To better understand and manage the interactions of agriculture and natural resources, for example under current increasing societal demands and climate changes, agro-environmental research must bring together an ever growing amount of data and information from multiple science domains. Data that is inherently large, multi-dimensional and heterogeneous, and requires computational intensive processing. Thus, agro-environmental researchers must deal with specific Big Data challenges in efficiently acquiring the data fit to their job while limiting the amount of computational, network and storage resources needed to practical levels. Automated procedures for collection, selection, annotation and indexing of data and metadata are indispensable in order to be able to effectively exploit the global network of available scientific information. This paper describes work performed in the EU FP7 Trees4Future and SemaGrow projects that contributes to development and evaluation of an infrastructure that allows efficient discovery and unified querying of agricultural and forestry resources using Linked Data and semantic technologies.