286
Data and Methods
Estimation of Aridity Index for Latin America, Changes,
and Trends
For this research, an aridity index was estimated using the UNEP Index P/PET,
where P is the annual average precipitation and PET is the annual potential evapotranspiration. Gridded precipitation and surface air temperature datasets were
obtained from the University of Delaware at 0.5° grid resolution for the period
1960–2017. Surface air temperature was used to estimate potential evapotranspiration. Potential evapotranspiration was estimated based on the Thornthwaite
method, which relates PET with temperature. For regions with average temperatures less than 0 °C ETP is zero. For regions with average annual temperatures
greater than zero and lower than 26.5 °C, the ETP is 0.16× temperature. For regions
with annual temperatures greater than 26.5 °C the PET is −415.85 + 32.24 T
−0.43 T2.
In order to assess changes in drylands, wetlands, and to analyze possible extent
growth of dry areas in the last years, averages of annual precipitation and temperature were estimated for two periods: 1961–1990 and 1991–2017. Aridity indices
were calculated based on these two period averages and the differences (or changes)
of aridity were estimated. In addition, annual aridity indices were estimated for the
full study period 1961–2017 and a simple linear regression model was applied. This
model allowed us to measure the temporal trends of aridity index for the whole
56-year data time series as well as their associated spatial patterns. Figure 16.1
shows the aridity index map for the period 1961–1990. Most areas of Argentina are
semi-arid and arid and most of central eastern area of Brazil is semi- arid. The western regions of Peru, Ecuador, and Chile show more areas of hyper- arid indices,
especially in Chile. Mexico has semi-arid, arid, and small areas of hyper-arid indices; the drier areas are in Baja California Peninsula.
Figure 16.2 exhibits the aridity index map for the period 1991–2017. Arid regions
in Argentina, and in western areas of Peru, Ecuador, and Chile show similar and
lower aridity indices compared to period 1961–1990. In Mexico, there are more
areas of Arid indices, especially for the central region of the country.
Changes of aridity indices areas were calculated by subtracting first the image of
the 1961–1990 period (Fig. 16.1) minus the image corresponding to the 1991–2017
period (Fig. 16.2). In addition a linear regression model was applied to estimate the
trend (slope of the model) in aridification for each grid cell. Resulting maps can be
seen in the results section of this chapter.
E. Nickl et al.
Data and Methods
Estimation of Aridity Index for Latin America, Changes,
and Trends
For this research, an aridity index was estimated using the UNEP Index P/PET,
where P is the annual average precipitation and PET is the annual potential evapotranspiration. Gridded precipitation and surface air temperature datasets were
obtained from the University of Delaware at 0.5° grid resolution for the period
1960–2017. Surface air temperature was used to estimate potential evapotranspiration. Potential evapotranspiration was estimated based on the Thornthwaite
method, which relates PET with temperature. For regions with average temperatures less than 0 °C ETP is zero. For regions with average annual temperatures
greater than zero and lower than 26.5 °C, the ETP is 0.16× temperature. For regions
with annual temperatures greater than 26.5 °C the PET is −415.85 + 32.24 T
−0.43 T2.
In order to assess changes in drylands, wetlands, and to analyze possible extent
growth of dry areas in the last years, averages of annual precipitation and temperature were estimated for two periods: 1961–1990 and 1991–2017. Aridity indices
were calculated based on these two period averages and the differences (or changes)
of aridity were estimated. In addition, annual aridity indices were estimated for the
full study period 1961–2017 and a simple linear regression model was applied. This
model allowed us to measure the temporal trends of aridity index for the whole
56-year data time series as well as their associated spatial patterns. Figure 16.1
shows the aridity index map for the period 1961–1990. Most areas of Argentina are
semi-arid and arid and most of central eastern area of Brazil is semi- arid. The western regions of Peru, Ecuador, and Chile show more areas of hyper- arid indices,
especially in Chile. Mexico has semi-arid, arid, and small areas of hyper-arid indices; the drier areas are in Baja California Peninsula.
Figure 16.2 exhibits the aridity index map for the period 1991–2017. Arid regions
in Argentina, and in western areas of Peru, Ecuador, and Chile show similar and
lower aridity indices compared to period 1961–1990. In Mexico, there are more
areas of Arid indices, especially for the central region of the country.
Changes of aridity indices areas were calculated by subtracting first the image of
the 1961–1990 period (Fig. 16.1) minus the image corresponding to the 1991–2017
period (Fig. 16.2). In addition a linear regression model was applied to estimate the
trend (slope of the model) in aridification for each grid cell. Resulting maps can be
seen in the results section of this chapter.
E. Nickl et al.