diff --git a/CHANGES.md b/CHANGES.md
index 98a373fa6..5ff51b16b 100644
--- a/CHANGES.md
+++ b/CHANGES.md
@@ -11,9 +11,11 @@
 
 * validation: we no longer exit with a non-zero exit code when the requested validation mode is "warn" [#1440][] (@jameshadfield)
 * validation: we no longer perform any validation when the requested validation mode is "skip" [#1440][] (@jameshadfield)
+* filter: Send all log messages to `stderr`. This allows output to be written to `stdout` (e.g. `--output-strains /dev/stdout`). [#1459][] (@victorlin)
 
 [#1440]: https://github.com/nextstrain/augur/pull/1440
 [#1445]: https://github.com/nextstrain/augur/pull/1445
+[#1459]: https://github.com/nextstrain/augur/pull/1459
 
 ## 24.3.0 (18 March 2024)
 
diff --git a/augur/filter/_run.py b/augur/filter/_run.py
index 3cab70cae..9a82c7f00 100644
--- a/augur/filter/_run.py
+++ b/augur/filter/_run.py
@@ -277,9 +277,9 @@ def run(args):
             raise AugurError(error)
 
         if (probabilistic_used):
-            print(f"Sampling probabilistically at {sequences_per_group:0.4f} sequences per group, meaning it is possible to have more than the requested maximum of {args.subsample_max_sequences} sequences after filtering.")
+            print_err(f"Sampling probabilistically at {sequences_per_group:0.4f} sequences per group, meaning it is possible to have more than the requested maximum of {args.subsample_max_sequences} sequences after filtering.")
         else:
-            print(f"Sampling at {sequences_per_group} per group.")
+            print_err(f"Sampling at {sequences_per_group} per group.")
 
         if queues_by_group is None:
             # We know all of the possible groups now from the first pass through
@@ -414,10 +414,10 @@ def run(args):
     total_strains_passed = len(valid_strains)
     total_strains_filtered = len(metadata_strains) + num_excluded_by_lack_of_metadata - total_strains_passed
 
-    print(f"{total_strains_filtered} {'strain was' if total_strains_filtered == 1 else 'strains were'} dropped during filtering")
+    print_err(f"{total_strains_filtered} {'strain was' if total_strains_filtered == 1 else 'strains were'} dropped during filtering")
 
     if num_excluded_by_lack_of_metadata:
-        print(f"\t{num_excluded_by_lack_of_metadata} had no metadata")
+        print_err(f"\t{num_excluded_by_lack_of_metadata} had no metadata")
 
     report_template_by_filter_name = {
         include_exclude_rules.filter_by_sequence_index.__name__: "{count} had no sequence data",
@@ -446,11 +446,11 @@ def run(args):
         parameters["count"] = count
         parameters["were"] = "was" if count == 1 else "were"
         parameters["they"] = "it"  if count == 1 else "they"
-        print("\t" + report_template_by_filter_name[filter_name].format(**parameters))
+        print_err("\t" + report_template_by_filter_name[filter_name].format(**parameters))
 
     if (group_by and args.sequences_per_group) or args.subsample_max_sequences:
         seed_txt = ", using seed {}".format(args.subsample_seed) if args.subsample_seed else ""
-        print(f"\t{num_excluded_subsamp} {'was' if num_excluded_subsamp == 1 else 'were'} dropped because of subsampling criteria{seed_txt}")
+        print_err(f"\t{num_excluded_subsamp} {'was' if num_excluded_subsamp == 1 else 'were'} dropped because of subsampling criteria{seed_txt}")
 
     if total_strains_passed == 0:
         empty_results_message = "All samples have been dropped! Check filter rules and metadata file format."
@@ -463,4 +463,4 @@ def run(args):
         else:
             raise ValueError(f"Encountered unhandled --empty-output-reporting method {args.empty_output_reporting!r}")
 
-    print(f"{total_strains_passed} {'strain' if total_strains_passed == 1 else 'strains'} passed all filters")
+    print_err(f"{total_strains_passed} {'strain' if total_strains_passed == 1 else 'strains'} passed all filters")
diff --git a/augur/io/print.py b/augur/io/print.py
index aff911db6..6ba31aab5 100644
--- a/augur/io/print.py
+++ b/augur/io/print.py
@@ -2,4 +2,6 @@
 
 
 def print_err(*args):
+    """Print to stderr. When data goes to stdout (most cases), this should be
+    used for any informational messages, not just errors/warnings."""
     print(*args, file=sys.stderr)
diff --git a/tests/functional/filter/cram/filter-empty-output-reporting.t b/tests/functional/filter/cram/filter-empty-output-reporting.t
index 5f90a8f62..61a7270f0 100644
--- a/tests/functional/filter/cram/filter-empty-output-reporting.t
+++ b/tests/functional/filter/cram/filter-empty-output-reporting.t
@@ -10,26 +10,30 @@ Test the default behavior for empty results is an error.
   >  --metadata "$TESTDIR/../data/metadata.tsv" \
   >  --exclude-all \
   >  --output-strains filtered_strains.txt > /dev/null
+  12 strains were dropped during filtering
+  	12 were dropped by `--exclude-all`
   ERROR: All samples have been dropped! Check filter rules and metadata file format.
   [2]
   $ wc -l filtered_strains.txt
   \s*0 .* (re)
 
 Repeat with the --empty-output-reporting=warn option.
-This whould output a warning message but no error.
+This should output a warning message but no error.
 
   $ ${AUGUR} filter \
   >  --metadata "$TESTDIR/../data/metadata.tsv" \
   >  --exclude-all \
   >  --output-strains filtered_strains.txt \
   >  --empty-output-reporting warn > /dev/null
+  12 strains were dropped during filtering
+  	12 were dropped by `--exclude-all`
   WARNING: All samples have been dropped! Check filter rules and metadata file format.
+  0 strains passed all filters
   $ wc -l filtered_strains.txt
   \s*0 .* (re)
 
 Ignore empty results with the --empty-output-reporting=silent option.
 Make sure all 3 output types are empty, except the metadata output should still include the header.
-This should not output any messages to stderr.
 
   $ ${AUGUR} filter \
   >  --metadata "$TESTDIR/../data/metadata.tsv" \
@@ -38,7 +42,7 @@ This should not output any messages to stderr.
   >  --output-sequences filtered_seqs.fasta \
   >  --output-metadata filtered_metadata.tsv \
   >  --output-strains filtered_strains.txt \
-  >  --empty-output-reporting silent > /dev/null
+  >  --empty-output-reporting silent 2>/dev/null
   $ wc -l filtered_seqs.fasta
   \s*0 .* (re)
   $ diff <(head -n 1 filtered_metadata.tsv) <(head -n 1 "$TESTDIR/../data/metadata.tsv")
diff --git a/tests/functional/filter/cram/filter-exclude-include.t b/tests/functional/filter/cram/filter-exclude-include.t
index 1fa508bce..2fd9acb7b 100644
--- a/tests/functional/filter/cram/filter-exclude-include.t
+++ b/tests/functional/filter/cram/filter-exclude-include.t
@@ -10,6 +10,6 @@ Force include one South American record by country to get two total records.
   >  --metadata "$TESTDIR/../data/metadata.tsv" \
   >  --exclude-where "region=South America" "region=North America" "region=Southeast Asia" \
   >  --include-where "country=Ecuador" \
-  >  --output-strains filtered_strains.txt > /dev/null
+  >  --output-strains filtered_strains.txt 2>/dev/null
   $ wc -l filtered_strains.txt
   \s*2 .* (re)
diff --git a/tests/functional/filter/cram/filter-exclude-where-multiple.t b/tests/functional/filter/cram/filter-exclude-where-multiple.t
index a9676d914..d96a03ad2 100644
--- a/tests/functional/filter/cram/filter-exclude-where-multiple.t
+++ b/tests/functional/filter/cram/filter-exclude-where-multiple.t
@@ -16,7 +16,7 @@ Scenario 1: Run command with one --exclude-where flag and multiple values
   $ ${AUGUR} filter \
   >   --metadata metadata.tsv \
   >  --exclude-where "region=A" "region=B" \
-  >  --output-strains filtered_strains.txt > /dev/null
+  >  --output-strains filtered_strains.txt 2>/dev/null
 
 Both exclusions are applied.
 
@@ -30,7 +30,7 @@ Scenario 2: Run command with two --exclude-where flags
   >   --metadata metadata.tsv \
   >  --exclude-where "region=A" \
   >  --exclude-where "region=B" \
-  >  --output-strains filtered_strains.txt > /dev/null
+  >  --output-strains filtered_strains.txt 2>/dev/null
 
 Both exclusions are applied.
 
diff --git a/tests/functional/filter/cram/filter-force-include-no-duplicates.t b/tests/functional/filter/cram/filter-force-include-no-duplicates.t
index 5e5f4dacc..2dbca85f3 100644
--- a/tests/functional/filter/cram/filter-force-include-no-duplicates.t
+++ b/tests/functional/filter/cram/filter-force-include-no-duplicates.t
@@ -37,7 +37,7 @@ Test all outputs with --include-where.
   >   --output-metadata metadata-filtered.tsv \
   >   --output-strains strains-filtered.txt \
   >   --output-sequences sequences-filtered.fasta \
-  >   > /dev/null 2>&1
+  >   2>/dev/null
   $ cat metadata-filtered.tsv | tail -n+2 | sort -k1
   a\t1 (esc)
   b\t2 (esc)
@@ -72,7 +72,7 @@ Test all outputs with --include.
   >   --output-metadata metadata-filtered.tsv \
   >   --output-strains strains-filtered.txt \
   >   --output-sequences sequences-filtered.fasta \
-  >   > /dev/null 2>&1
+  >   2>/dev/null
   $ cat metadata-filtered.tsv | tail -n+2 | sort -k1
   a\t1 (esc)
   b\t2 (esc)
diff --git a/tests/functional/filter/cram/filter-max-date.t b/tests/functional/filter/cram/filter-max-date.t
index 960ebc31b..5bea06dc0 100644
--- a/tests/functional/filter/cram/filter-max-date.t
+++ b/tests/functional/filter/cram/filter-max-date.t
@@ -16,7 +16,7 @@ Test that --max-date is inclusive.
   $ ${AUGUR} filter \
   >  --metadata metadata.tsv \
   >  --max-date 2020-03-01 \
-  >  --output-strains filtered_strains.txt > /dev/null
+  >  --output-strains filtered_strains.txt 2>/dev/null
   $ sort filtered_strains.txt
   SEQ_1
   SEQ_2
diff --git a/tests/functional/filter/cram/filter-metadata-date-formats.t b/tests/functional/filter/cram/filter-metadata-date-formats.t
index 7203fbae6..b267be762 100644
--- a/tests/functional/filter/cram/filter-metadata-date-formats.t
+++ b/tests/functional/filter/cram/filter-metadata-date-formats.t
@@ -16,7 +16,7 @@ Test that 2020 is evaluated as 2020-XX-XX.
   $ ${AUGUR} filter \
   >  --metadata metadata.tsv \
   >  --min-date 2020-02-01 \
-  >  --output-strains filtered_strains.txt > /dev/null
+  >  --output-strains filtered_strains.txt 2>/dev/null
   $ sort filtered_strains.txt
   SEQ_2
   SEQ_3
@@ -26,7 +26,7 @@ Test that 2020.0, 2020, and 2020-XX-XX all pass --min-date 2019
   $ ${AUGUR} filter \
   >  --metadata metadata.tsv \
   >  --min-date 2019 \
-  >  --output-strains filtered_strains.txt > /dev/null
+  >  --output-strains filtered_strains.txt 2>/dev/null
   $ sort filtered_strains.txt
   SEQ_1
   SEQ_2
diff --git a/tests/functional/filter/cram/filter-metadata-delimiter.t b/tests/functional/filter/cram/filter-metadata-delimiter.t
index 971a6c5e6..324dc9d9f 100644
--- a/tests/functional/filter/cram/filter-metadata-delimiter.t
+++ b/tests/functional/filter/cram/filter-metadata-delimiter.t
@@ -13,7 +13,7 @@ Comma-delimited metadata is allowed by default. However, the output metadata wil
   $ ${AUGUR} filter \
   >  --metadata metadata.txt \
   >  --exclude-where column=A \
-  >  --output-metadata filtered.txt > /dev/null
+  >  --output-metadata filtered.txt 2>/dev/null
   $ cat filtered.txt
   strain\tcolumn (esc)
   SEQ_2\tB (esc)
@@ -62,7 +62,7 @@ Allow colon-delimited metadata. However, the output metadata will be tab-delimit
   >  --metadata metadata.txt \
   >  --metadata-delimiters ':' \
   >  --exclude-where column=A \
-  >  --output-metadata filtered.txt > /dev/null
+  >  --output-metadata filtered.txt 2>/dev/null
   $ cat filtered.txt
   strain\tcolumn (esc)
   SEQ_2\tB (esc)
diff --git a/tests/functional/filter/cram/filter-min-date.t b/tests/functional/filter/cram/filter-min-date.t
index ea13bb0fc..7548fb7ec 100644
--- a/tests/functional/filter/cram/filter-min-date.t
+++ b/tests/functional/filter/cram/filter-min-date.t
@@ -16,7 +16,7 @@ Test that --min-date is inclusive.
   $ ${AUGUR} filter \
   >  --metadata metadata.tsv \
   >  --min-date 2020-02-26 \
-  >  --output-strains filtered_strains.txt > /dev/null
+  >  --output-strains filtered_strains.txt 2>/dev/null
   $ sort filtered_strains.txt
   SEQ_1
   SEQ_2
diff --git a/tests/functional/filter/cram/filter-mismatched-sequences-error.t b/tests/functional/filter/cram/filter-mismatched-sequences-error.t
index 1c92681ae..5930234fa 100644
--- a/tests/functional/filter/cram/filter-mismatched-sequences-error.t
+++ b/tests/functional/filter/cram/filter-mismatched-sequences-error.t
@@ -13,6 +13,9 @@ This should produce no results because the intersection of metadata and sequence
   >  --max-date 2020-01-30 \
   >  --output-strains filtered_strains.txt > /dev/null
   Note: You did not provide a sequence index, so Augur will generate one. You can generate your own index ahead of time with `augur index` and pass it with `augur filter --sequence-index`.
+  13 strains were dropped during filtering
+  	1 had no metadata
+  	12 had no sequence data
   ERROR: All samples have been dropped! Check filter rules and metadata file format.
   [2]
   $ wc -l filtered_strains.txt
@@ -27,6 +30,9 @@ Repeat with sequence and strain outputs. We should get the same results.
   >  --output-strains filtered_strains.txt \
   >  --output-sequences filtered.fasta > /dev/null
   Note: You did not provide a sequence index, so Augur will generate one. You can generate your own index ahead of time with `augur index` and pass it with `augur filter --sequence-index`.
+  13 strains were dropped during filtering
+  	1 had no metadata
+  	12 had no sequence data
   ERROR: All samples have been dropped! Check filter rules and metadata file format.
   [2]
   $ wc -l filtered_strains.txt
@@ -42,6 +48,9 @@ Since we expect metadata to be filtered by presence of strains in input sequence
   >  --metadata "$TESTDIR/../data/metadata.tsv" \
   >  --output-strains filtered_strains.txt > /dev/null
   Note: You did not provide a sequence index, so Augur will generate one. You can generate your own index ahead of time with `augur index` and pass it with `augur filter --sequence-index`.
+  13 strains were dropped during filtering
+  	1 had no metadata
+  	12 had no sequence data
   ERROR: All samples have been dropped! Check filter rules and metadata file format.
   [2]
   $ wc -l filtered_strains.txt
diff --git a/tests/functional/filter/cram/filter-numerical-ids.t b/tests/functional/filter/cram/filter-numerical-ids.t
index f6167a962..66998d12f 100644
--- a/tests/functional/filter/cram/filter-numerical-ids.t
+++ b/tests/functional/filter/cram/filter-numerical-ids.t
@@ -25,7 +25,7 @@ Test that nothing is filtered out due to missing sequence data.
   >   --metadata metadata.tsv \
   >   --sequences sequences.fasta \
   >   --output-strains filtered_strains.txt \
-  >   > /dev/null 2>&1
+  >   2>/dev/null
   $ sort filtered_strains.txt
   1
   2
diff --git a/tests/functional/filter/cram/filter-output-contents.t b/tests/functional/filter/cram/filter-output-contents.t
index 4f16002a4..a66ae4b09 100644
--- a/tests/functional/filter/cram/filter-output-contents.t
+++ b/tests/functional/filter/cram/filter-output-contents.t
@@ -15,7 +15,7 @@ The purpose of this test file is to check format and consistency among the
   >  --no-probabilistic-sampling \
   >  --output-metadata filtered_metadata.tsv \
   >  --output-strains filtered_strains.txt \
-  >  --output filtered.fasta > /dev/null
+  >  --output filtered.fasta 2>/dev/null
 
 Check that the header row is identical between input and output metadata.
 
diff --git a/tests/functional/filter/cram/filter-output-metadata-header.t b/tests/functional/filter/cram/filter-output-metadata-header.t
index 862977e33..15ff155e1 100644
--- a/tests/functional/filter/cram/filter-output-metadata-header.t
+++ b/tests/functional/filter/cram/filter-output-metadata-header.t
@@ -14,7 +14,7 @@ Quoted columns containing the tab delimiter are left unchanged.
 
   $ ${AUGUR} filter \
   >  --metadata metadata.tsv \
-  >  --output-metadata filtered_metadata.tsv > /dev/null
+  >  --output-metadata filtered_metadata.tsv 2>/dev/null
 
   $ head -n 1 filtered_metadata.tsv
   strain	"col	1"
@@ -28,7 +28,7 @@ Quoted columns without the tab delimiter are stripped of the quotes.
 
   $ ${AUGUR} filter \
   >  --metadata metadata.tsv \
-  >  --output-metadata filtered_metadata.tsv > /dev/null
+  >  --output-metadata filtered_metadata.tsv 2>/dev/null
 
   $ head -n 1 filtered_metadata.tsv
   strain	col1
@@ -42,7 +42,7 @@ Any other columns with quotes are quoted, and pre-existing quotes are escsaped b
 
   $ ${AUGUR} filter \
   >  --metadata metadata.tsv \
-  >  --output-metadata filtered_metadata.tsv > /dev/null
+  >  --output-metadata filtered_metadata.tsv 2>/dev/null
 
   $ head -n 1 filtered_metadata.tsv
   strain	"col""1"	"col2"""
diff --git a/tests/functional/filter/cram/filter-query-and-include-where.t b/tests/functional/filter/cram/filter-query-and-include-where.t
index 101b918aa..b61891978 100644
--- a/tests/functional/filter/cram/filter-query-and-include-where.t
+++ b/tests/functional/filter/cram/filter-query-and-include-where.t
@@ -17,7 +17,7 @@ Test that --include_where still works with filtering on query.
   >  --metadata metadata.tsv \
   >  --query "quality=='good' & location=='colorado'" \
   >  --include-where "location=nevada" \
-  >  --output-strains filtered_strains.txt > /dev/null
+  >  --output-strains filtered_strains.txt 2>/dev/null
   $ sort filtered_strains.txt
   SEQ_1
   SEQ_3
diff --git a/tests/functional/filter/cram/filter-query-and-include.t b/tests/functional/filter/cram/filter-query-and-include.t
index aeb603c25..a691f365b 100644
--- a/tests/functional/filter/cram/filter-query-and-include.t
+++ b/tests/functional/filter/cram/filter-query-and-include.t
@@ -20,7 +20,7 @@ Test that --include_where still works with filtering on query.
   >  --metadata metadata.tsv \
   >  --query "quality=='good' & location=='colorado'" \
   >  --include include.txt \
-  >  --output-strains filtered_strains.txt > /dev/null
+  >  --output-strains filtered_strains.txt 2>/dev/null
   $ sort filtered_strains.txt
   SEQ_1
   SEQ_3
diff --git a/tests/functional/filter/cram/filter-query-backtick-quoting.t b/tests/functional/filter/cram/filter-query-backtick-quoting.t
index 84c9e48a4..fd6157a39 100644
--- a/tests/functional/filter/cram/filter-query-backtick-quoting.t
+++ b/tests/functional/filter/cram/filter-query-backtick-quoting.t
@@ -17,7 +17,7 @@ The 'region name' column is query-able by backtick quoting.
   $ ${AUGUR} filter \
   >  --metadata metadata.tsv \
   >  --query '(`region name` == "A")' \
-  >  --output-strains filtered_strains.txt > /dev/null
+  >  --output-strains filtered_strains.txt 2>/dev/null
 
   $ sort filtered_strains.txt
   SEQ_1
diff --git a/tests/functional/filter/cram/filter-query-example.t b/tests/functional/filter/cram/filter-query-example.t
index 89f6e2a6e..e36e6e5e0 100644
--- a/tests/functional/filter/cram/filter-query-example.t
+++ b/tests/functional/filter/cram/filter-query-example.t
@@ -8,7 +8,7 @@ First, select strains from Brazil (there should be 1).
   $ ${AUGUR} filter \
   >  --metadata "$TESTDIR/../data/metadata.tsv" \
   >  --query "country == 'Brazil'" \
-  >  --output-strains filtered_strains.brazil.txt > /dev/null
+  >  --output-strains filtered_strains.brazil.txt 2>/dev/null
   $ wc -l filtered_strains.brazil.txt
   \s*1 .* (re)
 
@@ -17,7 +17,7 @@ Then, select strains from Colombia (there should be 3).
   $ ${AUGUR} filter \
   >  --metadata "$TESTDIR/../data/metadata.tsv" \
   >  --query "country == 'Colombia'" \
-  >  --output-strains filtered_strains.colombia.txt > /dev/null
+  >  --output-strains filtered_strains.colombia.txt 2>/dev/null
   $ wc -l filtered_strains.colombia.txt
   \s*3 .* (re)
 
@@ -29,7 +29,7 @@ Finally, exclude all sequences except those from the two sets of strains (there
   >  --metadata "$TESTDIR/../data/metadata.tsv" \
   >  --exclude-all \
   >  --include filtered_strains.brazil.txt filtered_strains.colombia.txt \
-  >  --output filtered.fasta > /dev/null
+  >  --output filtered.fasta 2>/dev/null
   $ grep "^>" filtered.fasta | wc -l
   \s*4 (re)
 
@@ -42,7 +42,7 @@ We should get the same outputs without building a sequence index on the fly, bec
   >  --exclude-all \
   >  --include filtered_strains.brazil.txt filtered_strains.colombia.txt \
   >  --output filtered.fasta \
-  >  --output-metadata filtered.tsv > /dev/null
+  >  --output-metadata filtered.tsv 2>/dev/null
   $ grep "^>" filtered.fasta | wc -l
   \s*4 (re)
 
@@ -58,6 +58,6 @@ Alternately, exclude the sequences from Brazil and Colombia (N=4) and records wi
   >  --sequence-index "$TESTDIR/../data/sequence_index.tsv" \
   >  --metadata "$TESTDIR/../data/metadata.tsv" \
   >  --exclude filtered_strains.brazil.txt filtered_strains.colombia.txt \
-  >  --output filtered.fasta > /dev/null
+  >  --output filtered.fasta 2>/dev/null
   $ grep "^>" filtered.fasta | wc -l
   \s*7 (re)
diff --git a/tests/functional/filter/cram/filter-query-numerical.t b/tests/functional/filter/cram/filter-query-numerical.t
index 5aeb142f8..875c04802 100644
--- a/tests/functional/filter/cram/filter-query-numerical.t
+++ b/tests/functional/filter/cram/filter-query-numerical.t
@@ -17,7 +17,7 @@ The 'coverage' column should be query-able by numerical comparisons.
   $ ${AUGUR} filter \
   >  --metadata metadata.tsv \
   >  --query "coverage >= 0.95" \
-  >  --output-strains filtered_strains.txt > /dev/null
+  >  --output-strains filtered_strains.txt 2>/dev/null
 
   $ sort filtered_strains.txt
   SEQ_2
@@ -52,7 +52,7 @@ However, that is still possible by explicitly specifying that it is a string col
   >  --metadata metadata.tsv \
   >  --query "coverage.str.endswith('.95')" \
   >  --query-columns coverage:str \
-  >  --output-strains filtered_strains.txt > /dev/null
+  >  --output-strains filtered_strains.txt 2>/dev/null
 
   $ sort filtered_strains.txt
   SEQ_2
diff --git a/tests/functional/filter/cram/filter-query-str.t b/tests/functional/filter/cram/filter-query-str.t
index a6b11b231..6ada43eb9 100644
--- a/tests/functional/filter/cram/filter-query-str.t
+++ b/tests/functional/filter/cram/filter-query-str.t
@@ -16,7 +16,7 @@ Create metadata file for testing.
   $ ${AUGUR} filter \
   >  --metadata metadata.tsv \
   >  --query "column.str.startswith('value')" \
-  >  --output-strains filtered_strains.txt > /dev/null
+  >  --output-strains filtered_strains.txt 2>/dev/null
 
   $ sort filtered_strains.txt
   SEQ_1
diff --git a/tests/functional/filter/cram/filter-sequences-vcf.t b/tests/functional/filter/cram/filter-sequences-vcf.t
index bbc433196..f6ddaa6cb 100644
--- a/tests/functional/filter/cram/filter-sequences-vcf.t
+++ b/tests/functional/filter/cram/filter-sequences-vcf.t
@@ -11,6 +11,10 @@ Filter TB strains from VCF and save as a list of filtered strains.
   >  --output filtered.vcf \
   >  --output-strains filtered_strains.txt > /dev/null
   Note: You did not provide a sequence index, so Augur will generate one. You can generate your own index ahead of time with `augur index` and pass it with `augur filter --sequence-index`.
+  162 strains were dropped during filtering
+  	155 had no sequence data
+  	7 were dropped because they were earlier than 2012.0 or missing a date
+  3 strains passed all filters
   $ wc -l filtered_strains.txt
   \s*3 .* (re)
 
diff --git a/tests/functional/filter/cram/filter-subsample-missing-date-parts.t b/tests/functional/filter/cram/filter-subsample-missing-date-parts.t
index 810fac97a..c70e62213 100644
--- a/tests/functional/filter/cram/filter-subsample-missing-date-parts.t
+++ b/tests/functional/filter/cram/filter-subsample-missing-date-parts.t
@@ -20,6 +20,10 @@ track which records were skipped for which reasons.
   >  --subsample-seed 0 \
   >  --output-log log.txt \
   >  --output-strains filtered_strains.txt > /dev/null
+  2 strains were dropped during filtering
+  	1 was dropped during grouping due to ambiguous year information
+  	1 was dropped because of subsampling criteria
+  1 strain passed all filters
   $ cat log.txt
   strain\tfilter\tkwargs (esc)
   SEQ_1\tskip_group_by_with_ambiguous_year\t"[[""date_column"", ""date""]]" (esc)
@@ -37,6 +41,11 @@ month information in their date fields.
   >  --subsample-seed 0 \
   >  --output-log log.txt \
   >  --output-strains filtered_strains.txt > /dev/null
+  2 strains were dropped during filtering
+  	1 was dropped during grouping due to ambiguous year information
+  	1 was dropped during grouping due to ambiguous month information
+  	0 were dropped because of subsampling criteria
+  1 strain passed all filters
   $ cat log.txt
   strain\tfilter\tkwargs (esc)
   SEQ_1\tskip_group_by_with_ambiguous_year\t"[[""date_column"", ""date""]]" (esc)
diff --git a/tests/functional/filter/cram/subsample-5-sequences-without-group-by-no-probabilistic-sampling.t b/tests/functional/filter/cram/subsample-5-sequences-without-group-by-no-probabilistic-sampling.t
index 5fc6c955f..ef3cfc0d1 100644
--- a/tests/functional/filter/cram/subsample-5-sequences-without-group-by-no-probabilistic-sampling.t
+++ b/tests/functional/filter/cram/subsample-5-sequences-without-group-by-no-probabilistic-sampling.t
@@ -13,6 +13,6 @@ This generates a dummy category and subsamples from there. With no-probabilistic
   >  --subsample-max-sequences 5 \
   >  --subsample-seed 314159 \
   >  --no-probabilistic-sampling \
-  >  --output filtered.fasta > /dev/null
+  >  --output filtered.fasta 2>/dev/null
   $ grep ">" filtered.fasta | wc -l
   \s*5 (re)
diff --git a/tests/functional/filter/cram/subsample-8-sequences-no-probabilistic-sampling.t b/tests/functional/filter/cram/subsample-8-sequences-no-probabilistic-sampling.t
index cd648093c..4c9655f2a 100644
--- a/tests/functional/filter/cram/subsample-8-sequences-no-probabilistic-sampling.t
+++ b/tests/functional/filter/cram/subsample-8-sequences-no-probabilistic-sampling.t
@@ -14,6 +14,6 @@ With 8 groups to subsample from (after filtering), this should produce one seque
   >  --subsample-max-sequences 8 \
   >  --subsample-seed 314159 \
   >  --no-probabilistic-sampling \
-  >  --output filtered.fasta > /dev/null
+  >  --output filtered.fasta 2>/dev/null
   $ grep ">" filtered.fasta | wc -l
   \s*8 (re)
diff --git a/tests/functional/filter/cram/subsample-ambiguous-dates-error.t b/tests/functional/filter/cram/subsample-ambiguous-dates-error.t
index ec4194a2a..10969be46 100644
--- a/tests/functional/filter/cram/subsample-ambiguous-dates-error.t
+++ b/tests/functional/filter/cram/subsample-ambiguous-dates-error.t
@@ -19,10 +19,10 @@ Metadata with ambiguous days on all strains should error when grouping by week.
   >   --subsample-seed 0 \
   >   --output-metadata metadata-filtered.tsv \
   >   --output-log filtered_log.tsv
-  ERROR: All samples have been dropped! Check filter rules and metadata file format.
   4 strains were dropped during filtering
   	4 were dropped during grouping due to ambiguous day information
   	0 were dropped because of subsampling criteria
+  ERROR: All samples have been dropped! Check filter rules and metadata file format.
   [2]
   $ cat filtered_log.tsv | grep "skip_group_by_with_ambiguous_day" | wc -l
   \s*4 (re)
@@ -46,10 +46,10 @@ Metadata with ambiguous months on all strains should error when grouping by mont
   >   --subsample-seed 0 \
   >   --output-metadata metadata-filtered.tsv \
   >   --output-log filtered_log.tsv
-  ERROR: All samples have been dropped! Check filter rules and metadata file format.
   4 strains were dropped during filtering
   	4 were dropped during grouping due to ambiguous month information
   	0 were dropped because of subsampling criteria
+  ERROR: All samples have been dropped! Check filter rules and metadata file format.
   [2]
   $ cat filtered_log.tsv | grep "skip_group_by_with_ambiguous_month" | wc -l
   \s*4 (re)
@@ -73,10 +73,10 @@ Metadata with ambiguous years on all strains should error when grouping by year.
   >   --subsample-seed 0 \
   >   --output-metadata metadata-filtered.tsv \
   >   --output-log filtered_log.tsv
-  ERROR: All samples have been dropped! Check filter rules and metadata file format.
   4 strains were dropped during filtering
   	4 were dropped during grouping due to ambiguous year information
   	0 were dropped because of subsampling criteria
+  ERROR: All samples have been dropped! Check filter rules and metadata file format.
   [2]
   $ cat filtered_log.tsv | grep "skip_group_by_with_ambiguous_year" | wc -l
   \s*4 (re)
diff --git a/tests/functional/filter/cram/subsample-group-by-empty-value.t b/tests/functional/filter/cram/subsample-group-by-empty-value.t
index 6a16e9f91..1fdbd4bd8 100644
--- a/tests/functional/filter/cram/subsample-group-by-empty-value.t
+++ b/tests/functional/filter/cram/subsample-group-by-empty-value.t
@@ -22,7 +22,7 @@ I.e. the groups here are:
   >   --sequences-per-group 1 \
   >   --subsample-seed 0 \
   >   --output-log filtered-log.tsv \
-  >   --output-strains filtered-strains.txt > /dev/null
+  >   --output-strains filtered-strains.txt 2>/dev/null
   $ cat filtered-strains.txt
   SEQ1
   SEQ3
diff --git a/tests/functional/filter/cram/subsample-group-by-region-1-sequence-per-group-seed.t b/tests/functional/filter/cram/subsample-group-by-region-1-sequence-per-group-seed.t
index cf56999ce..cecc84006 100644
--- a/tests/functional/filter/cram/subsample-group-by-region-1-sequence-per-group-seed.t
+++ b/tests/functional/filter/cram/subsample-group-by-region-1-sequence-per-group-seed.t
@@ -9,7 +9,7 @@ Filter with subsampling, requesting 1 sequence per group (for a group with 4 dis
   >  --group-by region \
   >  --sequences-per-group 1 \
   >  --subsample-seed 314159 \
-  >  --output-strains filtered_strains.txt > /dev/null
+  >  --output-strains filtered_strains.txt 2>/dev/null
   $ wc -l filtered_strains.txt
   \s*4 .* (re)
 
@@ -20,6 +20,6 @@ By setting the subsample seed above, we should guarantee that we get the same "r
   >  --group-by region \
   >  --sequences-per-group 1 \
   >  --subsample-seed 314159 \
-  >  --output-strains filtered_strains_repeated.txt > /dev/null
+  >  --output-strains filtered_strains_repeated.txt 2>/dev/null
 
   $ diff -u <(sort filtered_strains.txt) <(sort filtered_strains_repeated.txt)
diff --git a/tests/functional/filter/cram/subsample-group-by-with-custom-year-column.t b/tests/functional/filter/cram/subsample-group-by-with-custom-year-column.t
index 0f3dfe1e5..7a33eaeb5 100644
--- a/tests/functional/filter/cram/subsample-group-by-with-custom-year-column.t
+++ b/tests/functional/filter/cram/subsample-group-by-with-custom-year-column.t
@@ -22,6 +22,9 @@ Group by generated year column, and ensure all original columns are still in the
   >  --subsample-seed 0 \
   >  --output-metadata filtered_metadata.tsv > /dev/null
   WARNING: `--group-by year` uses a generated year value from the 'date' column. The custom 'year' column in the metadata is ignored for grouping purposes.
+  3 strains were dropped during filtering
+  	3 were dropped because of subsampling criteria
+  2 strains passed all filters
   $ cat filtered_metadata.tsv
   strain\tdate\tyear\tmonth (esc)
   SEQ1\t2021-01-01\todd\tJanuary (esc)
@@ -37,6 +40,9 @@ Group by generated year and month columns, and ensure all original columns are s
   >  --output-metadata filtered_metadata.tsv > /dev/null
   WARNING: `--group-by month` uses a generated month value from the 'date' column. The custom 'month' column in the metadata is ignored for grouping purposes.
   WARNING: `--group-by year` uses a generated year value from the 'date' column. The custom 'year' column in the metadata is ignored for grouping purposes.
+  2 strains were dropped during filtering
+  	2 were dropped because of subsampling criteria
+  3 strains passed all filters
   $ cat filtered_metadata.tsv
   strain\tdate\tyear\tmonth (esc)
   SEQ1\t2021-01-01\todd\tJanuary (esc)
diff --git a/tests/functional/filter/cram/subsample-group-by-without-force-included-strains.t b/tests/functional/filter/cram/subsample-group-by-without-force-included-strains.t
index 475d6635e..9c92eecc1 100644
--- a/tests/functional/filter/cram/subsample-group-by-without-force-included-strains.t
+++ b/tests/functional/filter/cram/subsample-group-by-without-force-included-strains.t
@@ -16,4 +16,4 @@ We don't filter these strains, so they could be considered for subsampling, but
   >   --include include_old_strains.txt \
   >   --group-by month year \
   >   --subsample-max-sequences 10 \
-  >   --output-metadata metadata-filtered.tsv > /dev/null
+  >   --output-metadata metadata-filtered.tsv 2>/dev/null
diff --git a/tests/functional/filter/cram/subsample-max-sequences-with-probabilistic-sampling-warning.t b/tests/functional/filter/cram/subsample-max-sequences-with-probabilistic-sampling-warning.t
index cb679e676..72ed589f2 100644
--- a/tests/functional/filter/cram/subsample-max-sequences-with-probabilistic-sampling-warning.t
+++ b/tests/functional/filter/cram/subsample-max-sequences-with-probabilistic-sampling-warning.t
@@ -15,6 +15,14 @@ Explicitly use probabilistic subsampling to handle the case when there are more
   >  --probabilistic-sampling \
   >  --output-strains filtered_strains_probabilistic.txt > /dev/null
   WARNING: Asked to provide at most 5 sequences, but there are 8 groups.
+  Sampling probabilistically at 0.6055 sequences per group, meaning it is possible to have more than the requested maximum of 5 sequences after filtering.
+  10 strains were dropped during filtering
+  	1 had no metadata
+  	1 had no sequence data
+  	1 was dropped because it was earlier than 2012.0 or missing a date
+  	1 was dropped during grouping due to ambiguous month information
+  	6 were dropped because of subsampling criteria, using seed 314159
+  3 strains passed all filters
 
 Using the default probabilistic subsampling, should work the same as the previous case.
 
@@ -28,6 +36,14 @@ Using the default probabilistic subsampling, should work the same as the previou
   >  --subsample-seed 314159 \
   >  --output-strains filtered_strains_default.txt > /dev/null
   WARNING: Asked to provide at most 5 sequences, but there are 8 groups.
+  Sampling probabilistically at 0.6055 sequences per group, meaning it is possible to have more than the requested maximum of 5 sequences after filtering.
+  10 strains were dropped during filtering
+  	1 had no metadata
+  	1 had no sequence data
+  	1 was dropped because it was earlier than 2012.0 or missing a date
+  	1 was dropped during grouping due to ambiguous month information
+  	6 were dropped because of subsampling criteria, using seed 314159
+  3 strains passed all filters
 
 By setting the subsample seed above, we should get the same results for both runs.
 
diff --git a/tests/functional/filter/cram/subsample-priority-file.t b/tests/functional/filter/cram/subsample-priority-file.t
index 1801c8d0c..cb5c5058d 100644
--- a/tests/functional/filter/cram/subsample-priority-file.t
+++ b/tests/functional/filter/cram/subsample-priority-file.t
@@ -11,6 +11,6 @@ The two highest priority strains are in these two years.
   >  --group-by year \
   >  --priority "$TESTDIR/../data/priorities.tsv" \
   >  --sequences-per-group 1 \
-  >  --output-strains filtered_strains.txt > /dev/null
+  >  --output-strains filtered_strains.txt 2>/dev/null
 
   $ diff -u <(sort -k 2,2rn -k 1,1 "$TESTDIR/../data/priorities.tsv" | head -n 2 | cut -f 1) <(sort -k 1,1 filtered_strains.txt)
diff --git a/tests/functional/filter/cram/subsample-priority-values.t b/tests/functional/filter/cram/subsample-priority-values.t
index 00fb0cd94..4fc9e0781 100644
--- a/tests/functional/filter/cram/subsample-priority-values.t
+++ b/tests/functional/filter/cram/subsample-priority-values.t
@@ -27,7 +27,7 @@ not have a priority score.
   >  --metadata metadata.tsv \
   >  --priority priorities.tsv \
   >  --subsample-max-sequences 5 \
-  >  --output-strains filtered_strains.txt > /dev/null
+  >  --output-strains filtered_strains.txt 2>/dev/null
   $ sort filtered_strains.txt
   SEQ_1
   SEQ_2
@@ -43,7 +43,7 @@ values).
   >  --metadata metadata.tsv \
   >  --priority priorities.tsv \
   >  --subsample-max-sequences 4 \
-  >  --output-strains filtered_strains.txt > /dev/null
+  >  --output-strains filtered_strains.txt 2>/dev/null
   $ sort filtered_strains.txt
   SEQ_1
   SEQ_2
@@ -56,7 +56,7 @@ Subsample 1 less strain. SEQ_4 should now be dropped.
   >  --metadata metadata.tsv \
   >  --priority priorities.tsv \
   >  --subsample-max-sequences 3 \
-  >  --output-strains filtered_strains.txt > /dev/null
+  >  --output-strains filtered_strains.txt 2>/dev/null
   $ sort filtered_strains.txt
   SEQ_1
   SEQ_2
@@ -68,7 +68,7 @@ Subsample 1 less strain. SEQ_1 should now be dropped.
   >  --metadata metadata.tsv \
   >  --priority priorities.tsv \
   >  --subsample-max-sequences 2 \
-  >  --output-strains filtered_strains.txt > /dev/null
+  >  --output-strains filtered_strains.txt 2>/dev/null
   $ sort filtered_strains.txt
   SEQ_2
   SEQ_3
@@ -79,6 +79,6 @@ Subsample 1 less strain. SEQ_1 should now be dropped.
   >  --metadata metadata.tsv \
   >  --priority priorities.tsv \
   >  --subsample-max-sequences 1 \
-  >  --output-strains filtered_strains.txt > /dev/null
+  >  --output-strains filtered_strains.txt 2>/dev/null
   $ sort filtered_strains.txt
   SEQ_3
diff --git a/tests/functional/filter/cram/subsample-skip-ambiguous-dates.t b/tests/functional/filter/cram/subsample-skip-ambiguous-dates.t
index fa183d0ca..b638bdd59 100644
--- a/tests/functional/filter/cram/subsample-skip-ambiguous-dates.t
+++ b/tests/functional/filter/cram/subsample-skip-ambiguous-dates.t
@@ -9,9 +9,16 @@ Strains with ambiguous years or months should be dropped and logged.
   >  --metadata "$TESTDIR/../data/metadata.tsv" \
   >  --group-by year month \
   >  --subsample-max-sequences 5 \
+  >  --subsample-seed 0 \
   >  --output-strains filtered_strains.txt \
   >  --output-log filtered_log.tsv > /dev/null
   WARNING: Asked to provide at most 5 sequences, but there are 6 groups.
+  Sampling probabilistically at 0.8203 sequences per group, meaning it is possible to have more than the requested maximum of 5 sequences after filtering.
+  8 strains were dropped during filtering
+  	1 was dropped during grouping due to ambiguous year information
+  	1 was dropped during grouping due to ambiguous month information
+  	6 were dropped because of subsampling criteria
+  4 strains passed all filters
   $ grep "SG_018" filtered_log.tsv | cut -f 1-2
   SG_018\tskip_group_by_with_ambiguous_month (esc)
   $ grep "COL/FLR_00024/2015" filtered_log.tsv | cut -f 1-2
@@ -36,7 +43,7 @@ Group by 'week'. Check the number of strains that have been dropped due to ambig
   >  --sequences-per-group 1 \
   >  --subsample-seed 0 \
   >  --output-strains filtered_strains.txt \
-  >  --output-log filtered_log.tsv > /dev/null
+  >  --output-log filtered_log.tsv 2>/dev/null
   $ grep "skip_group_by_with_ambiguous_year" filtered_log.tsv | wc -l
   \s*1 (re)
   $ grep "skip_group_by_with_ambiguous_month" filtered_log.tsv | wc -l