diff --git a/lciafmt/data/lcia.bib b/lciafmt/data/lcia.bib
index 1fd71f2..5f96129 100644
--- a/lciafmt/data/lcia.bib
+++ b/lciafmt/data/lcia.bib
@@ -16,3 +16,32 @@ @article{huijbregts_recipe_2017
 	year = {2017},
 	DOI = {10.1007/s11367-016-1246-y}
 }
+
+@incollection{forster_changes_2007,
+	address = {Cambridge, UK and New York, NY},
+	title = {Changes in {Atmospheric} {Constituents} and in {Radiative} {Forcing}},
+	booktitle = {Climate {Change} 2007: {The} {Physical} {Science} {Basis}. {Contribution} of {Working} {Group} {I} to the {Fourth} {Assessment} {Report} of the {Intergovernmental} {Panel} on {Climate} {Change}},
+	publisher = {Cambridge University Press},
+	author = {Forster, Piers and Ramaswamy, Venkatachalam},
+	year = {2007},
+}
+
+@incollection{myhre_anthropogenic_2013,
+	address = {Cambridge, UK and New York, NY},
+	title = {Anthropogenic and {Natural} {Radiative} {Forcing}},
+	url = {https://archive.ipcc.ch/pdf/assessment-report/ar5/wg1/WG1AR5_Chapter08_FINAL.pdf},
+	booktitle = {Climate {Change} 2013: {The} {Physical} {Science} {Basis}. {Contribution} of {Working} {Group} {I} to the {Fifth} {Assessment} {Report} of the {Intergovernmental} {Panel} on {Climate} {Change}},
+	publisher = {Cambridge University Press},
+	author = {Myhre, Gunnar and Shindell, Drew},
+	year = {2013},
+}
+
+@incollection{smith_earths_2021,
+	address = {Cambridge, UK and New York, NY},
+	title = {The {Earth}'s {Energy} {Budget}, {Climate} {Feedbacks}, and {Climate} {Sensitivity} {Supplementary} {Material}},
+	url = {https://www.ipcc.ch/report/ar6/wg1/downloads/report/IPCC_AR6_WGI_Chapter07_SM.pdf},
+	booktitle = {Climate {Change} 2021: {The} {Physical} {Science} {Basis}. {Contribution} of {Working} {Group} {I} to the {Sixth} {Assessment} {Report} of the {Intergovernmental} {Panel} on {Climate} {Change}},
+	publisher = {Cambridge University Press},
+	author = {Smith, Chris and Nicholls, Zebedee R. J. and Armour, Kyle and Collins, William and Forster, Piers and Meinshausen, Malte and Palmer, Matthew D. and Watanabe, Masahiro},
+	year = {2021},
+}
diff --git a/lciafmt/data/methods.json b/lciafmt/data/methods.json
index 92314dc..d971d5f 100644
--- a/lciafmt/data/methods.json
+++ b/lciafmt/data/methods.json
@@ -61,6 +61,10 @@
     "path": "ipcc",
     "case_insensitivity": "False",
     "url": "https://github.com/USEPA/LCIAformatter/blob/master/lciafmt/data/IPCC_GWP_values.csv",
+    "bib_id": {"AR6": "smith_earths_2021",
+               "AR5": "myhre_anthropogenic_2013",
+               "AR4": "forster_changes_2007"
+                },
     "citation": "Forster and Ramaswamy 2007 (AR4), Myhre and Shindell 2013 (AR5), Forster and Storelvmo 2021 (AR6)",
     "source_type": "csv"
   }
diff --git a/lciafmt/jsonld.py b/lciafmt/jsonld.py
index b2b0f0c..560b1cd 100644
--- a/lciafmt/jsonld.py
+++ b/lciafmt/jsonld.py
@@ -30,6 +30,7 @@ def __init__(self, zip_file: str):
         self.__indicators = {}
         self.__flows = {}
         self.__sources = {}
+        self.__sources_to_write = {}
         self.__bibids = {}
         self.__bibpath = datapath / 'lcia.bib'
 
@@ -47,11 +48,24 @@ def write(self, df: pd.DataFrame, write_flows=False):
         if 'category' not in df:
             df['category'] = df['Method']
 
-        for method in df['source_method'].unique():
+        methods = pd.unique(
+                df[['Method', 'source_method']].values.ravel('K'))
+        indicators = pd.unique(
+                df[['Indicator', 'source_indicator']].values.ravel('K'))
+
+        # identify all relevant bib_ids and sources
+        for method in methods:
             m = check_as_class(method)
+            if isinstance(m, str):
+                # not a recognized method, so no bib_id
+                continue
             bib = m.get_metadata().get('bib_id')
             if bib:
-                self.__bibids[bib] = m.value
+                if isinstance(bib, str):
+                    self.__bibids[bib] = m.value
+                elif isinstance(bib, dict):
+                    for k,v in bib.items():
+                        self.__bibids[v] = f'{m.value} {k}'
         for i in generate_sources(self.__bibpath, self.__bibids):
             self.__sources[i.id] = i
 
@@ -69,7 +83,7 @@ def write(self, df: pd.DataFrame, write_flows=False):
         dicts = [
             self.__indicators,
             self.__methods,
-            self.__sources
+            self.__sources_to_write
         ]
         if write_flows:
             dicts.append(self.__flows)
@@ -100,9 +114,14 @@ def __indicator(self, row) -> o.ImpactCategory:
                                                       row['source_indicator'])
         ind.impact_factors = []
         ind.version = pkg_version_number
-        source = self._return_source(row['Method'])
+        source = (self._return_source(row['source_method']) or
+                  self._return_source(row['Method'] + ' ' +
+                                      row['Indicator']) or
+                  self._return_source(row['source_method'] + ' ' +
+                                      row['source_indicator']))
         if source:
             ind.source = source.to_ref()
+            self.__sources_to_write[source.id] = source
         self.__indicators[uid] = ind
 
         method = self.__method(row)
@@ -121,6 +140,10 @@ def __method(self, row) -> o.ImpactMethod:
         m.id = uid
         m.name = row['Method']
         m.version = pkg_version_number
+        source = self._return_source(row['Method'])
+        if source:
+            m.source = source.to_ref()
+            self.__sources_to_write[source.id] = source
         m.impact_categories = []
         m.description = generate_method_description(row['Method'])
         self.__methods[uid] = m
@@ -160,6 +183,6 @@ def __flow(self, row):
 
     def _return_source(self, name):
         for uid, s in self.__sources.items():
-            if s.name == name:
+            if s.name == name or name.startswith(s.name):
                 return s
         return None