United States v. Metropolitan Water Reclamation , 792 F.3d 821 ( 2015 )

In the
United States Court of Appeals
For the Seventh Circuit
____________________
Nos.  14-­‐‑1776  &  14-­‐‑1777
UNITED  STATES  OF  AMERICA  and  STATE  OF  ILLINOIS,
Plaintiffs-­‐‑Appellees,
and
ALLIANCE  FOR  THE  GREAT  LAKES,  et  al.,
Intervening  Plaintiffs-­‐‑Appellants,
v.
METROPOLITAN   WATER   RECLAMATION   DISTRICT   OF   GREATER
CHICAGO,
Defendant-­‐‑Appellee.
____________________
Appeals  from  the  United  States  District  Court  for  the
Northern  District  of  Illinois,  Eastern  Division.
No.  11  C  8859  —  George  M.  Marovich,  Judge.
____________________
ARGUED  FEBRUARY  12,  2015  —  DECIDED  JULY  9,  2015
____________________
Before   EASTERBROOK,   KANNE,   and   HAMILTON,   Circuit
Judges.
EASTERBROOK,  Circuit  Judge.  So  much  of  the  Chicago  met-­‐‑
ropolitan   area   is   covered   with   concrete   or   other   impermea-­‐‑
ble   surfaces   that   the   remaining   ground   cannot   absorb   the
2   Nos.  14-­‐‑1776  &  14-­‐‑1777
water   from   heavy   rain.   The   excess   goes   into   a   combined
stormwater   and   sewer   system,   which   can   overflow   and   es-­‐‑
cape  through  outfalls  located  on  the  banks  of  canals  and  riv-­‐‑
ers.   In   1975   the   Metropolitan   Water   Reclamation   District,
which   manages   sewage   control   (including   purification
plants),   began   construction   on   an   ambitious   project   to   im-­‐‑
pound  water  until  it  can  be  cleaned  up  and  released  safely:
the  Tunnel  and  Reservoir  Plan,  sometimes  called  TARP  and
commonly  known  as  Deep  Tunnel.
About   110   miles   of   large-­‐‑diameter   tunnels,   as   much   as
350   feet   underground   (hence   “deep”   tunnel),   collect   runoff
water   and   sewage   during   rainfall.   But   these   tunnels,   large
and  extensive  as  they  are,  can  hold  “only”  2.3  billion  gallons
of  water,  and  heavy  or  extended  rain  may  exceed  that  capac-­‐‑
ity.  The  plan  therefore  includes  reservoirs,  to  which  the  tun-­‐‑
nels  direct  their  contents  during  high-­‐‑inflow  conditions.  Two
reservoirs,  which  between  them  can  hold  3.4  billion  gallons,
are   operational   today.   One   of   these   is   scheduled   to   be   re-­‐‑
placed  later  this  year  by  the  Thornton  Composite  Reservoir,
which   can   accommodate   4.8   billion   gallons   from   TARP.
(This   reservoir   can   hold   7.9   billion   gallons,   but   3.1   billion
gallons   of   that   capacity   is   for   overflow   from   Thorn   Creek
and  is  not  counted  as  part  of  the  Deep  Tunnel  system.)  The
final   piece   of   the   system,   the   McCook   Reservoir,   is   sched-­‐‑
uled  for  completion  in  2029  with  a  capacity  of  10  billion  gal-­‐‑
lons  (and  an  interim  capacity  of  3.5  billion  gallons  by  2017).
Deep  Tunnel’s  final  capacity  will  be  17.5  billion  gallons.
The  Thornton  and  McCook  reservoirs  have  taken  a  long
time   to   build   because   both   will   occupy   worked-­‐‑out   lime-­‐‑
stone   quarries.   The   demand   for   limestone,   which   has   de-­‐‑
clined  in  recent  years,  affects  the  date  of  completion.  (Paying
Nos.  14-­‐‑1776  &  14-­‐‑1777   3
to  have  limestone  dug  up  in  advance  of  demand  from  build-­‐‑
ing  and  roadwork  projects  not  only  would  be  expensive  but
also  would  require  the  acquisition  of  land  on  which  to  dump
huge  piles  of  limestone,  which  would  be  unsightly  and  also
create  environmental  problems  as  minerals  leached  from  the
limestone   during   rain.)   But   the   delay   in   finishing   the   reser-­‐‑
voirs,   plus   some   problems   in   the   design   of   the   tunnels   that
can   require   the   District   to   hold   them   below   capacity,   have
hindered   the   system’s   ability   to   prevent   stormwater   and
sewage   releases.   The   record   does   not   show   how   many   un-­‐‑
treated  releases  occur  annually  at  the  approximately  375  lo-­‐‑
cations  along  rivers  and  canals  where  this  is  possible.  It  does
show  that  flows  from  all  sources  (including  direct  rain  run-­‐‑
off)   are   so   substantial   approximately   once   a   year   that   the
rivers  and  canals  themselves  would  overflow  unless  permit-­‐‑
ted  to  discharge  into  Lake  Michigan  (normally  locks  prevent
this,  but  they  are  opened  when  necessary  to  keep  the  rivers’
levels  under  control).
In   2011   the   United   States   and   the   State   of   Illinois   jointly
filed  this  suit,  under  sections  301  and  309  of  the  Clean  Water
Act,   33   U.S.C.   §§  1311,   1319,   seeking   an   order   that   the   Dis-­‐‑
trict  improve  the  TARP’s  performance,  accelerate  its  comple-­‐‑
tion  date,  and  do  more  to  contain  and  mitigate  overflows  in
the  interim.  The  district  court  permitted  the  Alliance  for  the
Great   Lakes   and   four   other   environmental   organizations
(collectively   the   Alliance)   to   intervene   under   33   U.S.C.
§1365(b)(1)(B).   See   2012   U.S.   Dist.   LEXIS   111223   (N.D.   Ill.
Aug.  7,  2012).
The   complaint   was   accompanied   by   a   proposed   consent
decree,  which  the  pollution-­‐‑control  agencies  had  been  nego-­‐‑
tiating   with   the   District   for   four   years.   The   settlement   re-­‐‑
4   Nos.  14-­‐‑1776  &  14-­‐‑1777
quires  the  District  to  complete  the  Deep  Tunnel  project,  meet
operational   criteria   when   construction   has   been   completed,
monitor   the   system’s   performance,   develop   additional
measures   if   needed   to   attain   compliance   with   the   Act   and
applicable  permits  in  the  interim,  and  maintain  the  decree  in
force  until  the  district  court  concludes  that  compliance  with
the   Act   has   been   achieved.   The   Alliance   opposed   this   pro-­‐‑
posal,  contending  that  it  requires  the  District  to  do  too  little
and   takes   too   long   even   for   what   it   accomplishes.   In   a
lengthy  opinion,  the  district  court  rejected  the  Alliance’s  pro-­‐‑
test   and   entered   the   proposed   decree.   2014   U.S.   Dist.   LEXIS
2049   (N.D.   Ill.   Jan.   6,   2014).   The   opinion   contains   many   de-­‐‑
tails   about   the   settlement   that   we   do   not   need   to   recapitu-­‐‑
late,  though  we  describe  a  few  details  later.
The   district   judge   also   concluded   that   the   settlement
binds  the  Alliance,  and  we  start  with  this  decision.  A  consent
decree  is  at  base  a  contract,  see  United  States  v.  ITT  Continen-­‐‑
tal  Baking  Co.,  420  U.S.  223  (1975),  and  the  Alliance  asks  how
it  can  be  bound  by  a  contract  to  which  it  did  not  agree.  The
district   court’s   answer   was   “res   judicata”   (the   combination
of  issue  and  claim  preclusion),  but  that  doctrine  prevents  one
party  from  litigating  the  same  claim  or  issue  in  multiple  suits.
See  Robinson  v.  Harvey,  617  F.3d  915,  916  (7th  Cir.  2010).  The
Alliance  is  not  the  United  States  or  the  State  of  Illinois,  and
there  is  only  one  lawsuit.  The  district  court  relied  on  Friends
of   Milwaukee’s   Rivers   v.   Milwaukee   Metropolitan   Sewerage   Dis-­‐‑
trict,  382  F.3d  743,  757–65  (7th  Cir.  2004),  and  United  States  v.
Metropolitan   St.   Louis   Sewer   District,   952   F.2d   1040   (8th   Cir.
1992),   but   each   of   these   decisions   dealt   with   multiple   suits
(filed  only  a  few  hours  apart  in  Friends  of  Milwaukee’s  Rivers).
Nos.  14-­‐‑1776  &  14-­‐‑1777   5
Our  situation  is  governed  not  by  principles  of  res  judicata
but   by   the   fact   that   legislation   may   allocate   litigation   rights
to   an   agency   that   represents   the   whole   public.   See   Taylor   v.
Sturgell,  553  U.S.  880,  895  (2008)  (discussing  the  sixth  excep-­‐‑
tion  to  the  norm  that  principles  of  virtual  representation  do
not   generally   bind   one   person   to   the   results   of   another’s
suit);  Restatement  (Second)  of  Judgments  §41(1)(d).  The  statute
modifying  the  common  law  is  §1365(b),  which  provides:
No  action  may  be  commenced—
(1)   under   subsection   (a)(1)   of   this   section   [which   authorizes
private  suits  to  enforce  the  Clean  Water  Act]—
(A)  prior  to  sixty  days  after  the  plaintiff  has  given  no-­‐‑
tice  of  the  alleged  violation  (i)  to  the  Administrator  [of
the  EPA],  (ii)  to  the  State  in  which  the  alleged  violation
occurs,  and  (iii)  to  any  alleged  violator  of  the  standard,
limitation,  or  order,  or
(B)  if  the  Administrator  or  State  has  commenced  and  is
diligently   prosecuting   a   civil   or   criminal   action   in   a
court  of  the  United  States,  or  a  State  to  require  compli-­‐‑
ance  with  the  standard,  limitation,  or  order,  but  in  any
such  action  in  a  court  of  the  United  States  any  citizen
may  intervene  as  a  matter  of  right.
Section  1365(b)(1)(B)  tells  us  that  no  private  litigation  may  be
“commenced”  if  the  EPA  or  a  state  “has  commenced  and  is
diligently  prosecuting  a  civil  …  action”  about  the  same  mat-­‐‑
ter  the  private  litigant  wants  to  raise.
We  held  in  Friends  of  Milwaukee’s  Rivers  that  this  language
also   means   that   the   resolution   (including   a   settlement)   of   a
federal   or   state   suit   is   binding   on   a   private   litigant   whose
suit   was   filed   after   the   state   or   federal   government’s,   if   the
state   or   federal   action   was   diligently   prosecuted.   Accord,
Louisiana   Environmental   Action   Network   v.   Baton   Rouge,   677
6   Nos.  14-­‐‑1776  &  14-­‐‑1777
F.3d  737,  749–50  (5th  Cir.  2012);  Piney  Run  Preservation  Asso-­‐‑
ciation  v.  Carroll  County,  523  F.3d  453,  459–60  (4th  Cir.  2008).
The  United  States  and  Illinois  contend  that  the  outcome  of  a
governmental   suit   is   equally   conclusive   for   private   claims
asserted  by  intervenors.
This  leads  the  Alliance  to  ask  what  the  point  of  interven-­‐‑
tion  might  be—for  intervention,  no  less  than  a  ban  on  stand-­‐‑
alone   private   litigation,   is   part   of   this   scheme.   The   answer,
we  think,  is  that  intervention  carries  four  rights:  to  introduce
evidence  if  the  case  goes  to  trial;  to  object  to  a  proposed  set-­‐‑
tlement   (a   right   the   Alliance   has   exercised);   to   appeal   if   the
intervenor  thinks  that  the  government  has  accomplished  too
little   (another   right   the   Alliance   has   exercised);   and   to   en-­‐‑
force  any  judgment,  just  as  the  United  States  and  Illinois  can
do.   This   decree   orders   the   District   to   come   into   compliance
with   the   Act   and   its   permits,   providing   details   about   how
and   when.   If   the   District   falls   short,   either   in   implementing
the  interim  measures  or  achieving  compliance  by  2029,  then
the   Alliance   can   ask   the   district   court   for   relief.   And   if   the
District  asks  the  district  court  to  dissolve  the  decree  in  2030,
the  Alliance,  as  a  party,  can  protest  that  too  and  appeal  from
an  adverse  decision.
That  is  a  more  modest  role  than  a  full-­‐‑fledged  independ-­‐‑
ent   litigator   would   have,   but   §1365(b)(1)(B)   tells   us   that   a
private  party  is  not  supposed  to  be  a  full-­‐‑fledged  independ-­‐‑
ent   litigator,   if   the   state   or   federal   government   diligently
prosecutes  a  suit.  This  is  also  the  implication  of  the  Supreme
Court’s  observation  that  private  intervenors  are  supposed  to
“supplement   rather   than   to   supplant”   public   litigation.
Gwaltney  of  Smithfield,  Ltd.  v.  Chesapeake  Bay  Foundation,  Inc.,
484  U.S.  49,  60  (1987).  If  the  Alliance  could  carry  on  just  as  if
Nos.  14-­‐‑1776  &  14-­‐‑1777   7
it  were  the  plaintiff  in  a  separate  suit,  that  would  “supplant”
the   governmental   case.   Why   would   anyone   settle   with   the
EPA  or  a  state,  if  the  settlement  did  not  buy  peace?  The  Dis-­‐‑
trict  made  costly  promises,  but  if  the  Alliance  is  right  then  it
got  nothing  in  return,  for  the  Alliance  can  carry  on  with  the
suit.   And   if   the   Alliance   also   settled,   then   another   person
could   intervene   to   demand   still   more   relief.   Depriving   the
original   parties   of   their   ability   to   settle   disputes   is   not   con-­‐‑
sistent   with   the   observation   in   Gwaltney,   the   holding   of
Friends  of  Milwaukee’s  Rivers,  or  the  structure  of  §1365(b).
Section  1365(b)(1)(B)  blocks  stand-­‐‑alone  private  litigation
only  if  the  public  agencies  “diligently”  prosecute  their  suits.
The  Alliance  maintains  that  if  there  is  a  chance  that  the  con-­‐‑
sent  decree  will  fail  to  achieve  full  compliance  with  the  Act
and  all  permits,  then  the  government  has  not  prosecuted  the
claims  diligently.  That  can’t  be  right.  Even  the  most  diligent
litigator   may   conclude   that   settlement   is   the   best   option—if
only  because  it  frees  up  enforcement  resources  for  use  else-­‐‑
where—and   to   achieve   a   settlement   a   litigant   must   accept
something  less  than  the  most  favorable  outcome.
In   Friends   of   Milwaukee’s   Rivers   we   concluded   that   a   set-­‐‑
tlement   likely   to   achieve   the   principal   enforcement   goals
demonstrates  diligent  prosecution.  382  F.3d  at  759–60.  For  a
settled   case,   this   pretty   much   turns   the   diligence   issue   into
the  question  whether  the  proposed  consent  decree  is  reason-­‐‑
able.   We   remanded   in   Friends   of   Milwaukee’s   Rivers   because
the   district   court   had   approved   that   settlement   without   in-­‐‑
quiring   how   likely   it   was   to   be   an   adequate   solution.   Here,
by  contrast,  the  district  court  considered  the  proposal’s  ade-­‐‑
quacy  at  length  and  evaluated  the  Alliance’s  objections.  This
enables   us   to   merge   the   inquiries:   if   the   consent   decree   is   a
8   Nos.  14-­‐‑1776  &  14-­‐‑1777
reasonable  settlement  likely  to  bring  about  compliance  with
the  Act,  it  also  demonstrates  diligent  prosecution.
The  Clean  Water  Act  forbids  “the  discharge  of  any  pollu-­‐‑
tant”   (a   defined   phrase,   see   South   Florida   Water   Management
District   v.   Miccosukee   Tribe,   541   U.S.   95   (2004)),   except   when
authorized   by   permits.   The   District   holds   multiple   permits
that   govern   releases   from   its   system.   These   permits   recog-­‐‑
nize  the  inevitability  of  discharges  during  storms  but  impose
three   kinds   of   conditions:   the   discharged   water   must   have
minimum   oxygen   levels,   must   keep   solid   matter   (“floata-­‐‑
bles”)   under   a   specified   level,   and   must   provide   for   some
rudimentary   treatment   (what   the   permits   call   “primary
treatment  …  with  adequate  retention  time”).
The   federal   and   state   complaint   alleges   that   some   over-­‐‑
flows   do   not   meet   one   or   more   of   these   three   conditions.   If
the   Deep   Tunnel   plan,   augmented   by   the   provisions   of   the
consent  decree,  succeeds,  then  by  2029  (if  not  earlier)  it  will
be   possible   to   retain   the   water   long   enough   to   route   it
through   the   District’s   seven   treatment   plants   at   a   rate   that
will  allow  full  treatment.  But  the  Alliance  predicts  that  it  will
not  succeed.  The  district  court  rejected  the  Alliance’s  protests
in  part  because  the  alternatives  it  proposed  would  be  costly
(more  than  $1  billion),  much  of  which  would  turn  out  to  be
wasted   if   the   state   and   federal   government   are   right   about
the  plan’s  adequacy,  and  because  even  if  the  Alliance’s  pre-­‐‑
dictions   are   right   the   rate   of   backflows   from   rivers   and   ca-­‐‑
nals  into  Lake  Michigan  will  drop  from  once  a  year  to  once  a
decade  when  the  system  is  finished.
The  Alliance’s  main  argument  on  appeal  can  be  summed
up  as:  “It  just  won’t  work.”  The  argument  has  three  compo-­‐‑
nents.  First,  a  study  completed  in  1994,  and  the  testimony  of
Nos.  14-­‐‑1776  &  14-­‐‑1777   9
one   of   the   District’s   consultants   before   the   Illinois   Pollution
Control   Board,   state   that   heavy   rainfall   will   overwhelm   the
Deep   Tunnel   system.   But   the   EPA,   the   Pollution   Control
Board,   and   the   District   are   more   optimistic   about   the   sys-­‐‑
tem’s   likely   functioning   when   completed.   Dissatisfied   with
the  1994  study,  which  was  based  on  only  one  modeled  rain-­‐‑
fall   event   and   did   not   analyze   the   system’s   current   opera-­‐‑
tional  plan,  the  EPA  conducted  its  own.  This  new  study,  re-­‐‑
leased  in  2009,  concluded  that  the  finished  Deep  Tunnel  sys-­‐‑
tem   will   work,   at   least   during   average   and   above-­‐‑average
rainfall   years.   (The   study   did   not   consider   once-­‐‑in-­‐‑100-­‐‑year
events   and   other   huge   storms.)   Recent   data   covering   2005
through  2011  in  the  Upper  Des  Plaines  segment  of  the  TARP
system,  which  is  finished  (including  its  full  reservoir  capaci-­‐‑
ty),   show   an   average   of   fewer   than   two   overflow   events   a
year,   and   this   period   includes   2008,   which   had   the   largest
recorded   rainfall   in   the   Chicago   area’s   history.   The   district
court  was  entitled  to  think  a  wait-­‐‑and-­‐‑see  approach  reason-­‐‑
able.  The  best  way  to  decide  between  competing  predictions
is  to  see  what  happens.
But   the   Alliance   maintains   that   the   consent   decree   itself
concedes  that  the  system  won’t  work.  That  is  because  the  de-­‐‑
cree   authorizes   untreated   discharges   when   (a)   the   tunnels
are  full,  or  (b)  a  “transient  event”  occurs—roughly,  when  the
rate  of  inflow  is  so  great  that  allowing  it  to  continue  would
damage  the  tunnel  system,  and  the  District  must  close  one  or
more  gates  to  protect  it.  The  Alliance  says  that  this  will  occur
whenever  there  is  a  “heavy  rainstorm”  but  does  not  quantify
that  term  (how  heavy?  how  often?  how  much  untreated  wa-­‐‑
ter   will   be   discharged?).   More   important,   however,   is   the
structure   of   the   Clean   Water   Act:   discharges   are   forbidden
except   when   authorized.   If   the   EPA   and   the   Pollution   Con-­‐‑
10   Nos.  14-­‐‑1776  &  14-­‐‑1777
trol   Board   are   willing   to   authorize   untreated   discharges
when   there   is   no   alternative   (and   that’s   what   conditions   (a)
and  (b)  amount  to),  then  there’s  no  violation  of  the  Act.
It’s  not  as  if  the  Alliance  were  contending  that  the  tunnel
system  itself  should  be  enlarged  or  its  design  changed,  long
after   its   construction,   to   avoid   damage   from   particularly
large  inflows.  Water  weighs  62.4  pounds  per  cubic  foot,  or  8
pounds  a  gallon,  and  can  drop  a  long  distance  in  the  tunnel
system—greater   than   the   height   of   Niagara   Falls,   where
dropping   water   generates   vast   quantities   of   energy.   This
shows   the   power   of   gravity   and   the   need   to   protect   even
tunnels  with  thick  concrete  walls.  The  consent  decree  allows
the  District  to  protect  this  costly  pollution-­‐‑control  asset.
In   the   district   court,   the   Alliance   proposed   that   the   Dis-­‐‑
trict   build   treatment   plants   at   105   outfalls,   increasing   the
number   of   plants   from   7   large   ones   to   112   mostly   small
ones—just  in  case.  But  the  district  judge  sensibly  concluded
that  it  would  be  imprudent  to  spend  $1  billion  on  that  pro-­‐‑
ject   if   it   turns   out   that   the   system   will   be   largely   effective
when   the   reservoirs   are   completed,   for   by   the   time   the   105
new  plants  were  operational  they  might  be  unnecessary.  The
Alliance’s  appellate  brief  drops  this  proposal,  leaving  no  al-­‐‑
ternative  to  the  decree’s  toleration  of  the  inevitable.
The  Alliance’s  final  argument  that  the  decree  won’t  work
is   that   it   permits   the   release   of   floatables   in   excess   of   the
quantity   allowed   by   the   “Combined   Sewer   Overflow   Con-­‐‑
trol   Policy”   adopted   by   the   EPA   in   1994.   See   59   Fed.   Reg.
18,688   (Apr.   19,   1994).   The   district   court   found   otherwise,
concluding   that   after   the   Deep   Tunnel   system   is   complete
the  District  will  be  in  compliance  with  the  floatables  policy.
That  finding  is  not  clearly  erroneous.  If  the  Chicago  area  had
Nos.  14-­‐‑1776  &  14-­‐‑1777   11
a  different  kind  of  sewer  system,  or  if  Deep  Tunnel  had  been
designed  differently,  then  a  consent  decree  might  have  been
able  to  achieve  full  compliance  faster.  That  the  settlement  of
a   given   case   takes   the   world   as   it   exists,   however,   does   not
show  a  lack  of  diligent  prosecution  or  a  substantively  unrea-­‐‑
sonable  outcome.
The  EPA  and  the  State  sought  to  deal  with  the  limitations
imposed   by   the   design   of   the   sewer   system   and   the   Deep
Tunnel   project   by   using   realistically   available   options.   The
District  already  has  two  pontoon  boats  that  deal  with  floata-­‐‑
bles;   the   decree   requires   it   to   add   two   specially   designed
skimmer  boats  to  keep  floatables  under  control  (even  if  their
source  is  something  other  than  the  District’s  system)  and  to
put  a  boom  around  one  outfall  that  has  experienced  frequent
discharges.   (Booms   elsewhere   might   interfere   with   naviga-­‐‑
tion.)   Another   part   of   the   decree   requires   the   District   to
adopt  a  “green  infrastructure”  program  that  will  reduce  the
amount   of   water   flowing   into   the   system   during   rains.   The
District  must  supply  rain  barrels  to  catch  runoff  from  build-­‐‑
ings  and  increase  the  permeability  of  the  surface  so  that  the
ground   can   retain   more   water.   The   District   must   complete
the  reservoirs  on  schedule  and  pay  as  much  as  $5,000  a  day
for  failure  to  do  so;  a  desire  to  avoid  these  fines  may  lead  the
District  to  pay  the  limestone  miners  to  remove  the  rock  fast-­‐‑
er.  Once  each  of  the  new  reservoirs  is  completed,  the  District
must   provide   enough   on-­‐‑site   water   treatment   capacity   to
cover   “the   maximum   flow   accounting   for   all   hydraulic   and
hydrologic  factors  that  can  pass  through”  the  system.  It  must
provide   extra   pumps   if   needed   to   move   water   faster   from
reservoirs   to   treatment   plants,   so   that   storage   capacity   is
available  in  the  reservoirs  for  incoming  water.  And  if  moni-­‐‑
toring  reveals  that  these  steps  don’t  achieve  their  goal,  then
12   Nos.  14-­‐‑1776  &  14-­‐‑1777
the   District   must   come   up   with   and   implement   a   new   plan
that   will.   These   are   costly   promises   by   the   District,   and   the
district   court   did   not   abuse   its   discretion   (the   applicable
standard,  see  United  States  v.  George  A.  Whiting  Paper  Co.,  644
F.3d   368,   372   (7th   Cir.   2011))   in   concluding   that   the   decree
carries  a  reasonable  prospect  of  success.
As  the  Alliance  sees  things,  some  of  the  District’s  prom-­‐‑
ises  are  worthless  (or  at  least  incomplete)  because  the  decree
does   not   spell   out   “which   pollutants   will   be   monitored,   at
which   locations,   how   frequently,   or   by   what   method.”   And
it  does  not  specify  what,  precisely,  will  be  done  if  problems
remain  in  2029.  Some  other  consent  decrees  that  the  EPA  has
negotiated   do   include   these   things.   Yet   the   District   is   so
large,   and   the   locations   of   potential   outfalls   so   numerous,
that   it’s   just   not   practical   to   try   to   cover   all   details   in   one
document.  The  EPA  anticipates  working  out  details  as  time
passes   and   additional   reservoir   capacity   becomes   available
(which  influences  what  needs  to  be  looked  at  and  tested  for),
and   if   the   District   does   not   cooperate   the   court   can   afford
supplemental  relief.
As   for   what   happens   in   2029   or   later   if   untreated   dis-­‐‑
charges   continue   at   an   unacceptable   rate,   the   next   steps
ought   to   depend   on   what’s   not   then   working   well.   If   the
EPA  (or  a  court)  could  be  sure  in  2014  what  the  exact  nature
of   the   problem   (if   any)   would   be   in   2029,   then   it   would   be
sensible   to   start   planning   and   building   the   remedy   today;
but  if  either  there  won’t  be  a  serious  problem  in  2029,  or  the
problem   is   something   not   now   foreseen,   then   relying   on   a
2014  decree  for  the  solution  would  be  foolish.  Yogi  Berra  ob-­‐‑
served   that   it   is   hard   to   make   predictions,   especially   about
Nos.  14-­‐‑1776  &  14-­‐‑1777   13
the   future.   State   and   federal   agencies   are   entitled   to   rely
more  on  experience  and  less  on  predictions.
The  consent  decree  that  the  district  court  has  approved  is
reasonable  in  light  of  the  current  infrastructure,  the  costs  of
doing   things   differently   (no   one   proposes   to   build   a   new
sewer  system  or  redo  the  Deep  Tunnel  project),  and  the  lim-­‐‑
its  of  knowledge  about  what  will  happen  when  the  system  is
completed.  The  decree  is  the  outcome  of  diligent  prosecution
and   therefore   binds   would-­‐‑be   private   litigants   such   as   the
Alliance.
AFFIRMED