Abstract: This study quantitatively assessed the effects of various fertilization regimes on the herbage quality and yield of ‘Gannong No.3’ alfalfa in the main production areas of the Hexi Corridor. In a field experiment conducted in 2016 and 2017, a balanced fertilization regime (N 103.5 kg·ha^-1, P₂O₅ 105 kg·ha^-1, K₂O 90 kg·ha^-1) recommended for alfalfa forage production this area was used as a control and compared with nil fertilization, or with a suboptimal level of one of the three elements, N, P or K, and economic benefit was analyzed by data envelopment analysis (DEA). It was found that compared with no fertilization, all fertilizer treatments significantly increased alfalfa yield and total protein content, reduced the acid and neutral detergent fiber levels, improved relative feeding value, and increased the economic benefit. Compared with balanced fertilization, yield and quality of alfalfa herbage of treatments receiving the suboptimal fertilizer regimes were significantly lower than those of the balanced fertilization treatment, especially in the case of suboptimal P application. In 2016 and 2017, the total yield decreased by 25.9% and 25.7%, respectively, and the total protein concentration by 33.4%, 33.1%, respectively, relative to balanced fertilization, on the suboptimal P treatment. These results demonstrate that phosphorus is the nutrient limiting factor for alfalfa forage production in the desert irrigation area of Hexi, and that the importance-ranking of the three tested fertilizer elements is P>N>K. Analysis of the results using DEA showed that balanced fertilization gave the highest economic return (10679 CNY·ha^-1), while the nil fertilization treatment gave the lowest, and the treatments with suboptimal N, P or K were intermediate and ranked: Partial phosphorus deficiency<partial nitrogen deficiency<partial potassium deficiency.

Key words: Hexi desert irrigation area, alfalfa, fertilization effect, data envelopment analysis, economic benefit