Long-term P and K fertilisation strategies and balances affect soil availability indices,crop yield depression risk and N use

The last century has seen a large increase of fertiliser use, along with a subsequent rise of crop productivity. However, in many places its intensive use has become a burden to the environment, and legislation has been introduced to restrict nutrient applications. In combination with changing production scenarios as a result of climate change, this means an improved understanding is needed of how low nutrient availability and climatic stress factors affect yields and yield stability.We examined the long-term effects mineral and organic fertilisation on a nutrient-depleted field, and observed large annual variations: depending on the year, average spring barley yields under unfertilised management (U) were between 17–75% lower than the reference N_½P_½K_½ (60–10–60 kg ha^?1). Yields increased up to 174% under N₁P₁K₁ (120–20–120 kg ha^?1), while animal manure applications at an N availability level corresponding to N₁ were between 79 and 137%. No temporal yield trends could be observed, but long-term changes of Olsen-P and exchangeable K were related to the nutrient balances (inputs-offtake) (r² = 0.60 and 0.59, respectively, P < 0.001).Multiple linear regression analysis was used to examine the effects of the treatments in combination with annual weather variations. The results could be split into two outcomes, 1) a general relation between yields and temperatures for the periods of early spring (P < 0.01, multiple R² = 0.31) and summer (P < 0.001, multiple R² = 0.45), and 2) an interaction between temperature and nutrient applications during crop establishment, leading to a diverse response of relative yields (P < 0.001, multiple R² = 0.64), i.e. relative yield losses under the unfertilised treatment (U) were greater in years with lower spring temperatures, and, conversely, the increased nutrient availability in the fully mineral and organically fertilised treatments could partially alleviate the negative effects.After 13 years of repeated fertilisation, inputs were suspended for a single year and only N was applied to evaluate the residual effects. Yields were significantly affected by the different fertilisation histories (P < 0.001). Likewise, apparent nitrogen recovery tended to improve with previous inputs, but the observations were highly variable.Overall, the analyses agree with the notion that brief periods of stress at a critical stage may significantly affect yields, and confirmed that management of sufficient nutrient availability is critical for maintaining high and stable yields.