Introduction

Challenges associated with rapid population growth and limited arable land pose considerable challenge for global food security. It is necessary to address this issue by reducing constraints on food production, reducing yield fluctuations, reducing food waste, and changing diets1. However, higher yield and productivity in agricultural production require high agricultural inputs (including fertilisers, pesticides, irrigation, and agricultural machinery)2,3. Synthetic fertilisers play a crucial role, especially nitrogen (N) fertilisers4,5. Thus, farmers on the North China Plain often overuse N fertilisers to improve wheat yields6, and farmers in North America typically overuse N fertilisers to avoid the risk of reduced production7. Nonetheless, excessive use of N fertilisers and unreasonable agricultural management methods do not improve grain yield or agronomic efficiency; instead, they cause environmental problems, such as increased greenhouse gas (GHG) emissions, soil acidification, groundwater eutrophication, and soil salinisation8,9. According to previous studies, the average N application rate in China’s agricultural ecosystems is 305 kg N ha−1 yr−1, which is much higher than the global average N application rate (74 kg N ha−1 yr−1). In addition, N utilisation efficiencies in China, North America, and globally are 0.25, 0.65, and 0.42, respectively10. Therefore, the fertiliser utilisation efficiency in China still needs improvement.

Optimising crop nutrient management in China, improving fertiliser utilisation efficiency, and reducing environmental risks are all actions that are urgently required. Sustainable agricultural development can be achieved through practices such as improving fertiliser management, optimising crops management strategies, implementing integrated pest management, and adopting precise irrigation and water resource management4, which have gained wide attention from the public. To meet the growing demands of the global population, crop production will need to double from 2005 to 205011. Research has shown that owing to a combination of specific ___location factors, 79, 56, and 52% of rice, wheat, and maize, respectively, produced in China have experienced yield stagnation, and some crops in certain countries or regions have experienced yield reductions2. Therefore, it is necessary to determine whether and to what extent environmental costs may be reduced in the years to come while ensuring food security, which may be achieved long-term stable productivity of land resources12. Many studies have shown that grain yields can be increased and environmental impacts can be reduced by improving nutrient management and optimising agricultural practices10,13. Compared to traditional agriculture, crop production can be maintained or increased using organic and protective agricultural practices; nonetheless, the temporal stability of organic agriculture requires improvement14. Feng et al.15 indicated that intelligent climate agricultural practices (such as intensive planting, deep tillage, manure improvement, and N fertiliser optimisation) could be used to improve maize yields while reducing the carbon footprint.

There is enormous potential to increase grain yield and nitrogen use efficiency (NUE) by reducing N loss and, thus, environmental impacts16,17. Previous studies have shown that in Europe18, the Mediterranean region19, and the North China Plain20, optimising the amount of N fertiliser used for maize and wheat production (improving agronomic efficiency) reduces environmental impacts and improves ecological efficiency. Yao et al.21 compared the GHG emissions of maize from major maize-producing countries such as Canada, the United States, Mexico, Brazil, Greece, and Germany, and found significant differences in GHG emissions among these countries and regions. The climate and soil conditions in the subtropical plateau region of China are suitable for planting various food crops (rice, wheat, and maize), making this region an important grain producing area in China. Rice, wheat, and maize are among the most important food crops for humans worldwide. Yunnan is located in the low latitude subtropical plateau region and should thrive on the excellent conditions; however, its grain yield lags behind that of other regions, and fertilisation efficiency is imbalanced owing to improper nutrient management. Measures for optimising resource investments and reducing negative environmental impacts and emissions from crop production in subtropical plateau areas are currently unclear.

With increased public awareness of environmental damage and the associated government attention, it is necessary to quantify the grain yield levels and environmental costs at a regional scale. This will assist researchers and policymakers in developing emission reduction measures to achieve the sustainable development of crop production systems. Life cycle assessment (LCA) is often used to quantify and evaluate resource inputs and environmental impacts of crop production systems22. The potential for mitigating the environmental impacts of different crops in the subtropical Chinese plateau region remains unknown. Therefore, in this study, we conducted an LCA to analyse the environmental impacts of crop production in Yunnan Province from 2002 to 2021. The aims of this study were to (1) quantify the yield, nutrient input, and N fertiliser utilisation of the major grain crops in Yunnan Province; (2) assess GHG emissions and the AP and EP of crop production in the region; and (3) propose potential emission reduction strategies to reduce environmental costs in China’s subtropical plateau areas.

Materials and methods

Study area

This study was conducted in Yunnan Province (21°08′N to 29°15′Ν, 97°31′E to 106°11′E), including Kunming, Qujing, Yuxi, Zhaotong, Chuxiong, Honghe, Wenshan, Pu’er, Xishuangbanna, Dali, Baoshan, Dehong, Lijiang, Nujiang, Lincang, and Diqing city, and based on data from 2002 to 202123 (Fig. 1). Yunnan is located within a typical subtropical plateau monsoon climate zone and the average annual precipitation and temperature are 1064 mm and 17.4 °C, respectively24 (Fig. 2). The terrain is extremely complex, high in the northwest and low in the southeast, and the climate and main soil characteristics of the region vary significantly (Table 1). Approximately 5.40 million hectares of land is cultivated in Yunnan Province, and the total area of sown crops reached 4.19 million hectares in 2022. Rice, wheat, and maize are main grain crops in this region, and the total planting area and yield of the three crops in descending order is as follows: rice > maize > wheat.

Fig. 1
figure 1

The crop planting region and sub-regions in Yunnan. The green, yellow, and red lines represent the average production (104 Mg) of rice, wheat, and maize in Yunnan province from 2002 to 2021, respectively. Data for crops production were obtained from the NBSC25 from 2002 to 2021. The map was constructed using ArcMap 10.8 software (Esri, Redlands, CA) (available at: http://www.esri.com/ ).

Full size image
Fig. 2
figure 2

Changes in annual temperature (maximum temperature, minimum temperature and average temperature) and annual precipitation in Yunnan Province from 2002 to 2021. The blue bars represent total precipitation in each year during 2002–2021; the red, yellow and blue dots represent the maximum temperature, average temperature and minimum temperature of each year, respectively.

Full size image
Table 1 The climatic characteristics of in Yunnan in recent 20 years, and their major soil characteristics.
Full size table

Annual precipitation, annual temperature and annual illumination hours were obtained from the China Meteorological Data Service Center (CMDC)24. The major soil textures and soil pH were according to FAO and World Reference Base for Soil Resources (2018). Values are means ± SD. The soil N content was obtained from the China Soil Science Database (vdb3.soil.csdb.cn).

Data collection

Datasets of sowing area, yield, nutrient inputs (N, P, and K), pesticide inputs, and plastic films usage during the production of rice, wheat, and maize from 2002 to 2021 were obtained from the China Statistical Yearbook25 and a Compilation of Cost and Income of Agricultural Products in China26. Meteorological data were obtained from the China Meteorological Data Service Center24. All statistical yearbooks, including national, provincial, and prefecture levels, are available in the China Statistical Yearbook Database (CAJ) and the yearbook series set for 2002–2021. The data used in this study were regional- or local-level data from national summary data, ensuring the accuracy and representativeness of the research results. We divided the 20 years of data into four stages, 2002–2006, 2007–2011, 2012–2016, and 2017–2021, to reduce the impacts of inter-annual climate variation.

Statistical analysis

All primary data were processed using Microsoft Excel 2019. IBM SPSS Statistics (IBM Corp. Armonk, NY, USA) was used to conduct linear regression. Graphical plots were constructed using OriginPro 2021 and ArcGIS (version 10.8). A one-way analysis of variance was used to test the interactive and main impacts of subregions or time pe-riods on grain yield, partial factor productivity of nitrogen (PFP-N, in kilograms of grain per kilogram of N applied4), fertiliser rate, GHG emissions, AP, EP, and CF via SAS software (version 9.3; SAS Institute). Where treatment impacts were significant, means were compared using the least significant difference test at P < 0.05.

Functional units and system boundary

To reflect the food security and overall environmental sustainability of the system accurately, we evaluated the impact based on two functional units: per hectare and per mega grams (Mg) of crops. The system boundary used in this study was defined as “the cradle-to-gate of the field”27, which included all GHG emissions, AP, and EP from sowing to harvesting in relation to two stages (Fig. 3): the agricultural materials stage (MS), which includes the production and transportation of fertilisers, pesticides and other agricultural inputs (seed, diesel fuels, and plastic film) (Table S1), and the farming stage (FS), which includes fertilisers and pesticides application, diesel fuels consumption by machinery28. The emission factors of various agricultural inputs in the production and transportation processes are listed in Tables S2 and S3. The irrigation water was not included because of a lack of reliable data that would not affect the main results of this study. Due to the varying methods of crops straw treatment and utilization, such as burning, returning to the field, feed, or industrial raw material, which differ by region, climate, and management practices, obtaining consistent and reliable data becomes challenging20. Therefore, this study did not consider the environmental impact of crop straw.

Fig. 3
figure 3

System boundary of life cycle assessment for calculating environmental impacts in rice, wheat, and maize production systems.

Full size image

Impact assessment

The most widely used index for representing NUE is PFP-N (kg kg−1), reflecting the yield produced per unit of applied N29, which is calculated as follows:

$${\text{PFP}} - {\text{N}} = \frac{{\text{Y}}}{{\text{N}}}$$
(1)

where Y is the crop yield (Mg ha−1) and N is the rate of N application (kg N ha−1).

For the environmental assessment, we calculated the N balance (Nsurplus, kg N ha−1) from the difference between the amount of N input and the amount of N uptake, where it was absorbed by the aboveground biomass of crops30:

$${\text{N}}_{{{\text{surplus}}}} = {\text{N}}_{{{\text{input}}}} - {\text{N}}_{{{\text{uptake}}}}$$
(2)

where Ninput (kg N ha−1) represents the sum of the N fertiliser applied, and Nuptake (kg N ha−1) represents the total N uptake of the crop. Considering the rice, wheat, and maize yields of Yunnan over the past 20 years, the amounts of N used to produce a million grams (Mg) of rice, wheat, and maize were 21.0, 27.1, and 19.8 kg, respectively31,32.

Estimation of greenhouse gas emissions, acidification potential and eutrophication potential

Global warming, water eutrophication, and soil acidification are three significant environmental issues caused by excessive fertilisation, which directly impact ecosystems. Additionally, these environmental problems have received considerable attention in current agricultural production systems, particularly regarding grain production safety33. According to the system boundaries defined in Section “Functional units and system boundary”, the environmental impacts of crop production comprised emissions from fertilisers (N, P, and K), pesticides, and plastic films in the MS and FS. The calculation formulas for the environ-mental impacts (GHG emissions, AP, and EP) of the rice, wheat, and maize production systems were constructed as follows:

GHG emissions

The GHG emissions were calculated using the IPCC method and the following equation34:

$${\text{GHG}}_{{\text{T}}} = {\text{GHG}}_{{{\text{MS}}}} + {\text{GHG}}_{{{\text{FS}}}}$$
(3)

where GHGT (kg CO2-eq Mg−1) is the total GHG emissions per hectare per year from crop production. Furthermore, GHGMS and GHGFS (kg CO2-eq Mg−1) are the GHG emissions per hectare per year of crop planting area from MS and FS, respectively, which are obtained as follows:

$${\text{GHG}}_{{{\text{MS}}}} = \sum ({\text{AI}}_{{\text{i}}} \times {\text{EF}}_{{\text{i}}} )$$
(4)

where AIi (kg ha−1) is the application rate for the ith agricultural input comprising fertilisers (N, P, and K), pesticides, seed, diesel fuels, and plastic film; EFi (kg CO2-eq kg−1) is the GHG emissions factor associated with the ith agricultural input35. The CH4 emissions arising from the organic matter decomposition in anaerobic conditions (due to the flooding) are usually the main contributor, so CH4 emissions are only calculated for the rice season36. The GHGFS was calculated as follows:

$${\text{GHG}}_{{{\text{FS}}}} = {\text{Total N}}_{2} {\text{O}} \times 44/28 \times 265 + {\text{Total CH}}_{4} \times 27.9$$
(5)
$${\text{Total N}}_{2} {\text{O}} = {\text{N}}_{2} {\text{O}}_{{{\text{direct}}}} + 1.0{\text{\% }} \times {\text{NH}}_{3} + 2.5{\text{\% }} \times {\text{NO}}_{3}^{ - } {\text{ leaching}}$$
(6)

where the Total N2O is the N2O emission at the FS divided into direct and indirect N2O emissions from crop production. Total CH4 is the CH4 emission at the FS from rice paddies, which is based on the estimates by Huang et al.36,37 of CH4 emissions during rice production in various provinces of China. Furthermore, N2Odirect is the N2O emission from N fertiliser application10; 44/28 is the conversion coefficient of N2O-N to N2O; 265 and 27.9 are the global warming potential coefficients for N2O and CH4 from a period of 100 years36,38, respectively; 1.0% and 2.5% are the indirect N2O emission factors associated with NH3 volatilisation and nitrate leaching, respectively39.

Acidification potential (AP)

The AP of the entire life cycle of crop production systems mainly comprises emis-sions during the production and transportation of agricultural inputs, as well as NH3 emissions directly related to the application of N fertilisers in agricultural systems. The AP were calculated according to parameters used by Liang et al.40 and the following equation:

$${\text{AP}} = \mathop \sum \limits_{{{\text{n}} = 1}}^{{\text{m}}} {\text{MS}}_{{{\text{SO}}_{2} }} + 1.88 \times {\text{NH}}_{3} \times 17/14$$
(7)

where AP (kg SO2-eq Mg−1) is the acidification impact produced by the crops input by the production unit; MSSO2 is the SO2 emission from the production and transportation of various inputs during the MS, including fertilisers (N, P, and K), pesticides, seed, diesel fuels, and plastic film; NH3 is the volatilisation of NH3 during the FS; 17/14 is the conversion coefficient of N to NH3; and 1.88 is the conversion coefficient of SO2 acidification gas emission with 1 kg NH340.

Eutrophication potential (EP)

The substances that cause EP include NH3, NOx, NO3-, NH4-N, COD, and Ptot. The EP were calculated according to parameters used by Liang et al.40 and the following equation:

$${\text{EP}} = \mathop \sum \limits_{{{\text{n}} = 1}}^{{\text{m}}} {\text{MS}}_{{{\text{PO}}_{4} }} + 0.33 \times {\text{NH}}_{3} \times 17/14 + 0.42 \times {\text{NO}}_{3}^{ - } {\text{ leaching}} + 0.2{\text{\% }} \times {\text{P}}_{{{\text{input}}}}$$
(8)

where EP (kg PO4-eq Mg−1) is the eutrophication impact produced by the crops input by the production unit; MSPO4 is the PO4 emission from the production and transportation of various inputs during the MS, including fertilisers (N, P, and K), pesticides, seed, diesel fuels, and plastic film; NH3 and NO3- leaching represent the previously calculated NH3 volatilisation and NO3− leaching losses, respectively; 17/14 is the conversion coefficient of N to NH3; 0.33 and 0.42 are the conversion coefficients of PO4 with 1 kg NH3 and 1 kg NO3− leaching, respectively; and Pinput is the total amount of P fertiliser input40.

Emission-mitigation scenarios for greenhouse gas emissions, acidification potential and eutrophication potential

Enormous energy inputs from fertilisers occur throughout the life cycle of rice, wheat, and maize. Therefore, it is necessary to increase the efficiency of crop production farms by reducing the fertiliser rate while increasing yield. Specific efficient strategies and optimal scenarios were developed to explore the potential for mitigating the environmental impacts of crop production in the representative subtropical plateau region in China during 2002–2021, and to predict for the next 20 years to 2041. Scenario 1 (S1): ‘business as usual’, we projected the GHG emissions, AP, and EP for 2041 using the same average yield and N fertiliser input during 2002–2021 (Table S4). All other inputs for crop production were the same as those used in 2017–2021 (Table S1). Scenario 2 (S2): fertilisation-optimised scenario used recommended fertilisation rates, Wu et al.41 developed these rates based on the regional climatic, cultivation, and soil conditions of Chinese three major cereal crops. In S2, the optimal fertiliser application rate was based on the nutrient levels required for the target crop yield by Wu et al.41 (Table S4). All other inputs for crop production were the same as those for S1. Scenario 3 (S3): fertilisation-optimised scenario used recommended innovative N fertiliser, previous studies have indicated that controlled-release urea can improve crop yield and NUE while reducing N2O emissions, N leaching, and NH3 volatilisation42. Fertilisation rate based on a national meta-analysis conducted by Zhang et al.43, the application of controlled-release urea can increase the yield of rice (5.85%), wheat (8.93%), and maize (9.97%) compared to normal urea (Table S4), and it significantly reduced the accumulation of NH3 (36.61%) and N2O (28.18%). All other inputs for crop production were the same as those for S2. Scenario 4 (S4): integrated optimisation of fertilisation and crop management, it was previously employed, achieving 75% of the regional yield potential with optimised N application (Table S4) and crop management practices (optimising varieties, densities, and sowing dates, and realising the efficient use of nutrients through rhizosphere nutrient regulation)44.

Results

Crop yield, nutrient input, and PFP-N

Over the 20-year period, the average yields of rice, wheat, and maize were 6.03, 2.05, and 4.43 Mg ha−1, respectively, and the overall yield showed an upward trend over the four sub-periods (2002–2021) (Fig. 4a–c). The average PFP-N of rice and maize increased by 17.2% and 40%, respectively, whereas that of wheat decreased by 16.0% from 2002–2006 to 2017–2021 (Fig. 4d–f). The average PFP-N of rice, wheat, and maize were 29.0, 15.9, and 16.2 kg kg−1 during 2002–2021, respectively.

Fig. 4
figure 4

Grain yield (a–c) and PFP-N (d–f) of crops (rice, wheat, and maize) production systems in Yunnan Province from 2002 to 2021 (20 samples) were compared across five-year intervals. Different lowercase letters represent statistically significant differences (LSD test, P < 0.05).

Full size image

There were no significant differences in the total fertilisation rates for rice, wheat, and maize from 2007 to 2021, with the highest total fertilisation rates of crops in the four periods during 2002–2021 being 350, 190, and 398 kg ha−1, respectively (Fig. 5a–c). There were no significant differences in the application rates of N fertiliser for maize or P fertiliser for wheat between 2002 and 2021. And relative to rice and maize, wheat had a higher total fertilisation rate. However, the N fertiliser rate for maize was higher than those for rice and wheat, whereas the P and K fertiliser rates for rice were higher than those for wheat and maize (Fig. 5d–l).

Fig. 5
figure 5

Fertiliser input for crop (rice, wheat, and maize) production systems from 2002 to 2021 in Yunnan Province (20 samples). The rice (a) total fertiliser rate, wheat (b) total fertiliser rate, and maize (c) total fertiliser rates were compared across the five-year intervals. Different nutrient inputs (d–f, nitrogen (N) fertiliser rate; (g–i) phosphorus (P2O2) fertiliser rate; (j–l) potassium (K2O) fertiliser rate) for crops (rice, wheat, and maize) production systems were compared across five-year intervals. Different lowercase letters represent statistically significant differences (LSD test, P < 0.05).

Full size image

Greenhouse gas emissions, Acidification potential and Eutrophication potential

There were wide variations in the environmental impacts of crop production in Yunnan Province. During 2002–2021, the average GHG emissions from rice, wheat, and maize production in Yunnan were 898, 789, and 806 kg CO2-eq Mg−1, respectively, the average AP of the three crop types were 10.3, 18.2, and 18.5 kg SO2-eq Mg−1, respectively, and the average EP were 1.79, 3.14, and 3.20 kg PO4-eq Mg−1, respectively (Fig. 6). Over the 20-year period, GHG emissions, AP, and EP from the rice and maize production systems showed a decreasing trend, whereas the related impacts of the wheat production system showed an increasing trend. The application of agricultural inputs in the FS was the largest contributor to environmental emissions in crops production systems. For the same food crops, the trends of GHG emissions, AP and EP were the same in the four stages during 2002–2021(Fig. 6a–i).

Fig. 6
figure 6

Environmental impacts (global warming potential, a–c; acidification potential, d–f; eutrophication potential, g–i) of crops (rice, wheat, and maize) production systems in Yunnan Province from 2002 to 2021 (20 samples) were compared across five-year intervals. Different lowercase letters represent statistically significant differences (LSD test, P < 0.05). Vertical bars represent ± S.E. of the mean.

Full size image

Correlations between environmental impacts and influencing factors

To analyse the factors influencing the environmental impacts, a correlation analy-sis was conducted between the environmental factors and N surplus of the three crops with GHG emissions, AP, and EP. N surplus affected the growth of rice, wheat, and maize (Fig. 7). Notably, the environmental costs of rice production were greatly af-fected by annual temperature (Fig. 8). However, with changes in annual precipitation and temperature, the environmental impacts of wheat and maize showed no significant changes (Figs. S1–S3). The environmental impacts of rice, wheat, and maize production systems were positively correlated with N surplus, whereas the environmental impact of rice production was negatively correlated with temperature.

Fig. 7
figure 7

Correlations between N surplus and GHG emissions (a, b, c), the acidification potential (d, e, f), and the eutrophication potential (g, h, i). The black solid line represents the linear correlation; green, yellow, and orange dots represent environmental impacts of rice, wheat, and maize production in Yunnan Province from 2002 to 2021, respectively. P < 0.01 indicates the significance of the regression.

Full size image
Fig. 8
figure 8

Correlations between environmental impacts of rice production and annual temperature. The black solid line represents the linear correlation; green dots represent the average annual GHG emissions (a), AP (b), and EP (c) of rice production in Yunnan Province from 2002 to 2021. P < 0.05 indicates the significance of the regression.

Full size image

Environmental impacts and mitigation potential of crop production

The environmental risks of the grain crop production system on the plateau were significant; nevertheless, there was great potential to reduce such risks. If the yield and N fertiliser input of the crops maintained the average trend of the 20-year period stud-ied, the environmental impacts of the three crops under the designed S1 scenario would show an increasing trend. Based on the regional climatic, cultivation, and soil conditions (S2), the recommended fertilisation rates for crops are determined, leading to significant increases in the yields of rice, wheat, and maize. At the same time, GHG emissions, AP, and EP are expected to decrease by 22.8–37.4%, 45.4–57.7%, and 48.7–62.3%, respectively. In S3, the application of controlled-release urea can improve NUE and reduce N losses, with an expected yield increase of 5.85%-9.97%. Meanwhile, GHG emissions, AP, and EP are expected to decrease by 25.7–40.6%, 48.3–61.4%, and 51.5–65.8%, respectively. Compared to S1, the comprehensive scenario (S4) combines advanced crop and nutrient management practices, and rhizosphere nutrients regulation, which would achieve a 75% regional yield potential. GHG emissions, AP, and EP from the production of rice, wheat, and maize are expected to decrease by 43.0–59.5%, 51.5–64.5%, and 57.4–71.5%, respectively (Fig. 9).

Fig. 9
figure 9

Projected environmental impacts of rice, wheat, and maize production for 2041 in Yunnan, China. Red dots represent GHG emissions (a–c), AP (d–f), and EP (g–i) of crops production in Yunnan Province from 2002 to 2021. Red, orange, blue, and green dashed line represent environmental impacts of crops production in S1, S2, S3, and S4, respectively.

Full size image

Discussion

Differences in yield, nutrient input, and PFP-N

Owing to the adoption of various modern technologies in the agricultural sector, grain yield has significantly increased in recent years45. Currently, the most effective way of improving grain yield is to increase the amount of fertiliser applied, particularly N fertilisers46. However, this approach has not produced good results in Yunnan Province. In the present study, the average grain crop yields in Yunnan Province over the 20-year period studied were relatively low at 6.03, 2.05, and 4.3 Mg ha−1 for rice, wheat, and maize (Fig. 4a–c), respectively. These values were lower than the national average yields (rice:7.0 Mg ha−1, wheat:5.7 Mg ha−1, and maize:7.6 Mg ha−1) by 13.9, 64.0, and 41.7%, respectively4,47,48, and considerably lower than yields of rice in southern China, wheat in Germany, and maize in the United States49. We found that in the past, the amount of N fertiliser input for rice and maize was higher than the national average amount(Fig. 5)7, while the average PFP-N of grain crop production (28.95, 15.93, and 16.18 kg kg−1, respectively) were lower than the 35.7, 51.7, and 62.4% of the national average PFP-N levels of crop production (45, 33, and 43 kg kg−1, respectively) (Fig. 4d–f)4,50. Ni et al.51 collected all studies from the 1980s to 2020 on N fertilisation associated with plastic-shed vegetable production at 40 field sites in China, and they found that there was a significant and negative relationship between PFPN and fertiliser N rate for vegetables under plastic-shed production. Excessive N fertiliser application can lead to a decrease in crop N absorption and utilisation efficiency and cause associated problems, such as reduced crop production efficiency, increased resource consumption, and environmental risks52.

To meet grain demand in the subtropical plateau region, improvements in NUE are necessary to optimise crop production efficiency in the region in the coming decades. Jiang et al.53 assessed the spatial heterogeneity of the contributions of N input components on crop growth in the provincial scale in China, finding that the soil fertility N contribution on crop growth is relatively small in southwestern provinces, suggesting that reducing fertiliser use may be more suitable for sustainable agriculture. In addition, regarding the alkaline N and total N content in the tillage layer in different regions of China, the southwestern region is at a relatively high level (Table 1 and S5). Therefore, soil N is crucial for improving grain yields in the study area. It is important to note that crop NUE is related to changes in soil type, climate, crop rotation strategy, and variety of crop employed. When optimised fertiliser management techniques and agricultural management methods are applied to these good soil conditions and a reasonable planting density is ensured, then yield disparities can be reduced and fertiliser utilisation efficiency can be improved54,55. Zhou et al.56 reported that optimising fertilisation can improve NUE while increasing the yield by 13–15%. To improve the yield and resource utilisation efficiency in subtropical plateau areas, it is necessary to comprehensively consider the climate, species and varieties grown, cultivation measures, and management practices in the soil crop system.

Characteristics of environmental impacts from crop production

It is important to study the environmental footprint of crop production to achieve sustainable development and efficient resource management, address climate change, and guide policymaking. In this study, the GHG emissions, AP, and EP of rice and maize production showed decreasing trends (Fig. 6), whereas the environmental footprint of wheat production showed an increasing trend, which was opposite to the trend in the PFP-N of the crops. The environmental impacts of wheat and maize production in the study region were much higher than those in other countries and regions worldwide (Europe and America) (Table 2). Improvements in the NUE of crop production would reduce the environmental impacts, and one study found that optimising N fertiliser management reduced environmental risks and improved the ecological efficiency of crop production19. This is interesting to consider since wheat and maize yields in subtropical plateau regions are much lower than those in Europe (59.0%) and the United States (63.2%)33,57, resulting in lower environmental impacts in these countries and regions than in subtropical plateau regions (Table 2). A decrease in yield indicates a decrease in nutrient intake by crops, with N content in the soil exceeding the crop demand, leading to an increase in the soil N surplus. Higher nutrient inputs exacerbate that N surplus, thereby increasing environmental risks. In this study, N surplus was positively correlated with the environmental impacts on crops (Fig. 7). The N fertiliser input for maize production in this study area was higher than that in the United States58, and the remaining reactive N losses (N2O, NH3 volatilisation, and N leaching) were released into the air, water, and soil59,60. Nitrogen leaching can lead to an increase in the nitrate concentration in water, resulting in an increase in GHG, soil acidification, and water eutrophication61,62. The GHG emissions of the main grain crops in this study area ranged from 683 to 1134 kg CO2-eq Mg−1, while the weighted average GHG emissions of vegetables produced in China are 116 kg CO2-eq Mg−134; this difference could be because grain crops are more dependent on agricultural machinery than vegetable crops63. However, these differences are not fixed and the GHG emissions of different types of vegetables and food crops in different regions and under different cultivation management practices may vary significantly. In the study area, GHG emissions from the rice system were higher than those from the wheat and maize systems, which is consistent with the findings of Chen et al4. This may also lead to some differences due to specific local conditions and other management practices, such as soil types, agricultural management, variety characteristics, N supply, and environmental conditions64. In subtropical plateau regions, the combined effects of high temperatures and rainfall may pose significant environmental risks to crop production; for instance, there is a correlation between the environmental impact of rice production and annual temperature (Fig. 8).

Table 2 Comparison of the scope of environmental impacts between Yunnan Province and other regions.
Full size table

In agricultural systems, N fertilisation is the most critical factor for improving the environmental impact of crop production65. Reducing the N application rate and optimising N fertiliser management are necessary steps to reduce N surplus and envi-ronmental load. Notably, controlling N application does not necessarily reduce crop yield or crop production efficiency or increase environmental risks, as previously de-scribed33,66.

Mitigating the environmental impacts of crop production in subtropical plateau regions

Reducing the environmental impacts of growing major crops in subtropical plat-eau areas is a major challenge to sustainable agricultural development. However, the results of this study indicate that emissions from crop production in subtropical plat-eau areas can be effectively reduced by optimising the amount of N input and increas-ing crop yield based on the current production level and the amount of nutrients input in the study area. The results of the present study show that the N surplus in crop production is linearly correlated with environmental impacts (Fig. 7). Accordingly, N fertiliser application can be optimised to improve PFP and reduce environ-mental impacts4. Through long-term experimental observations, Hu et al.67 found that optimising the N rate for high-yield crops provided high NUE and low environmental risks. Certain measures reduce NH3 volatilisation, such as using controlled-release fertilisers, reducing the amount of N input, using urease inhibitors, and using deep N fertilisation68,69,70. In addition, pesticides and irrigation can have significant environmental impacts71. In crop production, it is necessary to control pesticide use and optimise irrigation to reduce environmental risks. Irrigation methods, including intermittent irrigation, intermittent irrigation drainage in mid-season frequent waterlogging, and no waterlogging drainage in mid-season intermittent irrigation, can be used to reduce CH4 emissions during the rice-growing season72. The results of scenario S4 in this study showed that selecting appropriate varieties, sowing dates, planting densities, and advanced nutrient management (ISSM) would offer the great-est potential to reduce emissions from crop production in high subtropical regions4 (Fig. 9). In a further consideration, the soil properties, microbial abundance and activity, and crop growth in farmland can be affected by differences in variety, water and nutrient management, fertiliser type, and tillage intensification among different cropping systems, leading to differences in environmental emissions73. Therefore, com-prehensive agricultural management practices should be adopted for crop production in subtropical plateau areas to promote sustainable development of agriculture and reduce environmental impacts and pollution. In such ways, farmers may achieve increased economic profits while also reducing environmental costs by minimizing agricultural inputs or adopting other emission reduction strategies33. The environmental costs, human health, and socioeconomic benefits of crops must be comprehensively considered, focusing on balancing various factors of crop production, reducing crop life cycle pollutant emissions, and achieving sustainability in the environment, human health, and ecosystem economy.

Uncertainty

In this study, we used an LCA to quantify the environmental performance of the main crop production systems in Yunnan Province, China, but there are certain limita-tions associated with this method. First, the data sources for the planting area, energy consumption, fertiliser and pesticide use, and crop yields were obtained from national statistical data24,25,26, and there were differences between these data and those from specific regions or farmlands, which could have affected the accuracy of the results. The second limitation is associated with the uncertainty and variability of the EFs. This study selected IPCC emission factors to evaluate the environmental impacts of crop production in subtropical plateau areas. However, LCA of crop production typically requires considering changes at temporal and spatial scales, and a subtropical plateau region can introduce uncertainty in emission factors due to its unique climate characteristics, geographical environments, soil types, and agricultural management practices35,74. The third limitation is the accuracy in determining the system boundary. LCA is reliant on the scope and boundaries employed to conduct the assessment. Crop production involves a series of links, such as soil management, agricultural material manufacturing, planting methods, product processing, and transportation75; therefore, selecting certain boundaries may cause a degree of uncertainty. The environmental risk predictions in this study are based on current various optimization measures. However, could the long-term optimisation management measures for fertilisers (such as the application of innovative fertilisers, tailored fertilisation, and deep N application) potentially maintain stable effects on enhancing efficiency and reducing emissions? Research indicates that the optimisation effects of carbon sequestration and emission reduction measures could be influenced by the physical and chemical properties of the soil76. Furthermore, with the intensification of future climate change and increasing competition for land, water resources, labor, and energy, the long-term effectiveness of carbon sequestration remains uncertain. Over time, changes in local culture, operational conditions, and socio-economic factors (such as trade, infrastructure, and agricultural policies) may also affect the implementation of these measures. Therefore, future research needs to clarify the interactions among various uncertain factors. Crop production research should focus on addressing these limitations to improve the reliability and applicability of LCA. Nevertheless, this study provides basic information for understanding the impacts of pollutant emissions on crop production in subtropical plateau areas (Fig. S4).

Conclusions

LCA was used to evaluate the environmental impacts of major food crops in Yunnan Province based on statistical data on crop production from 2002 to 2021. To explore emission reduction strategies for food crop production, further analyses were conducted to determine the main factors driving the environmental impacts. The re-sults indicate that, over the 20-year period, the yield of grain crops and the amount of fertiliser used in Yunnan Province showed an increasing trend, and the GHG emissions associated with grain crop production mainly originated from the application of agricultural inputs during the MS. The environmental footprint of crop production in Yunnan Province was higher than that in other regions, and there was a highly sig-nificant correlation between N surplus and environmental impacts. Temperature also influenced the environmental impacts associated with rice growth. A scenario analysis showed that, through advanced crop and nutrient management practices, food crop yields could be increased while reducing environmental risks, and GHG emissions, AP, and EP from the production of rice, wheat, and maize are expected to decrease by 43.0–59.5%, 51.5–64.5%, and 57.4–71.5%, respectively (scenario 4, S4). This study provides a reference for managing the sustainable development of main crop planting systems in subtropical plateau regions.