ERSSTv3 data you can compare the northern and southern hemispheres with and without land. LO is the Land and Oceans monthly data and O is the Oceans only monthly data. N and S are the hemispheres north and south using the 0 to 90 data sets that are easily download. If there is no land modification impact, there would be no significant change in the LO-O for Land and Ocean (LO) temperature imbalance or differential between hemispheres and O, Ocean only temperature differential during the instrumental period. On the right axis scale, the differential increases from approximately -0.2 in late 1800s and near 1960 to approximately 0.1 C during 2000 to 2010. There is an approximately 0.3 C impact which would produce between 0.15C and 0.3C "Global" impact depending on choice of data set, smoothing and start-end dates. There is no way to separate out glacial melt from UHI to agriculture, but there is a significant impact due to some modification of land area.
"Green Revolution refers to a series of research, development, and technology transfer
initiatives, occurring between the 1940s and the late 1960s, that
increased agriculture production worldwide, particularly in the
developing world, beginning most markedly in the late 1960s. The initiatives, led by Norman Borlaug,
the "Father of the Green Revolution" credited with saving over a
billion people from starvation, involved the development of
high-yielding varieties of cereal grains, expansion of irrigation
infrastructure, modernization of management techniques, distribution of
hybridized seeds, synthetic fertilizers, and pesticides to farmers."
The correlation between the green revolution and the land modification impact on global climate is at least interesting except for the most devoted minions of the Great and Powerful Carbon. The is other land use data available, but the data is very coarse, just decade averages in most cases. I may see if I can add some of that during half time.
UPDATE: Since many seem to be in love with the Berkley Earth Surface Temperature data here is a comparison of the BEST NH land only with the ERSST3 Northern Oceans 20N-90N which shows more variation than the full hemisphere data.
Just to show how baseline dependent BEST is, this chart uses Tmax and Tmin with the 2000-2012 baseline and produces a new Tave.