CLIMATE VARIABILITY, AGRICULTURE AND REMOTE SENSING - Food Security, Food Prices and Climate Variability - page 54

Temperature changes are much easier to measure than changes in precipitation, due to the
mixing of the atmosphere and the relative ease of making accurate temperature measurements
( Plate 5 ). Regarding the impact of observed warming on precipitation, in 2007 the fifth Inter-
governmental Panel on Climate Change (IPCC, 2007a) assessment stated:
the increased atmospheric moisture content associated with warming might be expected to
lead to increased global mean precipitation. Global annual land mean precipitation showed
a small, but uncertain upward trend over the 20th century of approximately 1.1 mm per
decade. However, the record is characterized by large inter-decadal variability, and global
annual land mean precipitation shows a non-signiicant decrease since 1950.
To estimate changes in precipitation globally, one of the best datasets to use is the Global
Precipitation Climatology Project (GPCP) precipitation merged dataset for 1979-2010,
version 2.2, which combines precipitation estimates from multiple satellite observations
acquired from Special Sensor Microwave Imager emission and scattering algorithms, GOES
Precipitation Index, Outgoing Longwave Precipitation Index, rain gauges and TIROS Oper-
ational Vertical Sounders on NOAA polar orbiting satellites (Adler et al ., 2003) ( Plate 6 ).
Merged satellite and gauge datasets such as the GPCP have distinct advantages over gauge data
alone, as they provide information in places and times when ground data are not available,
and have a greater chance of observing a rainfall event due to their use of data from multiple
sensors and observations.
Deriving trends and extreme rainfall events from any global precipitation dataset remains
difficult, however, due to the lack of adequate ground observations over much of the devel-
oping world and the high temporal and spatial variability of rainfall (Dinku et al ., 2008). Plate
7 shows the anomaly of the GPCP for 2010 from the 1988 to 2004 base period. The image
has white for non-signiicant trends and colors for where the trends are significant. What the
reader will notice is how few places have rainfall trends, and how most trends are in the
ocean. The analysis does not do well in capturing trends that are societally relevant, for
example changes to the start of the growing season in tropical countries, or changes in rainfall
intensity causing more floods. Documenting these more meaningful rainfall trends requires
extremely careful analysis, lots of very high quality data and a great deal of specific knowledge
of local vulnerabilities, hazards and livelihood strategies.
Rainfall variability
New assessments using proxy datasets such as those derived from the Gravity Recovery and
Climate Experiment show balances in groundwater and global runoff into the oceans have
trends that are quantifiable and significant for local agricultural conditions (Rodell et al ., 2009;
Gosling and Arnell, 2011; Syed et al ., 2010). By merging models with observations, different
disciplines are beginning to demonstrate increasing precipitation intensity and large-scale
changes in drought and flood event occurrence. A recent assessment of ocean salinity
observations from 1950-2000 shows that the global water cycle is indeed intensifying at a rate
of 8 percent per degree of surface warming (Durack et al ., 2012), likely leading to increased
droughts and intense rainfall events over land even if we cannot yet observe these changes.
We can expect increasing weather impacts for agriculture as the climate changes (Ohring and
Gruber, 2001; Turvey, 2001).