecoinvent v3.3 and IPCC factors.

- Mechanization and income: Improved practices increased net income by 30–50% versus traditional practices. In DS, NIV was 962.7 US$ ha⁻¹ (IPRc) vs 747.5 US$ ha⁻¹ (FP); in WS NIV was 81.7 US$ ha⁻¹ (IPR) vs −50.8 US$ ha⁻¹ (FP). - Harvesting loss: WS (no combine) had high harvesting losses (16.0% IPR vs 28.2% FP1w vs 23.6% FP4w). DS with combine IPRc had 0.9–1.7% vs FP 4.0–9.3%. - Grain quality: In WS2014, discoloration was 3.8% (IPR) vs 6.8–7.9% (FP1w/FP4w). HRR was 47.2% (IPR) vs 27.3% (FP1w) and 17.2% (FP4w). In DS, IPRc HRR 54.8–64.0% vs FP 48.1–57.5%. - Outputs (examples): Whole grains per ha: WS IPR 1149 kg vs FP1w 567 kg vs FP4w 380 kg; DS IPRc 2628–3437 kg vs FP 2130–2997 kg. Broken grains reduced markedly under IPRc in DS (e.g., 98–311 kg ha⁻¹) relative to FP (141–504 kg ha⁻¹). - Energy: Despite higher mechanization energy, total life-cycle inputs did not significantly increase for IPR vs FP. NEV in WS: 17.56 vs 13.62 GJ ha⁻¹ (IPR vs FP); DS: 45.58–47.52 GJ ha⁻¹ (no significant difference). NER ranged 1.95–3.23, not significantly different between IPR and FP or seasons. - GHGE: Total GHGE in WS were 5.30 Mg CO₂-eq ha⁻¹ (IPR) and 5.73 Mg CO₂-eq ha⁻¹ (FP). In DS they were 2.04 Mg (IPRc) and 2.93 Mg (FP). Cultivation contributed 70–90% of GHGE, harvest 5–15%, post-harvest 5–10%. Accounting for product recovery, higher losses raised GHGE per ha by about 30–50%. - Sensitivity: Reducing harvesting loss from 14% to 2% increased NIV by ~US$78→182 (WS) and US$744→950 (DS); NIR increased accordingly (WS 0.10→0.25; DS 0.91→1.14). - Specific improvements: Combine harvesting reduced harvesting loss by 3–7 percentage points; flatbed drying plus hermetic storage reduced discoloration by 3–4% and increased HRR by 20–30% (by weight).

Mechanized post-harvest practices (combine harvesting, flatbed drying, hermetic storage) substantially decreased harvest and post-harvest losses, improving grain quality and profitability without increasing life-cycle energy use or GHGE. Loss reductions increased the rice product recovery ratio, which in turn lowered GHGE per hectare despite similar or slightly higher operational energy inputs. In DS, higher yields and lower losses amplified NEV and NIV compared to WS, consistent with farmer perceptions of low WS profitability. The analysis shows that the economic advantage of IPR arises primarily from avoided losses rather than from reduced direct costs, and that GHGE are dominated by cultivation with loss-driven increases in emissions intensity. Proper implementation (timely harvest, skilled combine operation, prompt drying, correct storage, calibrated milling) is critical to realize these benefits. The findings also provide evidence to guide service providers and policymakers in developing mechanization services and policies that yield economic gains without increasing environmental footprints.

Improved post-harvest practices increased net income by 30–50% compared with farmer practices and significantly reduced harvest and post-harvest losses while not increasing total life-cycle energy use or GHGE. Combine harvesting reduced harvesting loss by 3–7 percentage points; flatbed drying and hermetic storage reduced discoloration by 3–4% and increased HRR by 20–30%. With higher yields, lower losses, and better product recovery in DS, NEV and NIV were 30–50% higher than in WS, and GHGE were 40–60% lower. Energy required to manufacture and operate machinery is offset by reduced production losses. Future research should extend similar attributional LCA assessments across diverse rice-producing regions and systems to refine recommendations and support scalable, context-specific mechanization strategies.

- Cultivation-stage data were sourced from secondary studies, not measured in this experiment. - Experimental plots were relatively small (approximately 300–500 m²), which may influence harvesting loss estimates; sensitivity analysis was used to assess impacts on cost-benefits. - Energy associated with straw was accounted for, but in-field GHGE were taken from prior studies rather than measured contemporaneously. - LCA conversion factors for materials and electricity were from Ecoinvent (rest-of-world averages), which may not perfectly reflect Myanmar-specific supply chains or grid mixes. - Losses during drying and storage handling, and losses to birds/rodents, were not fully quantified within the experiment (rodent losses referenced separately).