2 Materials and Methods
2.1 Literature search
We considered all articles returned by a search using the Web of Science and the China National Knowledge Infrastructure return with the following terms: [topic* AND (potassium OR straw return OR residue cycle) AND (soil potassium fertility OR soil available OR potassium balance OR potassium-bearing minerals OR yield) AND (rice OR wheat OR winter oilseed rape)]. All papers were published between January 1990 and June 2019. The search was limited to the following principles: all tests were conducted under field conditions, all fertilizer inputs were from simple chemical fertilizers, and the planting mode was single cropping or rotation cropping in the Yangtze River basin. After sorting, we obtained 55 papers representing 215 cases for rice, 157 for wheat, and 95 for winter oilseed rape (Table S1). Another 227 cases for soil available K content in different K management strategies were collected for a meta-analysis (Table S2). Finally, datasets for soil available K during the period 2007-2008 were compiled from the Chinese national soil testing and fertilizer recommendation program database.
2.2 Soil sampling and analysis
All 228 soil samples (10 cores per site) were collected for soil K distribution from the top 20 cm of the soil profile at all sites after crop harvest in the Yangtze River basin in China during 2018-2019. All samples were air dried and crushed to pass through 1 mm sieve for chemical analysis. Soil available K was measured using NH4OAc extraction and a flame photometer method (Jackson, 1958).
2.3 Data analysis
ArcGIS 10.2 software was used to map soil K distribution. A meta-analysis using MetaWin 2.1 was conducted to assess the effects of K fertilization, straw return, and concurrent use both practices on crop yields and soil K content. Statistical analyses were performed to evaluate the range of variability and standard deviation (S.D.) using SPSS 18.0. All figures were drawn uisng Origin 8.0.