deletions | additions       diff --git a/untitled.tex b/untitled.tex  index eb723ab..e69351f 100644  --- a/untitled.tex  +++ b/untitled.tex 
...
The El Niño events in the Tropical Andes produce a below-average precipitation pattern \cite{Garreaud2009}. Yet, this response is not uniform in the Ecuadorian Andes \cite{Villacis2003}. No relationship exists between ENSO and precipitation during the wet season (October-May) \cite{Villacis2003}. Yet, there is a consistent trend with below-average precipitation during the dry season (June-August) \cite{Villacis2003}. This may occur because precipitation has some Atlantic influence as well \cite{Vuille_2000}. Additionally, ENSO can largely explain the inter-annual variability of temperature \cite{Vuille_2000}. El Niño events are associated with higher temperatures in most of the Ecuadorian Andes \cite{Vuille_2000}. However, these studies relied on a relatively short period —approximately 30 years— of weather station data. Moreover, they include only some wildfire variables —temperature and precipitation—. In spite of this, they show a consistent agreement on the effect of El Niño over environmental conditions conducive to wildfires. This agrees with our overall results; nevertheless this should be carefully analyzed considering the use of reanalysis data in this study.  The general limitations of reanalysis data were discussed in the data and methods section. The results shown in Table 3 demonstrate that the  precipitation produced by ENSO's teleconnections is not well represented in by  the 20CR data. Evethough this variable should be strongly influenced by ENSO, the linear correlations do not show any link. This is especially unexpected for the dry season. The rest of the wildfire variables show a relationship with ENSO in spite of their biases in comparison to weather station data. This biases, and the course resolution of the reanalysis data brings uncertainty to the results. The quantification of these uncertainties is beyond the scope of this study. This could be accomplished examining each 20CR realization. Another approach could be the use of alternative reanalysis products. Although, the limitations may be similar. Another source of uncertainty is the selected wildfire index.