Objectives Long-term unbalanced fertilization is a major factor leading to nutrient imbalance in tobacco-growing soils, thereby threatening soil quality and tobacco production. This study aimed to construct an evaluation model for tobacco-growing soil quality, along with a Minimum Data Set (MDS) and its corresponding evaluation model, to improve the efficiency of soil quality assessment while reducing costs. The constructed models were applied to a field experiment with long-term unbalanced fertilization to explore the effects of long-term unbalanced nutrient input on soil properties and overall quality in tobacco fields.

Methods Based on the tobacco-growing zones in Guizhou in 2022, 100 representative soil samples were collected from major tobacco-growing areas to measure physical, chemical, and biological indicators. Principal component analysis (PCA) combined with index membership functions was used to construct a Total Data Set (TDS) evaluation model. PCA and stepwise regression analysis were employed to screen the MDS, and a tobacco-growing soil quality evaluation model based on the MDS was developed. Additionally, a long-term unbalanced nutrient input experiment, including treatments of no fertilization (CK), nitrogen-only (N), nitrogen and phosphorus (NP), nitrogen and potassium (NK), and conventional fertilization (NPK), was conducted. Soil quality evaluation indicators were measured and input into the models for analysis.

Results 1) Soil organic matter (SOM), clay content, pH, silt content, and total nitrogen (TN) were high-weight indicators for evaluating tobacco-growing soil quality and were important environmental factors affecting it. 2) The soil quality in Guizhou’s tobacco-growing areas was good, with the Total Data Set Soil Quality Index (TDSSQI) ranging from 34.71 to 75.40, averaging 61.16, which is suitable for flue-cured tobacco cultivation. 3) Compared with stepwise regression analysis, the Soil Quality Index (MDSSQI) obtained from the MDS constructed based on PCA was closer to the TDSSQI (correlation coefficient r=0.81, P< 0.001). The MDS indicators, in descending order of weight, were exchangeable calcium (Ca), available phosphorus (AP), available nitrogen (AN), silt content, clay content, mean weight diameter (MWD), microbial biomass carbon (MBC), available iron (Fe), total potassium (TK), available potassium (AK), carbon-to-nitrogen ratio (C/N ratio), and available molybdenum (Mo). The MDS evaluation model was: MDSSQI=9.918*f_(Ca)+9.619*f_(AP)+9.357*f_(AN)+9.354*f_(silt)+9.923*f_(clay)+8.94*f_(MWD)+8.746*f_(MBC)+8.178*f_(Fe)+7.834*f_(TK)+7.455*f_(AK)+7.312*f_(CNratio)+4.057*f_(Mo). 4) Compared with the balanced fertilization treatment (NPK), long-term nutrient imbalance significantly increased soil pH (P<0.05) and decreased microbial activity. The NK treatment significantly reduced SOM, available sulfur (S), available Fe, available copper (Cu), available zinc (Zn), and phosphorus (TP, AP) contents (P<0.05). The NP treatment significantly decreased silt content, AK, and multiple trace element contents while significantly increasing clay content (P<0.05). The N and CK treatments significantly reduced SOM, available nutrients, and multiple trace element contents (P<0.05) but had minor effects on exchangeable Ca, C/N ratio, and MWD in the MDS. 5) Compared with the NPK treatment, the TDSSQI and MDSSQI of the NK, NP, N, and CK treatments were significantly reduced by 9.12%, 12.17%, 19.08%, 22.60%, and 13.97%, 12.11%, 28.29%, 33.12%, respectively (P<0.05).

Conclusions The minimum dataset, constructed using 12 core indicators, could ensure the reliability of soil quality assessment while substantially reducing evaluation costs. According to the assessment results based on the minimum dataset, long-term nutrient imbalance alters soil pH, weakens microbial activity, and significantly changes soil texture and nutrient content, thereby degrading the quality of tobacco field soils and impairing their sustainable productivity in Guizhou.