2.1 Data collection
For the datasets used by meta-analysis, we selected peer-reviewed literature on the effects of LUCC on SOC and TN storage in southwest China from 1990 to 2018 by searching major academic databases, e.g., CNKI, CAB Abstracts, Elsevier, China Knowledge Resource Integrated Database, and by sorting the reference lists. The keywords and phrases used for the literature search were: ‘Karst area of China’, ‘land use/land cover change’, ‘soil organic carbon’, ‘total nitrogen’, ‘soil quality’ in title, abstract, or keywords. In addition, each study was further screened for integrity, relevance and scientific merits based on the following steps: (1) Select an appropriate classification for each study: defined as the experimental group and the control group. In the vegetation destruction mode, the land use mode with the minimum human disturbance intensity was selected as the control group (usually forest land and grassland). In the model of returning farmland to forest and grass, farmland was selected as the control group; (2) Ensure that each study was independent and published after 1990. We excluded studies before 1990 because land use patterns did not change much before that year; (3) Make sure the soil sampling depth was more than 10 cm (the minimum soil stratification in this study), and the same study cannot be classified as both vegetation destruction and conversion of farmland to forest and grass (unless clearly stated in the study); (4) Make sure all publications specified the results of SOC or TN concentrations. If the data was graphically represented, the GetData Graph Digitizer (Russian Federation version 2.22) was utilized to obtain the data.
In total, 89 published papers and 939 sets of data were included, consisting of 471 sets of SOC data and 468 sets of TN data (SOC or TN concentration and storage, and soil mass). More specifically, 253 sets of SOC data and 240 sets of TN data belonged to deforestation; 218 groups SOC data and 228 groups TN data belonged to land restoration. According to the land use status classification criteria issued by the Ministry of Land and Resources on November 1, 2017 (GB/T21010-2017), the land cover in the Karst area was categorized into native forest, grassland (including natural grassland, artificial grassland and abandoned farmland), secondary forest, shrubby land, cropland (including dry land and paddy fields), plantation and fruit tree. It is noted that DF-others mainly refers to the transformation of grassland into farmland, while LR-others mainly refers to the transformation of grassland into woodland. Spatial distribution of the study sites where the datasets were collected is shown in Fig. 1.
In this study, we classified the collected data into three categories: (1) different patterns of LUCCs (vegetation destruction; returning farmland to forest and grass); (2) SOC and TN storages accumulated in different soil layers (0-20, 20-30, 30-40, 40-60, and 60-100 cm) in response to LUCCs; (3) four specific impact factors as: climate factors (annual average temperature and annual precipitation), geographical factors (slope and elevation), the duration of LUCCs, soil physicochemical properties (initial SOC/TN storage, BD, TN, total phosphorus, total potassium and pH, clay and water content), which were selected in order to further explain the cumulative SOC/TN storage changes affected by LUCC.