143
12.2 A Systematic Approach
This systematic map relies on a data set compiled as part of the CSA Compendium
“The Compendium”. The Compendium created search terms relevant to one of 102
technologies including new inputs and farm management practices (58 agronomic,
15 agroforestry, 19 livestock, 5 energy, or 5 postharvest management practices) on
more than 57 outcomes in productivity, resilience or mitigation, such as, yields,
gender differentiated labour use, or soil organic carbon, respectively. Studies were
included based on four inclusion criteria: (1) conducted in a tropical developing
country, (2) included conventional control practice and a practice being suggested
as CSA, (3) contained primary data on the impacts on at least one of the indicators
of interest and (4) conducted in the field (i.e., no modelling studies). Lists of the
search terms for practices and outcomes and additional details on the inclusion criteria can be found in the systematic review protocol (Rosenstock et al. 2015b).
Studies were identified by searching the Web of Science and Scopus databases
using search terms indicative of practices and outcomes. Our search found 150,367
candidate studies, 7497 of which were included in the final Compendium library
based initially on abstract/title reviews and then full text reviews. Out of these, 313
studies were conducted in one of the five countries. Data were compiled into an
Excel database manually from each study. Data retrieved from the selected studies
include information on location, climate, soils, crops, livestock species and outcome values for both conventional (non-CSA control) and treatment practices.
Frequency and distribution of components in the data set (i.e., practices, outcomes
and products) are analysed by summary statistics.
12.3 The Evidence
More than 150 studies met our inclusion criteria for this paper and were included in
the data set analysed here. The data set contains 12,509 data points that compare a
conventional practice with a potential CSA practice in a specific time and place. For
example, the comparison of conservation agriculture versus conventional agriculture at Chitedze Agricultural Research Station, Malawi in 2007 (see Thierfelder
et al. 2013). Studies were unevenly distributed across the five countries with a tenfold difference in the number of studies conducted in the most studied country
(Tanzania) versus the least studied country (Mozambique) (Fig. 12.1). The studies
were primarily conducted on research stations where 58% of data was generated
compared with 42% on farmers’ fields or in household surveys. This is significant
because research on station under scientist-controlled conditions often outperforms
the same practice in farmers’ fields due to the higher quality of implementing the
practices and historical management of the site (Cook et al. 2013). Thus, the evidence will generally reflect the upper bound of what can be achieved by farmers.
12 What Is the Evidence Base for Climate-Smart Agriculture in East and Southern…
12.2 A Systematic Approach
This systematic map relies on a data set compiled as part of the CSA Compendium
“The Compendium”. The Compendium created search terms relevant to one of 102
technologies including new inputs and farm management practices (58 agronomic,
15 agroforestry, 19 livestock, 5 energy, or 5 postharvest management practices) on
more than 57 outcomes in productivity, resilience or mitigation, such as, yields,
gender differentiated labour use, or soil organic carbon, respectively. Studies were
included based on four inclusion criteria: (1) conducted in a tropical developing
country, (2) included conventional control practice and a practice being suggested
as CSA, (3) contained primary data on the impacts on at least one of the indicators
of interest and (4) conducted in the field (i.e., no modelling studies). Lists of the
search terms for practices and outcomes and additional details on the inclusion criteria can be found in the systematic review protocol (Rosenstock et al. 2015b).
Studies were identified by searching the Web of Science and Scopus databases
using search terms indicative of practices and outcomes. Our search found 150,367
candidate studies, 7497 of which were included in the final Compendium library
based initially on abstract/title reviews and then full text reviews. Out of these, 313
studies were conducted in one of the five countries. Data were compiled into an
Excel database manually from each study. Data retrieved from the selected studies
include information on location, climate, soils, crops, livestock species and outcome values for both conventional (non-CSA control) and treatment practices.
Frequency and distribution of components in the data set (i.e., practices, outcomes
and products) are analysed by summary statistics.
12.3 The Evidence
More than 150 studies met our inclusion criteria for this paper and were included in
the data set analysed here. The data set contains 12,509 data points that compare a
conventional practice with a potential CSA practice in a specific time and place. For
example, the comparison of conservation agriculture versus conventional agriculture at Chitedze Agricultural Research Station, Malawi in 2007 (see Thierfelder
et al. 2013). Studies were unevenly distributed across the five countries with a tenfold difference in the number of studies conducted in the most studied country
(Tanzania) versus the least studied country (Mozambique) (Fig. 12.1). The studies
were primarily conducted on research stations where 58% of data was generated
compared with 42% on farmers’ fields or in household surveys. This is significant
because research on station under scientist-controlled conditions often outperforms
the same practice in farmers’ fields due to the higher quality of implementing the
practices and historical management of the site (Cook et al. 2013). Thus, the evidence will generally reflect the upper bound of what can be achieved by farmers.
12 What Is the Evidence Base for Climate-Smart Agriculture in East and Southern…